diff --git a/SCENARIO_CATALOG.md b/SCENARIO_CATALOG.md index 78ccc6a22ca197027b034a175d153d85f8ec4bac..0dab667657d05f93a578e794211f601cf4ab71ae 100644 --- a/SCENARIO_CATALOG.md +++ b/SCENARIO_CATALOG.md @@ -1,3 +1,13 @@ +> **RETIRED (curation decision).** The auto-generated 200-level `cat-*` +> corpus described below has been removed from the repo. It scored low +> on discrimination/rigor (see `SCENARIO_QUALITY.md`) and added +> memorization surface without distinct capability. The benchmark is now +> a small **curated, hand-authored** set (Art-of-War, economy/building +> planning, strategy, rush-hour, adversarial, strict-spec, perception, +> reasoning), each designed deliberately one at a time with a +> non-gameable win+fail and a named external-benchmark analogue. This +> file is kept only as historical design notes. + # OpenRA-Bench Scenario Catalog (research-grounded, 200 levels) Goal: a benchmark whose scenarios **generalize on reasoning** — we diff --git a/openra_bench/scenarios/packs/cat-c1-00.yaml b/openra_bench/scenarios/packs/cat-c1-00.yaml deleted file mode 100644 index c5764bfcd3d754fe709f3eb113d959954c807da7..0000000000000000000000000000000000000000 --- a/openra_bench/scenarios/packs/cat-c1-00.yaml +++ /dev/null @@ -1,63 +0,0 @@ -meta: - id: cat-c1-00 - title: 'Frontier Scouting #1' - capability: perception - real_world_meaning: Pathfinding is solved; deciding WHICH unexplored region to commit - to under partial information and a deadline is the real search problem. - robotics_analogue: 'UAV/UGV frontier selection: choose which unknown region to commit - a time-limited searcher to.' - author: catalog-generator -base_map: rush-hour-arena -base: - agent: - faction: allies - enemy: - faction: soviet - tools: - - move_units - - attack_unit - - stop - planning: true - termination: - max_ticks: 30000 - actors: - - type: jeep - owner: agent - position: - - 6 - - 6 - count: 2 - - type: e1 - owner: agent - position: - - 7 - - 8 - count: 2 -levels: - easy: - description: "Frontier Scouting: easy \u2014 the decision is harder (less info\ - \ / tighter clock / committed choice)." - win_condition: - all_of: - - explored_pct_gte: 38 - - within_ticks: 16000 - max_turns: 45 - medium: - description: "Frontier Scouting: medium \u2014 the decision is harder (less info\ - \ / tighter clock / committed choice)." - win_condition: - all_of: - - explored_pct_gte: 47 - - within_ticks: 12500 - max_turns: 53 - hard: - description: "Frontier Scouting: hard \u2014 the decision is harder (less info\ - \ / tighter clock / committed choice)." - win_condition: - all_of: - - explored_pct_gte: 56 - - units_lost_lte: 0 - - within_ticks: 9000 - max_turns: 61 - fail_condition: - units_lost_lte: -1 diff --git a/openra_bench/scenarios/packs/cat-c1-01.yaml b/openra_bench/scenarios/packs/cat-c1-01.yaml deleted file mode 100644 index f458a29c3b554337b992baf52965c6eeff3010bd..0000000000000000000000000000000000000000 --- a/openra_bench/scenarios/packs/cat-c1-01.yaml +++ /dev/null @@ -1,63 +0,0 @@ -meta: - id: cat-c1-01 - title: 'Frontier Scouting #2' - capability: perception - real_world_meaning: Pathfinding is solved; deciding WHICH unexplored region to commit - to under partial information and a deadline is the real search problem. - robotics_analogue: 'UAV/UGV frontier selection: choose which unknown region to commit - a time-limited searcher to.' - author: catalog-generator -base_map: rush-hour-arena -base: - agent: - faction: allies - enemy: - faction: soviet - tools: - - move_units - - attack_unit - - stop - planning: true - termination: - max_ticks: 30000 - actors: - - type: jeep - owner: agent - position: - - 6 - - 6 - count: 2 - - type: e1 - owner: agent - position: - - 7 - - 8 - count: 2 -levels: - easy: - description: "Frontier Scouting: easy \u2014 the decision is harder (less info\ - \ / tighter clock / committed choice)." - win_condition: - all_of: - - explored_pct_gte: 39 - - within_ticks: 15800 - max_turns: 45 - medium: - description: "Frontier Scouting: medium \u2014 the decision is harder (less info\ - \ / tighter clock / committed choice)." - win_condition: - all_of: - - explored_pct_gte: 48 - - within_ticks: 12300 - max_turns: 53 - hard: - description: "Frontier Scouting: hard \u2014 the decision is harder (less info\ - \ / tighter clock / committed choice)." - win_condition: - all_of: - - explored_pct_gte: 57 - - units_lost_lte: 0 - - within_ticks: 8800 - max_turns: 61 - fail_condition: - units_lost_lte: -1 diff --git a/openra_bench/scenarios/packs/cat-c1-02.yaml b/openra_bench/scenarios/packs/cat-c1-02.yaml deleted file mode 100644 index 8d5d821f855cf5e09d02b32b6432b9a3a43aa9fe..0000000000000000000000000000000000000000 --- a/openra_bench/scenarios/packs/cat-c1-02.yaml +++ /dev/null @@ -1,65 +0,0 @@ -meta: - id: cat-c1-02 - title: 'Frontier Scouting #3' - capability: perception - real_world_meaning: Pathfinding is solved; deciding WHICH unexplored region to commit to under partial - information and a deadline is the real search problem. - robotics_analogue: 'UAV/UGV frontier selection: choose which unknown region to commit a time-limited - searcher to.' - author: catalog-generator - status: quarantine - quarantine_reason: 'redundant cat-* over-generation (audit: keep 2 representatives/category)' -base_map: rush-hour-arena -base: - agent: - faction: allies - enemy: - faction: soviet - tools: - - move_units - - attack_unit - - stop - planning: true - termination: - max_ticks: 30000 - actors: - - type: jeep - owner: agent - position: - - 6 - - 6 - count: 2 - - type: e1 - owner: agent - position: - - 7 - - 8 - count: 2 -levels: - easy: - description: 'Frontier Scouting: easy — the decision is harder (less info / tighter clock / committed - choice).' - win_condition: - all_of: - - explored_pct_gte: 40 - - within_ticks: 15600 - max_turns: 45 - medium: - description: 'Frontier Scouting: medium — the decision is harder (less info / tighter clock / committed - choice).' - win_condition: - all_of: - - explored_pct_gte: 49 - - within_ticks: 12100 - max_turns: 53 - hard: - description: 'Frontier Scouting: hard — the decision is harder (less info / tighter clock / committed - choice).' - win_condition: - all_of: - - explored_pct_gte: 58 - - units_lost_lte: 0 - - within_ticks: 8600 - max_turns: 61 - fail_condition: - units_lost_lte: -1 diff --git a/openra_bench/scenarios/packs/cat-c1-03.yaml b/openra_bench/scenarios/packs/cat-c1-03.yaml deleted file mode 100644 index 45d28e607b98ea5a73a93491aeacb53e4eae3154..0000000000000000000000000000000000000000 --- a/openra_bench/scenarios/packs/cat-c1-03.yaml +++ /dev/null @@ -1,65 +0,0 @@ -meta: - id: cat-c1-03 - title: 'Frontier Scouting #4' - capability: perception - real_world_meaning: Pathfinding is solved; deciding WHICH unexplored region to commit to under partial - information and a deadline is the real search problem. - robotics_analogue: 'UAV/UGV frontier selection: choose which unknown region to commit a time-limited - searcher to.' - author: catalog-generator - status: quarantine - quarantine_reason: 'redundant cat-* over-generation (audit: keep 2 representatives/category)' -base_map: rush-hour-arena -base: - agent: - faction: allies - enemy: - faction: soviet - tools: - - move_units - - attack_unit - - stop - planning: true - termination: - max_ticks: 30000 - actors: - - type: jeep - owner: agent - position: - - 6 - - 6 - count: 2 - - type: e1 - owner: agent - position: - - 7 - - 8 - count: 2 -levels: - easy: - description: 'Frontier Scouting: easy — the decision is harder (less info / tighter clock / committed - choice).' - win_condition: - all_of: - - explored_pct_gte: 41 - - within_ticks: 15400 - max_turns: 45 - medium: - description: 'Frontier Scouting: medium — the decision is harder (less info / tighter clock / committed - choice).' - win_condition: - all_of: - - explored_pct_gte: 50 - - within_ticks: 11900 - max_turns: 53 - hard: - description: 'Frontier Scouting: hard — the decision is harder (less info / tighter clock / committed - choice).' - win_condition: - all_of: - - explored_pct_gte: 59 - - units_lost_lte: 0 - - within_ticks: 8400 - max_turns: 61 - fail_condition: - units_lost_lte: -1 diff --git a/openra_bench/scenarios/packs/cat-c1-04.yaml b/openra_bench/scenarios/packs/cat-c1-04.yaml deleted file mode 100644 index 722a9ec4d73d1f2918ec53f437a9ac645e3b80dd..0000000000000000000000000000000000000000 --- a/openra_bench/scenarios/packs/cat-c1-04.yaml +++ /dev/null @@ -1,65 +0,0 @@ -meta: - id: cat-c1-04 - title: 'Frontier Scouting #5' - capability: perception - real_world_meaning: Pathfinding is solved; deciding WHICH unexplored region to commit to under partial - information and a deadline is the real search problem. - robotics_analogue: 'UAV/UGV frontier selection: choose which unknown region to commit a time-limited - searcher to.' - author: catalog-generator - status: quarantine - quarantine_reason: 'redundant cat-* over-generation (audit: keep 2 representatives/category)' -base_map: rush-hour-arena -base: - agent: - faction: allies - enemy: - faction: soviet - tools: - - move_units - - attack_unit - - stop - planning: true - termination: - max_ticks: 30000 - actors: - - type: jeep - owner: agent - position: - - 6 - - 6 - count: 2 - - type: e1 - owner: agent - position: - - 7 - - 8 - count: 2 -levels: - easy: - description: 'Frontier Scouting: easy — the decision is harder (less info / tighter clock / committed - choice).' - win_condition: - all_of: - - explored_pct_gte: 42 - - within_ticks: 15200 - max_turns: 45 - medium: - description: 'Frontier Scouting: medium — the decision is harder (less info / tighter clock / committed - choice).' - win_condition: - all_of: - - explored_pct_gte: 51 - - within_ticks: 11700 - max_turns: 53 - hard: - description: 'Frontier Scouting: hard — the decision is harder (less info / tighter clock / committed - choice).' - win_condition: - all_of: - - explored_pct_gte: 60 - - units_lost_lte: 0 - - within_ticks: 8200 - max_turns: 61 - fail_condition: - units_lost_lte: -1 diff --git a/openra_bench/scenarios/packs/cat-c1-05.yaml b/openra_bench/scenarios/packs/cat-c1-05.yaml deleted file mode 100644 index 4ce0f8d17ed71d0f789c22d1af63a6efcf61997d..0000000000000000000000000000000000000000 --- a/openra_bench/scenarios/packs/cat-c1-05.yaml +++ /dev/null @@ -1,65 +0,0 @@ -meta: - id: cat-c1-05 - title: 'Frontier Scouting #6' - capability: perception - real_world_meaning: Pathfinding is solved; deciding WHICH unexplored region to commit to under partial - information and a deadline is the real search problem. - robotics_analogue: 'UAV/UGV frontier selection: choose which unknown region to commit a time-limited - searcher to.' - author: catalog-generator - status: quarantine - quarantine_reason: 'redundant cat-* over-generation (audit: keep 2 representatives/category)' -base_map: rush-hour-arena -base: - agent: - faction: allies - enemy: - faction: soviet - tools: - - move_units - - attack_unit - - stop - planning: true - termination: - max_ticks: 30000 - actors: - - type: jeep - owner: agent - position: - - 6 - - 6 - count: 2 - - type: e1 - owner: agent - position: - - 7 - - 8 - count: 2 -levels: - easy: - description: 'Frontier Scouting: easy — the decision is harder (less info / tighter clock / committed - choice).' - win_condition: - all_of: - - explored_pct_gte: 43 - - within_ticks: 15000 - max_turns: 45 - medium: - description: 'Frontier Scouting: medium — the decision is harder (less info / tighter clock / committed - choice).' - win_condition: - all_of: - - explored_pct_gte: 52 - - within_ticks: 11500 - max_turns: 53 - hard: - description: 'Frontier Scouting: hard — the decision is harder (less info / tighter clock / committed - choice).' - win_condition: - all_of: - - explored_pct_gte: 61 - - units_lost_lte: 0 - - within_ticks: 8000 - max_turns: 61 - fail_condition: - units_lost_lte: -1 diff --git a/openra_bench/scenarios/packs/cat-c10-00.yaml b/openra_bench/scenarios/packs/cat-c10-00.yaml deleted file mode 100644 index d68ecad5e7bb18f2fa89d59f0f31470a3cb76d12..0000000000000000000000000000000000000000 --- a/openra_bench/scenarios/packs/cat-c10-00.yaml +++ /dev/null @@ -1,80 +0,0 @@ -meta: - id: cat-c10-00 - title: 'Force Coordination #1' - capability: action - real_world_meaning: 'Coordinated multi-robot rendezvous: steer separated teams to - a common staging point on time.' - robotics_analogue: Drive split groups to converge simultaneously and survive. - author: catalog-generator -base_map: rush-hour-arena -base: - agent: - faction: allies - enemy: - faction: soviet - tools: - - move_units - - attack_unit - - stop - planning: true - termination: - max_ticks: 30000 - actors: - - type: jeep - owner: agent - position: - - 6 - - 6 - count: 2 - - type: jeep - owner: agent - position: - - 6 - - 34 - count: 2 - - type: e1 - owner: enemy - position: - - 110 - - 20 - count: 2 - stance: 2 -levels: - easy: - description: "Force Coordination: easy \u2014 the decision is harder (less info\ - \ / tighter clock / committed choice)." - win_condition: - all_of: - - all_units_in_region: - x: 55 - y: 20 - radius: 7 - - units_lost_lte: 1 - - within_ticks: 13000 - max_turns: 45 - medium: - description: "Force Coordination: medium \u2014 the decision is harder (less info\ - \ / tighter clock / committed choice)." - win_condition: - all_of: - - all_units_in_region: - x: 55 - y: 20 - radius: 6 - - units_lost_lte: 1 - - within_ticks: 10000 - max_turns: 53 - hard: - description: "Force Coordination: hard \u2014 the decision is harder (less info\ - \ / tighter clock / committed choice)." - win_condition: - all_of: - - all_units_in_region: - x: 55 - y: 20 - radius: 5 - - units_lost_lte: 0 - - within_ticks: 7000 - max_turns: 61 - fail_condition: - units_lost_lte: -1 diff --git a/openra_bench/scenarios/packs/cat-c10-01.yaml b/openra_bench/scenarios/packs/cat-c10-01.yaml deleted file mode 100644 index 03049c78f06cc4836de9c12d365d586c27d35a78..0000000000000000000000000000000000000000 --- a/openra_bench/scenarios/packs/cat-c10-01.yaml +++ /dev/null @@ -1,80 +0,0 @@ -meta: - id: cat-c10-01 - title: 'Force Coordination #2' - capability: action - real_world_meaning: 'Coordinated multi-robot rendezvous: steer separated teams to - a common staging point on time.' - robotics_analogue: Drive split groups to converge simultaneously and survive. - author: catalog-generator -base_map: rush-hour-arena -base: - agent: - faction: allies - enemy: - faction: soviet - tools: - - move_units - - attack_unit - - stop - planning: true - termination: - max_ticks: 30000 - actors: - - type: jeep - owner: agent - position: - - 6 - - 6 - count: 2 - - type: jeep - owner: agent - position: - - 6 - - 34 - count: 2 - - type: e1 - owner: enemy - position: - - 110 - - 20 - count: 2 - stance: 2 -levels: - easy: - description: "Force Coordination: easy \u2014 the decision is harder (less info\ - \ / tighter clock / committed choice)." - win_condition: - all_of: - - all_units_in_region: - x: 58 - y: 20 - radius: 7 - - units_lost_lte: 1 - - within_ticks: 13000 - max_turns: 45 - medium: - description: "Force Coordination: medium \u2014 the decision is harder (less info\ - \ / tighter clock / committed choice)." - win_condition: - all_of: - - all_units_in_region: - x: 58 - y: 20 - radius: 6 - - units_lost_lte: 1 - - within_ticks: 10000 - max_turns: 53 - hard: - description: "Force Coordination: hard \u2014 the decision is harder (less info\ - \ / tighter clock / committed choice)." - win_condition: - all_of: - - all_units_in_region: - x: 58 - y: 20 - radius: 5 - - units_lost_lte: 0 - - within_ticks: 7000 - max_turns: 61 - fail_condition: - units_lost_lte: -1 diff --git a/openra_bench/scenarios/packs/cat-c10-02.yaml b/openra_bench/scenarios/packs/cat-c10-02.yaml deleted file mode 100644 index e57684ba37b32d892a258103344f339c18903b0f..0000000000000000000000000000000000000000 --- a/openra_bench/scenarios/packs/cat-c10-02.yaml +++ /dev/null @@ -1,82 +0,0 @@ -meta: - id: cat-c10-02 - title: 'Force Coordination #3' - capability: action - real_world_meaning: 'Coordinated multi-robot rendezvous: steer separated teams to a common staging point - on time.' - robotics_analogue: Drive split groups to converge simultaneously and survive. - author: catalog-generator - status: quarantine - quarantine_reason: 'redundant cat-* over-generation (audit: keep 2 representatives/category)' -base_map: rush-hour-arena -base: - agent: - faction: allies - enemy: - faction: soviet - tools: - - move_units - - attack_unit - - stop - planning: true - termination: - max_ticks: 30000 - actors: - - type: jeep - owner: agent - position: - - 6 - - 6 - count: 2 - - type: jeep - owner: agent - position: - - 6 - - 34 - count: 2 - - type: e1 - owner: enemy - position: - - 110 - - 20 - count: 2 - stance: 2 -levels: - easy: - description: 'Force Coordination: easy — the decision is harder (less info / tighter clock / committed - choice).' - win_condition: - all_of: - - all_units_in_region: - x: 61 - y: 20 - radius: 7 - - units_lost_lte: 1 - - within_ticks: 13000 - max_turns: 45 - medium: - description: 'Force Coordination: medium — the decision is harder (less info / tighter clock / committed - choice).' - win_condition: - all_of: - - all_units_in_region: - x: 61 - y: 20 - radius: 6 - - units_lost_lte: 1 - - within_ticks: 10000 - max_turns: 53 - hard: - description: 'Force Coordination: hard — the decision is harder (less info / tighter clock / committed - choice).' - win_condition: - all_of: - - all_units_in_region: - x: 61 - y: 20 - radius: 5 - - units_lost_lte: 0 - - within_ticks: 7000 - max_turns: 61 - fail_condition: - units_lost_lte: -1 diff --git a/openra_bench/scenarios/packs/cat-c10-03.yaml b/openra_bench/scenarios/packs/cat-c10-03.yaml deleted file mode 100644 index fb49d9609468e2ede6a1953d2e6decf0e5e1aac5..0000000000000000000000000000000000000000 --- a/openra_bench/scenarios/packs/cat-c10-03.yaml +++ /dev/null @@ -1,82 +0,0 @@ -meta: - id: cat-c10-03 - title: 'Force Coordination #4' - capability: action - real_world_meaning: 'Coordinated multi-robot rendezvous: steer separated teams to a common staging point - on time.' - robotics_analogue: Drive split groups to converge simultaneously and survive. - author: catalog-generator - status: quarantine - quarantine_reason: 'redundant cat-* over-generation (audit: keep 2 representatives/category)' -base_map: rush-hour-arena -base: - agent: - faction: allies - enemy: - faction: soviet - tools: - - move_units - - attack_unit - - stop - planning: true - termination: - max_ticks: 30000 - actors: - - type: jeep - owner: agent - position: - - 6 - - 6 - count: 2 - - type: jeep - owner: agent - position: - - 6 - - 34 - count: 2 - - type: e1 - owner: enemy - position: - - 110 - - 20 - count: 2 - stance: 2 -levels: - easy: - description: 'Force Coordination: easy — the decision is harder (less info / tighter clock / committed - choice).' - win_condition: - all_of: - - all_units_in_region: - x: 64 - y: 20 - radius: 7 - - units_lost_lte: 1 - - within_ticks: 13000 - max_turns: 45 - medium: - description: 'Force Coordination: medium — the decision is harder (less info / tighter clock / committed - choice).' - win_condition: - all_of: - - all_units_in_region: - x: 64 - y: 20 - radius: 6 - - units_lost_lte: 1 - - within_ticks: 10000 - max_turns: 53 - hard: - description: 'Force Coordination: hard — the decision is harder (less info / tighter clock / committed - choice).' - win_condition: - all_of: - - all_units_in_region: - x: 64 - y: 20 - radius: 5 - - units_lost_lte: 0 - - within_ticks: 7000 - max_turns: 61 - fail_condition: - units_lost_lte: -1 diff --git a/openra_bench/scenarios/packs/cat-c10-04.yaml b/openra_bench/scenarios/packs/cat-c10-04.yaml deleted file mode 100644 index 71b9d30a44a950f2d9420fe290447e109ba266ab..0000000000000000000000000000000000000000 --- a/openra_bench/scenarios/packs/cat-c10-04.yaml +++ /dev/null @@ -1,82 +0,0 @@ -meta: - id: cat-c10-04 - title: 'Force Coordination #5' - capability: action - real_world_meaning: 'Coordinated multi-robot rendezvous: steer separated teams to a common staging point - on time.' - robotics_analogue: Drive split groups to converge simultaneously and survive. - author: catalog-generator - status: quarantine - quarantine_reason: 'redundant cat-* over-generation (audit: keep 2 representatives/category)' -base_map: rush-hour-arena -base: - agent: - faction: allies - enemy: - faction: soviet - tools: - - move_units - - attack_unit - - stop - planning: true - termination: - max_ticks: 30000 - actors: - - type: jeep - owner: agent - position: - - 6 - - 6 - count: 2 - - type: jeep - owner: agent - position: - - 6 - - 34 - count: 2 - - type: e1 - owner: enemy - position: - - 110 - - 20 - count: 2 - stance: 2 -levels: - easy: - description: 'Force Coordination: easy — the decision is harder (less info / tighter clock / committed - choice).' - win_condition: - all_of: - - all_units_in_region: - x: 67 - y: 20 - radius: 7 - - units_lost_lte: 1 - - within_ticks: 13000 - max_turns: 45 - medium: - description: 'Force Coordination: medium — the decision is harder (less info / tighter clock / committed - choice).' - win_condition: - all_of: - - all_units_in_region: - x: 67 - y: 20 - radius: 6 - - units_lost_lte: 1 - - within_ticks: 10000 - max_turns: 53 - hard: - description: 'Force Coordination: hard — the decision is harder (less info / tighter clock / committed - choice).' - win_condition: - all_of: - - all_units_in_region: - x: 67 - y: 20 - radius: 5 - - units_lost_lte: 0 - - within_ticks: 7000 - max_turns: 61 - fail_condition: - units_lost_lte: -1 diff --git a/openra_bench/scenarios/packs/cat-c10-05.yaml b/openra_bench/scenarios/packs/cat-c10-05.yaml deleted file mode 100644 index 5ec7e12f2962dda5a2ff1317c545e5057b33e73d..0000000000000000000000000000000000000000 --- a/openra_bench/scenarios/packs/cat-c10-05.yaml +++ /dev/null @@ -1,82 +0,0 @@ -meta: - id: cat-c10-05 - title: 'Force Coordination #6' - capability: action - real_world_meaning: 'Coordinated multi-robot rendezvous: steer separated teams to a common staging point - on time.' - robotics_analogue: Drive split groups to converge simultaneously and survive. - author: catalog-generator - status: quarantine - quarantine_reason: 'redundant cat-* over-generation (audit: keep 2 representatives/category)' -base_map: rush-hour-arena -base: - agent: - faction: allies - enemy: - faction: soviet - tools: - - move_units - - attack_unit - - stop - planning: true - termination: - max_ticks: 30000 - actors: - - type: jeep - owner: agent - position: - - 6 - - 6 - count: 2 - - type: jeep - owner: agent - position: - - 6 - - 34 - count: 2 - - type: e1 - owner: enemy - position: - - 110 - - 20 - count: 2 - stance: 2 -levels: - easy: - description: 'Force Coordination: easy — the decision is harder (less info / tighter clock / committed - choice).' - win_condition: - all_of: - - all_units_in_region: - x: 70 - y: 20 - radius: 7 - - units_lost_lte: 1 - - within_ticks: 13000 - max_turns: 45 - medium: - description: 'Force Coordination: medium — the decision is harder (less info / tighter clock / committed - choice).' - win_condition: - all_of: - - all_units_in_region: - x: 70 - y: 20 - radius: 6 - - units_lost_lte: 1 - - within_ticks: 10000 - max_turns: 53 - hard: - description: 'Force Coordination: hard — the decision is harder (less info / tighter clock / committed - choice).' - win_condition: - all_of: - - all_units_in_region: - x: 70 - y: 20 - radius: 5 - - units_lost_lte: 0 - - within_ticks: 7000 - max_turns: 61 - fail_condition: - units_lost_lte: -1 diff --git a/openra_bench/scenarios/packs/cat-c11-00.yaml b/openra_bench/scenarios/packs/cat-c11-00.yaml deleted file mode 100644 index d8eea82c8186ea8390262e938395adbf15a66cf9..0000000000000000000000000000000000000000 --- a/openra_bench/scenarios/packs/cat-c11-00.yaml +++ /dev/null @@ -1,87 +0,0 @@ -meta: - id: cat-c11-00 - title: 'Tempo / Timing Window #1' - capability: reasoning - real_world_meaning: 'Acting only inside a valid time window (takt-time): too early - or too late both fail.' - robotics_analogue: "Strike inside a window that opens late and closes \u2014 not\ - \ earliest, not biggest." - author: catalog-generator -base_map: rush-hour-arena -base: - agent: - faction: allies - enemy: - faction: soviet - tools: - - move_units - - attack_unit - - attack_move - - stop - planning: true - termination: - max_ticks: 30000 - actors: - - type: 2tnk - owner: agent - position: - - 8 - - 18 - count: 2 - - type: e1 - owner: agent - position: - - 9 - - 20 - count: 3 - - type: e1 - owner: enemy - position: - - 100 - - 20 - count: 2 - stance: 2 - - type: e3 - owner: enemy - position: - - 104 - - 24 - count: 1 - stance: 2 - - type: e1 - owner: enemy - position: - - 40 - - 10 - count: 1 - stance: 2 -levels: - easy: - description: "Tempo / Timing Window: easy \u2014 the decision is harder (less\ - \ info / tighter clock / committed choice)." - win_condition: - all_of: - - after_ticks: 1500 - - units_killed_gte: 2 - - within_ticks: 12000 - max_turns: 45 - medium: - description: "Tempo / Timing Window: medium \u2014 the decision is harder (less\ - \ info / tighter clock / committed choice)." - win_condition: - all_of: - - after_ticks: 2300 - - units_killed_gte: 3 - - within_ticks: 9500 - max_turns: 53 - hard: - description: "Tempo / Timing Window: hard \u2014 the decision is harder (less\ - \ info / tighter clock / committed choice)." - win_condition: - all_of: - - after_ticks: 3100 - - units_killed_gte: 4 - - within_ticks: 7000 - max_turns: 61 - fail_condition: - units_lost_lte: -1 diff --git a/openra_bench/scenarios/packs/cat-c11-01.yaml b/openra_bench/scenarios/packs/cat-c11-01.yaml deleted file mode 100644 index b89a5a9e847be92c1b9b51e213e3efae0ec27afc..0000000000000000000000000000000000000000 --- a/openra_bench/scenarios/packs/cat-c11-01.yaml +++ /dev/null @@ -1,87 +0,0 @@ -meta: - id: cat-c11-01 - title: 'Tempo / Timing Window #2' - capability: reasoning - real_world_meaning: 'Acting only inside a valid time window (takt-time): too early - or too late both fail.' - robotics_analogue: "Strike inside a window that opens late and closes \u2014 not\ - \ earliest, not biggest." - author: catalog-generator -base_map: rush-hour-arena -base: - agent: - faction: allies - enemy: - faction: soviet - tools: - - move_units - - attack_unit - - attack_move - - stop - planning: true - termination: - max_ticks: 30000 - actors: - - type: 2tnk - owner: agent - position: - - 8 - - 18 - count: 2 - - type: e1 - owner: agent - position: - - 9 - - 20 - count: 3 - - type: e1 - owner: enemy - position: - - 100 - - 20 - count: 2 - stance: 2 - - type: e3 - owner: enemy - position: - - 104 - - 24 - count: 1 - stance: 2 - - type: e1 - owner: enemy - position: - - 40 - - 10 - count: 1 - stance: 2 -levels: - easy: - description: "Tempo / Timing Window: easy \u2014 the decision is harder (less\ - \ info / tighter clock / committed choice)." - win_condition: - all_of: - - after_ticks: 1500 - - units_killed_gte: 2 - - within_ticks: 12000 - max_turns: 45 - medium: - description: "Tempo / Timing Window: medium \u2014 the decision is harder (less\ - \ info / tighter clock / committed choice)." - win_condition: - all_of: - - after_ticks: 2300 - - units_killed_gte: 3 - - within_ticks: 9500 - max_turns: 53 - hard: - description: "Tempo / Timing Window: hard \u2014 the decision is harder (less\ - \ info / tighter clock / committed choice)." - win_condition: - all_of: - - after_ticks: 3100 - - units_killed_gte: 4 - - within_ticks: 7000 - max_turns: 61 - fail_condition: - units_lost_lte: -1 diff --git a/openra_bench/scenarios/packs/cat-c11-02.yaml b/openra_bench/scenarios/packs/cat-c11-02.yaml deleted file mode 100644 index ceb68f773ff6924895d4e6a265be70a1ac26374a..0000000000000000000000000000000000000000 --- a/openra_bench/scenarios/packs/cat-c11-02.yaml +++ /dev/null @@ -1,88 +0,0 @@ -meta: - id: cat-c11-02 - title: 'Tempo / Timing Window #3' - capability: reasoning - real_world_meaning: 'Acting only inside a valid time window (takt-time): too early or too late both - fail.' - robotics_analogue: Strike inside a window that opens late and closes — not earliest, not biggest. - author: catalog-generator - status: quarantine - quarantine_reason: 'redundant cat-* over-generation (audit: keep 2 representatives/category)' -base_map: rush-hour-arena -base: - agent: - faction: allies - enemy: - faction: soviet - tools: - - move_units - - attack_unit - - attack_move - - stop - planning: true - termination: - max_ticks: 30000 - actors: - - type: 2tnk - owner: agent - position: - - 8 - - 18 - count: 2 - - type: e1 - owner: agent - position: - - 9 - - 20 - count: 3 - - type: e1 - owner: enemy - position: - - 100 - - 20 - count: 2 - stance: 2 - - type: e3 - owner: enemy - position: - - 104 - - 24 - count: 1 - stance: 2 - - type: e1 - owner: enemy - position: - - 40 - - 10 - count: 1 - stance: 2 -levels: - easy: - description: 'Tempo / Timing Window: easy — the decision is harder (less info / tighter clock / committed - choice).' - win_condition: - all_of: - - after_ticks: 1500 - - units_killed_gte: 2 - - within_ticks: 12000 - max_turns: 45 - medium: - description: 'Tempo / Timing Window: medium — the decision is harder (less info / tighter clock / - committed choice).' - win_condition: - all_of: - - after_ticks: 2300 - - units_killed_gte: 3 - - within_ticks: 9500 - max_turns: 53 - hard: - description: 'Tempo / Timing Window: hard — the decision is harder (less info / tighter clock / committed - choice).' - win_condition: - all_of: - - after_ticks: 3100 - - units_killed_gte: 4 - - within_ticks: 7000 - max_turns: 61 - fail_condition: - units_lost_lte: -1 diff --git a/openra_bench/scenarios/packs/cat-c11-03.yaml b/openra_bench/scenarios/packs/cat-c11-03.yaml deleted file mode 100644 index ea8f54b508a8a36d37135b9aaab88f855f3f0c24..0000000000000000000000000000000000000000 --- a/openra_bench/scenarios/packs/cat-c11-03.yaml +++ /dev/null @@ -1,88 +0,0 @@ -meta: - id: cat-c11-03 - title: 'Tempo / Timing Window #4' - capability: reasoning - real_world_meaning: 'Acting only inside a valid time window (takt-time): too early or too late both - fail.' - robotics_analogue: Strike inside a window that opens late and closes — not earliest, not biggest. - author: catalog-generator - status: quarantine - quarantine_reason: 'redundant cat-* over-generation (audit: keep 2 representatives/category)' -base_map: rush-hour-arena -base: - agent: - faction: allies - enemy: - faction: soviet - tools: - - move_units - - attack_unit - - attack_move - - stop - planning: true - termination: - max_ticks: 30000 - actors: - - type: 2tnk - owner: agent - position: - - 8 - - 18 - count: 2 - - type: e1 - owner: agent - position: - - 9 - - 20 - count: 3 - - type: e1 - owner: enemy - position: - - 100 - - 20 - count: 2 - stance: 2 - - type: e3 - owner: enemy - position: - - 104 - - 24 - count: 1 - stance: 2 - - type: e1 - owner: enemy - position: - - 40 - - 10 - count: 1 - stance: 2 -levels: - easy: - description: 'Tempo / Timing Window: easy — the decision is harder (less info / tighter clock / committed - choice).' - win_condition: - all_of: - - after_ticks: 1500 - - units_killed_gte: 2 - - within_ticks: 12000 - max_turns: 45 - medium: - description: 'Tempo / Timing Window: medium — the decision is harder (less info / tighter clock / - committed choice).' - win_condition: - all_of: - - after_ticks: 2300 - - units_killed_gte: 3 - - within_ticks: 9500 - max_turns: 53 - hard: - description: 'Tempo / Timing Window: hard — the decision is harder (less info / tighter clock / committed - choice).' - win_condition: - all_of: - - after_ticks: 3100 - - units_killed_gte: 4 - - within_ticks: 7000 - max_turns: 61 - fail_condition: - units_lost_lte: -1 diff --git a/openra_bench/scenarios/packs/cat-c11-04.yaml b/openra_bench/scenarios/packs/cat-c11-04.yaml deleted file mode 100644 index f3d6bdea8faf5c06d169f1aea235cdf14e08ce52..0000000000000000000000000000000000000000 --- a/openra_bench/scenarios/packs/cat-c11-04.yaml +++ /dev/null @@ -1,88 +0,0 @@ -meta: - id: cat-c11-04 - title: 'Tempo / Timing Window #5' - capability: reasoning - real_world_meaning: 'Acting only inside a valid time window (takt-time): too early or too late both - fail.' - robotics_analogue: Strike inside a window that opens late and closes — not earliest, not biggest. - author: catalog-generator - status: quarantine - quarantine_reason: 'redundant cat-* over-generation (audit: keep 2 representatives/category)' -base_map: rush-hour-arena -base: - agent: - faction: allies - enemy: - faction: soviet - tools: - - move_units - - attack_unit - - attack_move - - stop - planning: true - termination: - max_ticks: 30000 - actors: - - type: 2tnk - owner: agent - position: - - 8 - - 18 - count: 2 - - type: e1 - owner: agent - position: - - 9 - - 20 - count: 3 - - type: e1 - owner: enemy - position: - - 100 - - 20 - count: 2 - stance: 2 - - type: e3 - owner: enemy - position: - - 104 - - 24 - count: 1 - stance: 2 - - type: e1 - owner: enemy - position: - - 40 - - 10 - count: 1 - stance: 2 -levels: - easy: - description: 'Tempo / Timing Window: easy — the decision is harder (less info / tighter clock / committed - choice).' - win_condition: - all_of: - - after_ticks: 1500 - - units_killed_gte: 2 - - within_ticks: 12000 - max_turns: 45 - medium: - description: 'Tempo / Timing Window: medium — the decision is harder (less info / tighter clock / - committed choice).' - win_condition: - all_of: - - after_ticks: 2300 - - units_killed_gte: 3 - - within_ticks: 9500 - max_turns: 53 - hard: - description: 'Tempo / Timing Window: hard — the decision is harder (less info / tighter clock / committed - choice).' - win_condition: - all_of: - - after_ticks: 3100 - - units_killed_gte: 4 - - within_ticks: 7000 - max_turns: 61 - fail_condition: - units_lost_lte: -1 diff --git a/openra_bench/scenarios/packs/cat-c12-00.yaml b/openra_bench/scenarios/packs/cat-c12-00.yaml deleted file mode 100644 index 781f578e5c98c4aa582655682514e6e6a1e4c3f2..0000000000000000000000000000000000000000 --- a/openra_bench/scenarios/packs/cat-c12-00.yaml +++ /dev/null @@ -1,86 +0,0 @@ -meta: - id: cat-c12-00 - title: 'Error Recovery / Replan #1' - capability: reasoning - real_world_meaning: "Detecting plan failure and repairing it under reduced means\ - \ and a deadline \u2014 receding-horizon replanning." - robotics_analogue: After a setback, rebuild/repair and re-establish capability under - a hard clock. - author: catalog-generator -base_map: rush-hour-arena -base: - agent: - faction: allies - enemy: - faction: soviet - tools: - - build - - place_building - - move_units - - deploy - - stop - planning: true - termination: - max_ticks: 40000 - actors: - - type: fact - owner: agent - position: - - 10 - - 18 - - type: powr - owner: agent - position: - - 14 - - 18 - - type: tent - owner: agent - position: - - 12 - - 21 - - type: jeep - owner: agent - position: - - 8 - - 16 - - type: e1 - owner: enemy - position: - - 114 - - 34 - count: 1 - stance: 2 -levels: - easy: - description: "Error Recovery / Replan: easy \u2014 the decision is harder (less\ - \ info / tighter clock / committed choice)." - win_condition: - all_of: - - has_building: tent - - building_total_gte: 4 - - after_ticks: 1200 - - within_ticks: 22000 - max_turns: 45 - medium: - description: "Error Recovery / Replan: medium \u2014 the decision is harder (less\ - \ info / tighter clock / committed choice)." - win_condition: - all_of: - - has_building: tent - - building_total_gte: 5 - - after_ticks: 1200 - - within_ticks: 17000 - max_turns: 53 - hard: - description: "Error Recovery / Replan: hard \u2014 the decision is harder (less\ - \ info / tighter clock / committed choice)." - win_condition: - all_of: - - has_building: tent - - building_total_gte: 6 - - after_ticks: 1200 - - within_ticks: 12000 - max_turns: 61 - fail_condition: - units_lost_lte: -1 -starting_cash: 3000 diff --git a/openra_bench/scenarios/packs/cat-c12-01.yaml b/openra_bench/scenarios/packs/cat-c12-01.yaml deleted file mode 100644 index 5413038a321f642d5b980a7fb849870d5987db31..0000000000000000000000000000000000000000 --- a/openra_bench/scenarios/packs/cat-c12-01.yaml +++ /dev/null @@ -1,86 +0,0 @@ -meta: - id: cat-c12-01 - title: 'Error Recovery / Replan #2' - capability: reasoning - real_world_meaning: "Detecting plan failure and repairing it under reduced means\ - \ and a deadline \u2014 receding-horizon replanning." - robotics_analogue: After a setback, rebuild/repair and re-establish capability under - a hard clock. - author: catalog-generator -base_map: rush-hour-arena -base: - agent: - faction: allies - enemy: - faction: soviet - tools: - - build - - place_building - - move_units - - deploy - - stop - planning: true - termination: - max_ticks: 40000 - actors: - - type: fact - owner: agent - position: - - 10 - - 18 - - type: powr - owner: agent - position: - - 14 - - 18 - - type: tent - owner: agent - position: - - 12 - - 21 - - type: jeep - owner: agent - position: - - 8 - - 16 - - type: e1 - owner: enemy - position: - - 114 - - 34 - count: 1 - stance: 2 -levels: - easy: - description: "Error Recovery / Replan: easy \u2014 the decision is harder (less\ - \ info / tighter clock / committed choice)." - win_condition: - all_of: - - has_building: tent - - building_total_gte: 4 - - after_ticks: 1200 - - within_ticks: 22000 - max_turns: 45 - medium: - description: "Error Recovery / Replan: medium \u2014 the decision is harder (less\ - \ info / tighter clock / committed choice)." - win_condition: - all_of: - - has_building: tent - - building_total_gte: 5 - - after_ticks: 1200 - - within_ticks: 17000 - max_turns: 53 - hard: - description: "Error Recovery / Replan: hard \u2014 the decision is harder (less\ - \ info / tighter clock / committed choice)." - win_condition: - all_of: - - has_building: tent - - building_total_gte: 6 - - after_ticks: 1200 - - within_ticks: 12000 - max_turns: 61 - fail_condition: - units_lost_lte: -1 -starting_cash: 3000 diff --git a/openra_bench/scenarios/packs/cat-c12-02.yaml b/openra_bench/scenarios/packs/cat-c12-02.yaml deleted file mode 100644 index 9e19f34fa27195184f6e9b8c41fa21a01d49be77..0000000000000000000000000000000000000000 --- a/openra_bench/scenarios/packs/cat-c12-02.yaml +++ /dev/null @@ -1,87 +0,0 @@ -meta: - id: cat-c12-02 - title: 'Error Recovery / Replan #3' - capability: reasoning - real_world_meaning: Detecting plan failure and repairing it under reduced means and a deadline — receding-horizon - replanning. - robotics_analogue: After a setback, rebuild/repair and re-establish capability under a hard clock. - author: catalog-generator - status: quarantine - quarantine_reason: 'redundant cat-* over-generation (audit: keep 2 representatives/category)' -base_map: rush-hour-arena -base: - agent: - faction: allies - enemy: - faction: soviet - tools: - - build - - place_building - - move_units - - deploy - - stop - planning: true - termination: - max_ticks: 40000 - actors: - - type: fact - owner: agent - position: - - 10 - - 18 - - type: powr - owner: agent - position: - - 14 - - 18 - - type: tent - owner: agent - position: - - 12 - - 21 - - type: jeep - owner: agent - position: - - 8 - - 16 - - type: e1 - owner: enemy - position: - - 114 - - 34 - count: 1 - stance: 2 -levels: - easy: - description: 'Error Recovery / Replan: easy — the decision is harder (less info / tighter clock / - committed choice).' - win_condition: - all_of: - - has_building: tent - - building_total_gte: 4 - - after_ticks: 1200 - - within_ticks: 22000 - max_turns: 45 - medium: - description: 'Error Recovery / Replan: medium — the decision is harder (less info / tighter clock - / committed choice).' - win_condition: - all_of: - - has_building: tent - - building_total_gte: 5 - - after_ticks: 1200 - - within_ticks: 17000 - max_turns: 53 - hard: - description: 'Error Recovery / Replan: hard — the decision is harder (less info / tighter clock / - committed choice).' - win_condition: - all_of: - - has_building: tent - - building_total_gte: 6 - - after_ticks: 1200 - - within_ticks: 12000 - max_turns: 61 - fail_condition: - units_lost_lte: -1 -starting_cash: 3000 diff --git a/openra_bench/scenarios/packs/cat-c12-03.yaml b/openra_bench/scenarios/packs/cat-c12-03.yaml deleted file mode 100644 index fc1c94ea80c4e8386ec090f166e2886cdd721c30..0000000000000000000000000000000000000000 --- a/openra_bench/scenarios/packs/cat-c12-03.yaml +++ /dev/null @@ -1,87 +0,0 @@ -meta: - id: cat-c12-03 - title: 'Error Recovery / Replan #4' - capability: reasoning - real_world_meaning: Detecting plan failure and repairing it under reduced means and a deadline — receding-horizon - replanning. - robotics_analogue: After a setback, rebuild/repair and re-establish capability under a hard clock. - author: catalog-generator - status: quarantine - quarantine_reason: 'redundant cat-* over-generation (audit: keep 2 representatives/category)' -base_map: rush-hour-arena -base: - agent: - faction: allies - enemy: - faction: soviet - tools: - - build - - place_building - - move_units - - deploy - - stop - planning: true - termination: - max_ticks: 40000 - actors: - - type: fact - owner: agent - position: - - 10 - - 18 - - type: powr - owner: agent - position: - - 14 - - 18 - - type: tent - owner: agent - position: - - 12 - - 21 - - type: jeep - owner: agent - position: - - 8 - - 16 - - type: e1 - owner: enemy - position: - - 114 - - 34 - count: 1 - stance: 2 -levels: - easy: - description: 'Error Recovery / Replan: easy — the decision is harder (less info / tighter clock / - committed choice).' - win_condition: - all_of: - - has_building: tent - - building_total_gte: 4 - - after_ticks: 1200 - - within_ticks: 22000 - max_turns: 45 - medium: - description: 'Error Recovery / Replan: medium — the decision is harder (less info / tighter clock - / committed choice).' - win_condition: - all_of: - - has_building: tent - - building_total_gte: 5 - - after_ticks: 1200 - - within_ticks: 17000 - max_turns: 53 - hard: - description: 'Error Recovery / Replan: hard — the decision is harder (less info / tighter clock / - committed choice).' - win_condition: - all_of: - - has_building: tent - - building_total_gte: 6 - - after_ticks: 1200 - - within_ticks: 12000 - max_turns: 61 - fail_condition: - units_lost_lte: -1 -starting_cash: 3000 diff --git a/openra_bench/scenarios/packs/cat-c12-04.yaml b/openra_bench/scenarios/packs/cat-c12-04.yaml deleted file mode 100644 index c5cea10498e92d43058aa8ce5e57ea7d843a1c2b..0000000000000000000000000000000000000000 --- a/openra_bench/scenarios/packs/cat-c12-04.yaml +++ /dev/null @@ -1,87 +0,0 @@ -meta: - id: cat-c12-04 - title: 'Error Recovery / Replan #5' - capability: reasoning - real_world_meaning: Detecting plan failure and repairing it under reduced means and a deadline — receding-horizon - replanning. - robotics_analogue: After a setback, rebuild/repair and re-establish capability under a hard clock. - author: catalog-generator - status: quarantine - quarantine_reason: 'redundant cat-* over-generation (audit: keep 2 representatives/category)' -base_map: rush-hour-arena -base: - agent: - faction: allies - enemy: - faction: soviet - tools: - - build - - place_building - - move_units - - deploy - - stop - planning: true - termination: - max_ticks: 40000 - actors: - - type: fact - owner: agent - position: - - 10 - - 18 - - type: powr - owner: agent - position: - - 14 - - 18 - - type: tent - owner: agent - position: - - 12 - - 21 - - type: jeep - owner: agent - position: - - 8 - - 16 - - type: e1 - owner: enemy - position: - - 114 - - 34 - count: 1 - stance: 2 -levels: - easy: - description: 'Error Recovery / Replan: easy — the decision is harder (less info / tighter clock / - committed choice).' - win_condition: - all_of: - - has_building: tent - - building_total_gte: 4 - - after_ticks: 1200 - - within_ticks: 22000 - max_turns: 45 - medium: - description: 'Error Recovery / Replan: medium — the decision is harder (less info / tighter clock - / committed choice).' - win_condition: - all_of: - - has_building: tent - - building_total_gte: 5 - - after_ticks: 1200 - - within_ticks: 17000 - max_turns: 53 - hard: - description: 'Error Recovery / Replan: hard — the decision is harder (less info / tighter clock / - committed choice).' - win_condition: - all_of: - - has_building: tent - - building_total_gte: 6 - - after_ticks: 1200 - - within_ticks: 12000 - max_turns: 61 - fail_condition: - units_lost_lte: -1 -starting_cash: 3000 diff --git a/openra_bench/scenarios/packs/cat-c2-00.yaml b/openra_bench/scenarios/packs/cat-c2-00.yaml deleted file mode 100644 index c91e2f41fda82212dab29b2f8f9f3bf0878c534f..0000000000000000000000000000000000000000 --- a/openra_bench/scenarios/packs/cat-c2-00.yaml +++ /dev/null @@ -1,94 +0,0 @@ -meta: - id: cat-c2-00 - title: 'Threat Enumeration #1' - capability: perception - real_world_meaning: "Estimating hidden adversary state from partial sightings before\ - \ acting \u2014 perception under occlusion and distractors." - robotics_analogue: Scene state-estimation under fog and decoys before committing - force. - author: catalog-generator -base_map: rush-hour-arena -base: - agent: - faction: allies - enemy: - faction: soviet - tools: - - move_units - - attack_unit - - attack_move - - stop - planning: true - termination: - max_ticks: 30000 - actors: - - type: 2tnk - owner: agent - position: - - 8 - - 18 - count: 2 - - type: e1 - owner: agent - position: - - 9 - - 20 - count: 3 - - type: e1 - owner: enemy - position: - - 100 - - 20 - count: 2 - stance: 2 - - type: e3 - owner: enemy - position: - - 104 - - 24 - count: 1 - stance: 2 - - type: e1 - owner: enemy - position: - - 40 - - 10 - count: 1 - stance: 2 - - type: e1 - owner: enemy - position: - - 48 - - 10 - count: 1 - stance: 2 -levels: - easy: - description: "Threat Enumeration: easy \u2014 the decision is harder (less info\ - \ / tighter clock / committed choice)." - win_condition: - all_of: - - enemies_discovered_gte: 2 - - units_lost_lte: 1 - - within_ticks: 14000 - max_turns: 45 - medium: - description: "Threat Enumeration: medium \u2014 the decision is harder (less info\ - \ / tighter clock / committed choice)." - win_condition: - all_of: - - enemies_discovered_gte: 3 - - units_lost_lte: 1 - - within_ticks: 11000 - max_turns: 53 - hard: - description: "Threat Enumeration: hard \u2014 the decision is harder (less info\ - \ / tighter clock / committed choice)." - win_condition: - all_of: - - enemies_discovered_gte: 4 - - units_lost_lte: 0 - - within_ticks: 8000 - max_turns: 61 - fail_condition: - units_lost_lte: -1 diff --git a/openra_bench/scenarios/packs/cat-c2-01.yaml b/openra_bench/scenarios/packs/cat-c2-01.yaml deleted file mode 100644 index c37b663fb66275dea2e793c34b4d9aaf172c6321..0000000000000000000000000000000000000000 --- a/openra_bench/scenarios/packs/cat-c2-01.yaml +++ /dev/null @@ -1,94 +0,0 @@ -meta: - id: cat-c2-01 - title: 'Threat Enumeration #2' - capability: perception - real_world_meaning: "Estimating hidden adversary state from partial sightings before\ - \ acting \u2014 perception under occlusion and distractors." - robotics_analogue: Scene state-estimation under fog and decoys before committing - force. - author: catalog-generator -base_map: rush-hour-arena -base: - agent: - faction: allies - enemy: - faction: soviet - tools: - - move_units - - attack_unit - - attack_move - - stop - planning: true - termination: - max_ticks: 30000 - actors: - - type: 2tnk - owner: agent - position: - - 8 - - 18 - count: 2 - - type: e1 - owner: agent - position: - - 9 - - 20 - count: 3 - - type: e1 - owner: enemy - position: - - 100 - - 20 - count: 2 - stance: 2 - - type: e3 - owner: enemy - position: - - 104 - - 24 - count: 1 - stance: 2 - - type: e1 - owner: enemy - position: - - 40 - - 10 - count: 1 - stance: 2 - - type: e1 - owner: enemy - position: - - 48 - - 10 - count: 1 - stance: 2 -levels: - easy: - description: "Threat Enumeration: easy \u2014 the decision is harder (less info\ - \ / tighter clock / committed choice)." - win_condition: - all_of: - - enemies_discovered_gte: 2 - - units_lost_lte: 1 - - within_ticks: 13850 - max_turns: 45 - medium: - description: "Threat Enumeration: medium \u2014 the decision is harder (less info\ - \ / tighter clock / committed choice)." - win_condition: - all_of: - - enemies_discovered_gte: 3 - - units_lost_lte: 1 - - within_ticks: 10850 - max_turns: 53 - hard: - description: "Threat Enumeration: hard \u2014 the decision is harder (less info\ - \ / tighter clock / committed choice)." - win_condition: - all_of: - - enemies_discovered_gte: 4 - - units_lost_lte: 0 - - within_ticks: 7850 - max_turns: 61 - fail_condition: - units_lost_lte: -1 diff --git a/openra_bench/scenarios/packs/cat-c2-02.yaml b/openra_bench/scenarios/packs/cat-c2-02.yaml deleted file mode 100644 index 7c0dbf4e0714693bddef3843626ab736e00c449a..0000000000000000000000000000000000000000 --- a/openra_bench/scenarios/packs/cat-c2-02.yaml +++ /dev/null @@ -1,95 +0,0 @@ -meta: - id: cat-c2-02 - title: 'Threat Enumeration #3' - capability: perception - real_world_meaning: Estimating hidden adversary state from partial sightings before acting — perception - under occlusion and distractors. - robotics_analogue: Scene state-estimation under fog and decoys before committing force. - author: catalog-generator - status: quarantine - quarantine_reason: 'redundant cat-* over-generation (audit: keep 2 representatives/category)' -base_map: rush-hour-arena -base: - agent: - faction: allies - enemy: - faction: soviet - tools: - - move_units - - attack_unit - - attack_move - - stop - planning: true - termination: - max_ticks: 30000 - actors: - - type: 2tnk - owner: agent - position: - - 8 - - 18 - count: 2 - - type: e1 - owner: agent - position: - - 9 - - 20 - count: 3 - - type: e1 - owner: enemy - position: - - 100 - - 20 - count: 2 - stance: 2 - - type: e3 - owner: enemy - position: - - 104 - - 24 - count: 1 - stance: 2 - - type: e1 - owner: enemy - position: - - 40 - - 10 - count: 1 - stance: 2 - - type: e1 - owner: enemy - position: - - 48 - - 10 - count: 1 - stance: 2 -levels: - easy: - description: 'Threat Enumeration: easy — the decision is harder (less info / tighter clock / committed - choice).' - win_condition: - all_of: - - enemies_discovered_gte: 2 - - units_lost_lte: 1 - - within_ticks: 13700 - max_turns: 45 - medium: - description: 'Threat Enumeration: medium — the decision is harder (less info / tighter clock / committed - choice).' - win_condition: - all_of: - - enemies_discovered_gte: 3 - - units_lost_lte: 1 - - within_ticks: 10700 - max_turns: 53 - hard: - description: 'Threat Enumeration: hard — the decision is harder (less info / tighter clock / committed - choice).' - win_condition: - all_of: - - enemies_discovered_gte: 4 - - units_lost_lte: 0 - - within_ticks: 7700 - max_turns: 61 - fail_condition: - units_lost_lte: -1 diff --git a/openra_bench/scenarios/packs/cat-c2-03.yaml b/openra_bench/scenarios/packs/cat-c2-03.yaml deleted file mode 100644 index 3943773de2f5abc5533851d43378e9173dba7b5f..0000000000000000000000000000000000000000 --- a/openra_bench/scenarios/packs/cat-c2-03.yaml +++ /dev/null @@ -1,95 +0,0 @@ -meta: - id: cat-c2-03 - title: 'Threat Enumeration #4' - capability: perception - real_world_meaning: Estimating hidden adversary state from partial sightings before acting — perception - under occlusion and distractors. - robotics_analogue: Scene state-estimation under fog and decoys before committing force. - author: catalog-generator - status: quarantine - quarantine_reason: 'redundant cat-* over-generation (audit: keep 2 representatives/category)' -base_map: rush-hour-arena -base: - agent: - faction: allies - enemy: - faction: soviet - tools: - - move_units - - attack_unit - - attack_move - - stop - planning: true - termination: - max_ticks: 30000 - actors: - - type: 2tnk - owner: agent - position: - - 8 - - 18 - count: 2 - - type: e1 - owner: agent - position: - - 9 - - 20 - count: 3 - - type: e1 - owner: enemy - position: - - 100 - - 20 - count: 2 - stance: 2 - - type: e3 - owner: enemy - position: - - 104 - - 24 - count: 1 - stance: 2 - - type: e1 - owner: enemy - position: - - 40 - - 10 - count: 1 - stance: 2 - - type: e1 - owner: enemy - position: - - 48 - - 10 - count: 1 - stance: 2 -levels: - easy: - description: 'Threat Enumeration: easy — the decision is harder (less info / tighter clock / committed - choice).' - win_condition: - all_of: - - enemies_discovered_gte: 2 - - units_lost_lte: 1 - - within_ticks: 13550 - max_turns: 45 - medium: - description: 'Threat Enumeration: medium — the decision is harder (less info / tighter clock / committed - choice).' - win_condition: - all_of: - - enemies_discovered_gte: 3 - - units_lost_lte: 1 - - within_ticks: 10550 - max_turns: 53 - hard: - description: 'Threat Enumeration: hard — the decision is harder (less info / tighter clock / committed - choice).' - win_condition: - all_of: - - enemies_discovered_gte: 4 - - units_lost_lte: 0 - - within_ticks: 7550 - max_turns: 61 - fail_condition: - units_lost_lte: -1 diff --git a/openra_bench/scenarios/packs/cat-c2-04.yaml b/openra_bench/scenarios/packs/cat-c2-04.yaml deleted file mode 100644 index 7adf3fabac6242d64e3de03a8a45c8c822042572..0000000000000000000000000000000000000000 --- a/openra_bench/scenarios/packs/cat-c2-04.yaml +++ /dev/null @@ -1,95 +0,0 @@ -meta: - id: cat-c2-04 - title: 'Threat Enumeration #5' - capability: perception - real_world_meaning: Estimating hidden adversary state from partial sightings before acting — perception - under occlusion and distractors. - robotics_analogue: Scene state-estimation under fog and decoys before committing force. - author: catalog-generator - status: quarantine - quarantine_reason: 'redundant cat-* over-generation (audit: keep 2 representatives/category)' -base_map: rush-hour-arena -base: - agent: - faction: allies - enemy: - faction: soviet - tools: - - move_units - - attack_unit - - attack_move - - stop - planning: true - termination: - max_ticks: 30000 - actors: - - type: 2tnk - owner: agent - position: - - 8 - - 18 - count: 2 - - type: e1 - owner: agent - position: - - 9 - - 20 - count: 3 - - type: e1 - owner: enemy - position: - - 100 - - 20 - count: 2 - stance: 2 - - type: e3 - owner: enemy - position: - - 104 - - 24 - count: 1 - stance: 2 - - type: e1 - owner: enemy - position: - - 40 - - 10 - count: 1 - stance: 2 - - type: e1 - owner: enemy - position: - - 48 - - 10 - count: 1 - stance: 2 -levels: - easy: - description: 'Threat Enumeration: easy — the decision is harder (less info / tighter clock / committed - choice).' - win_condition: - all_of: - - enemies_discovered_gte: 2 - - units_lost_lte: 1 - - within_ticks: 13400 - max_turns: 45 - medium: - description: 'Threat Enumeration: medium — the decision is harder (less info / tighter clock / committed - choice).' - win_condition: - all_of: - - enemies_discovered_gte: 3 - - units_lost_lte: 1 - - within_ticks: 10400 - max_turns: 53 - hard: - description: 'Threat Enumeration: hard — the decision is harder (less info / tighter clock / committed - choice).' - win_condition: - all_of: - - enemies_discovered_gte: 4 - - units_lost_lte: 0 - - within_ticks: 7400 - max_turns: 61 - fail_condition: - units_lost_lte: -1 diff --git a/openra_bench/scenarios/packs/cat-c2-05.yaml b/openra_bench/scenarios/packs/cat-c2-05.yaml deleted file mode 100644 index 08b6894521b750d255c76384d19ea529e7cd681d..0000000000000000000000000000000000000000 --- a/openra_bench/scenarios/packs/cat-c2-05.yaml +++ /dev/null @@ -1,95 +0,0 @@ -meta: - id: cat-c2-05 - title: 'Threat Enumeration #6' - capability: perception - real_world_meaning: Estimating hidden adversary state from partial sightings before acting — perception - under occlusion and distractors. - robotics_analogue: Scene state-estimation under fog and decoys before committing force. - author: catalog-generator - status: quarantine - quarantine_reason: 'redundant cat-* over-generation (audit: keep 2 representatives/category)' -base_map: rush-hour-arena -base: - agent: - faction: allies - enemy: - faction: soviet - tools: - - move_units - - attack_unit - - attack_move - - stop - planning: true - termination: - max_ticks: 30000 - actors: - - type: 2tnk - owner: agent - position: - - 8 - - 18 - count: 2 - - type: e1 - owner: agent - position: - - 9 - - 20 - count: 3 - - type: e1 - owner: enemy - position: - - 100 - - 20 - count: 2 - stance: 2 - - type: e3 - owner: enemy - position: - - 104 - - 24 - count: 1 - stance: 2 - - type: e1 - owner: enemy - position: - - 40 - - 10 - count: 1 - stance: 2 - - type: e1 - owner: enemy - position: - - 48 - - 10 - count: 1 - stance: 2 -levels: - easy: - description: 'Threat Enumeration: easy — the decision is harder (less info / tighter clock / committed - choice).' - win_condition: - all_of: - - enemies_discovered_gte: 2 - - units_lost_lte: 1 - - within_ticks: 13250 - max_turns: 45 - medium: - description: 'Threat Enumeration: medium — the decision is harder (less info / tighter clock / committed - choice).' - win_condition: - all_of: - - enemies_discovered_gte: 3 - - units_lost_lte: 1 - - within_ticks: 10250 - max_turns: 53 - hard: - description: 'Threat Enumeration: hard — the decision is harder (less info / tighter clock / committed - choice).' - win_condition: - all_of: - - enemies_discovered_gte: 4 - - units_lost_lte: 0 - - within_ticks: 7250 - max_turns: 61 - fail_condition: - units_lost_lte: -1 diff --git a/openra_bench/scenarios/packs/cat-c3-00.yaml b/openra_bench/scenarios/packs/cat-c3-00.yaml deleted file mode 100644 index a9d1ffea51cd171f392183535019ef2df8e71b3c..0000000000000000000000000000000000000000 --- a/openra_bench/scenarios/packs/cat-c3-00.yaml +++ /dev/null @@ -1,80 +0,0 @@ -meta: - id: cat-c3-00 - title: 'Tech Critical Path #1' - capability: reasoning - real_world_meaning: "Precondition-ordered construction (B needs A) to hit a capability\ - \ deadline \u2014 topological planning under a cost+time budget." - robotics_analogue: Task-graph scheduling with prerequisites under a makespan. - author: catalog-generator -base_map: rush-hour-arena -base: - agent: - faction: allies - enemy: - faction: soviet - tools: - - build - - place_building - - move_units - - deploy - - stop - planning: true - termination: - max_ticks: 40000 - actors: - - type: fact - owner: agent - position: - - 10 - - 18 - - type: powr - owner: agent - position: - - 14 - - 18 - - type: jeep - owner: agent - position: - - 8 - - 16 - - type: e1 - owner: enemy - position: - - 114 - - 34 - count: 1 - stance: 2 -levels: - easy: - description: "Tech Critical Path: easy \u2014 the decision is harder (less info\ - \ / tighter clock / committed choice)." - win_condition: - all_of: - - has_building: tent - - building_total_gte: 4 - - power_surplus_gte: 0 - - within_ticks: 22000 - max_turns: 45 - medium: - description: "Tech Critical Path: medium \u2014 the decision is harder (less info\ - \ / tighter clock / committed choice)." - win_condition: - all_of: - - has_building: tent - - building_total_gte: 5 - - power_surplus_gte: 0 - - within_ticks: 17000 - max_turns: 53 - hard: - description: "Tech Critical Path: hard \u2014 the decision is harder (less info\ - \ / tighter clock / committed choice)." - win_condition: - all_of: - - has_building: tent - - building_total_gte: 6 - - power_surplus_gte: 0 - - within_ticks: 12000 - max_turns: 61 - fail_condition: - units_lost_lte: -1 -starting_cash: 6000 diff --git a/openra_bench/scenarios/packs/cat-c3-01.yaml b/openra_bench/scenarios/packs/cat-c3-01.yaml deleted file mode 100644 index 5911f27467847db905546f7496b269b369199b49..0000000000000000000000000000000000000000 --- a/openra_bench/scenarios/packs/cat-c3-01.yaml +++ /dev/null @@ -1,80 +0,0 @@ -meta: - id: cat-c3-01 - title: 'Tech Critical Path #2' - capability: reasoning - real_world_meaning: "Precondition-ordered construction (B needs A) to hit a capability\ - \ deadline \u2014 topological planning under a cost+time budget." - robotics_analogue: Task-graph scheduling with prerequisites under a makespan. - author: catalog-generator -base_map: rush-hour-arena -base: - agent: - faction: allies - enemy: - faction: soviet - tools: - - build - - place_building - - move_units - - deploy - - stop - planning: true - termination: - max_ticks: 40000 - actors: - - type: fact - owner: agent - position: - - 10 - - 18 - - type: powr - owner: agent - position: - - 14 - - 18 - - type: jeep - owner: agent - position: - - 8 - - 16 - - type: e1 - owner: enemy - position: - - 114 - - 34 - count: 1 - stance: 2 -levels: - easy: - description: "Tech Critical Path: easy \u2014 the decision is harder (less info\ - \ / tighter clock / committed choice)." - win_condition: - all_of: - - has_building: tent - - building_total_gte: 4 - - power_surplus_gte: 0 - - within_ticks: 21700 - max_turns: 45 - medium: - description: "Tech Critical Path: medium \u2014 the decision is harder (less info\ - \ / tighter clock / committed choice)." - win_condition: - all_of: - - has_building: tent - - building_total_gte: 5 - - power_surplus_gte: 0 - - within_ticks: 16700 - max_turns: 53 - hard: - description: "Tech Critical Path: hard \u2014 the decision is harder (less info\ - \ / tighter clock / committed choice)." - win_condition: - all_of: - - has_building: tent - - building_total_gte: 6 - - power_surplus_gte: 0 - - within_ticks: 11700 - max_turns: 61 - fail_condition: - units_lost_lte: -1 -starting_cash: 6000 diff --git a/openra_bench/scenarios/packs/cat-c3-02.yaml b/openra_bench/scenarios/packs/cat-c3-02.yaml deleted file mode 100644 index 44afdc165f8efa025638a83876fc0d40cc3fe6db..0000000000000000000000000000000000000000 --- a/openra_bench/scenarios/packs/cat-c3-02.yaml +++ /dev/null @@ -1,82 +0,0 @@ -meta: - id: cat-c3-02 - title: 'Tech Critical Path #3' - capability: reasoning - real_world_meaning: Precondition-ordered construction (B needs A) to hit a capability deadline — topological - planning under a cost+time budget. - robotics_analogue: Task-graph scheduling with prerequisites under a makespan. - author: catalog-generator - status: quarantine - quarantine_reason: 'redundant cat-* over-generation (audit: keep 2 representatives/category)' -base_map: rush-hour-arena -base: - agent: - faction: allies - enemy: - faction: soviet - tools: - - build - - place_building - - move_units - - deploy - - stop - planning: true - termination: - max_ticks: 40000 - actors: - - type: fact - owner: agent - position: - - 10 - - 18 - - type: powr - owner: agent - position: - - 14 - - 18 - - type: jeep - owner: agent - position: - - 8 - - 16 - - type: e1 - owner: enemy - position: - - 114 - - 34 - count: 1 - stance: 2 -levels: - easy: - description: 'Tech Critical Path: easy — the decision is harder (less info / tighter clock / committed - choice).' - win_condition: - all_of: - - has_building: tent - - building_total_gte: 4 - - power_surplus_gte: 0 - - within_ticks: 21400 - max_turns: 45 - medium: - description: 'Tech Critical Path: medium — the decision is harder (less info / tighter clock / committed - choice).' - win_condition: - all_of: - - has_building: tent - - building_total_gte: 5 - - power_surplus_gte: 0 - - within_ticks: 16400 - max_turns: 53 - hard: - description: 'Tech Critical Path: hard — the decision is harder (less info / tighter clock / committed - choice).' - win_condition: - all_of: - - has_building: tent - - building_total_gte: 6 - - power_surplus_gte: 0 - - within_ticks: 11400 - max_turns: 61 - fail_condition: - units_lost_lte: -1 -starting_cash: 6000 diff --git a/openra_bench/scenarios/packs/cat-c3-03.yaml b/openra_bench/scenarios/packs/cat-c3-03.yaml deleted file mode 100644 index 170dd9e798ad510847f9bda76ce90866d34a8e62..0000000000000000000000000000000000000000 --- a/openra_bench/scenarios/packs/cat-c3-03.yaml +++ /dev/null @@ -1,82 +0,0 @@ -meta: - id: cat-c3-03 - title: 'Tech Critical Path #4' - capability: reasoning - real_world_meaning: Precondition-ordered construction (B needs A) to hit a capability deadline — topological - planning under a cost+time budget. - robotics_analogue: Task-graph scheduling with prerequisites under a makespan. - author: catalog-generator - status: quarantine - quarantine_reason: 'redundant cat-* over-generation (audit: keep 2 representatives/category)' -base_map: rush-hour-arena -base: - agent: - faction: allies - enemy: - faction: soviet - tools: - - build - - place_building - - move_units - - deploy - - stop - planning: true - termination: - max_ticks: 40000 - actors: - - type: fact - owner: agent - position: - - 10 - - 18 - - type: powr - owner: agent - position: - - 14 - - 18 - - type: jeep - owner: agent - position: - - 8 - - 16 - - type: e1 - owner: enemy - position: - - 114 - - 34 - count: 1 - stance: 2 -levels: - easy: - description: 'Tech Critical Path: easy — the decision is harder (less info / tighter clock / committed - choice).' - win_condition: - all_of: - - has_building: tent - - building_total_gte: 4 - - power_surplus_gte: 0 - - within_ticks: 21100 - max_turns: 45 - medium: - description: 'Tech Critical Path: medium — the decision is harder (less info / tighter clock / committed - choice).' - win_condition: - all_of: - - has_building: tent - - building_total_gte: 5 - - power_surplus_gte: 0 - - within_ticks: 16100 - max_turns: 53 - hard: - description: 'Tech Critical Path: hard — the decision is harder (less info / tighter clock / committed - choice).' - win_condition: - all_of: - - has_building: tent - - building_total_gte: 6 - - power_surplus_gte: 0 - - within_ticks: 11100 - max_turns: 61 - fail_condition: - units_lost_lte: -1 -starting_cash: 6000 diff --git a/openra_bench/scenarios/packs/cat-c3-04.yaml b/openra_bench/scenarios/packs/cat-c3-04.yaml deleted file mode 100644 index 332f45d12e387a88c08aa02fec79689b84582fca..0000000000000000000000000000000000000000 --- a/openra_bench/scenarios/packs/cat-c3-04.yaml +++ /dev/null @@ -1,82 +0,0 @@ -meta: - id: cat-c3-04 - title: 'Tech Critical Path #5' - capability: reasoning - real_world_meaning: Precondition-ordered construction (B needs A) to hit a capability deadline — topological - planning under a cost+time budget. - robotics_analogue: Task-graph scheduling with prerequisites under a makespan. - author: catalog-generator - status: quarantine - quarantine_reason: 'redundant cat-* over-generation (audit: keep 2 representatives/category)' -base_map: rush-hour-arena -base: - agent: - faction: allies - enemy: - faction: soviet - tools: - - build - - place_building - - move_units - - deploy - - stop - planning: true - termination: - max_ticks: 40000 - actors: - - type: fact - owner: agent - position: - - 10 - - 18 - - type: powr - owner: agent - position: - - 14 - - 18 - - type: jeep - owner: agent - position: - - 8 - - 16 - - type: e1 - owner: enemy - position: - - 114 - - 34 - count: 1 - stance: 2 -levels: - easy: - description: 'Tech Critical Path: easy — the decision is harder (less info / tighter clock / committed - choice).' - win_condition: - all_of: - - has_building: tent - - building_total_gte: 4 - - power_surplus_gte: 0 - - within_ticks: 20800 - max_turns: 45 - medium: - description: 'Tech Critical Path: medium — the decision is harder (less info / tighter clock / committed - choice).' - win_condition: - all_of: - - has_building: tent - - building_total_gte: 5 - - power_surplus_gte: 0 - - within_ticks: 15800 - max_turns: 53 - hard: - description: 'Tech Critical Path: hard — the decision is harder (less info / tighter clock / committed - choice).' - win_condition: - all_of: - - has_building: tent - - building_total_gte: 6 - - power_surplus_gte: 0 - - within_ticks: 10800 - max_turns: 61 - fail_condition: - units_lost_lte: -1 -starting_cash: 6000 diff --git a/openra_bench/scenarios/packs/cat-c3-05.yaml b/openra_bench/scenarios/packs/cat-c3-05.yaml deleted file mode 100644 index f26ffcb55a02dc26e769fef494f49c639a0ff27e..0000000000000000000000000000000000000000 --- a/openra_bench/scenarios/packs/cat-c3-05.yaml +++ /dev/null @@ -1,82 +0,0 @@ -meta: - id: cat-c3-05 - title: 'Tech Critical Path #6' - capability: reasoning - real_world_meaning: Precondition-ordered construction (B needs A) to hit a capability deadline — topological - planning under a cost+time budget. - robotics_analogue: Task-graph scheduling with prerequisites under a makespan. - author: catalog-generator - status: quarantine - quarantine_reason: 'redundant cat-* over-generation (audit: keep 2 representatives/category)' -base_map: rush-hour-arena -base: - agent: - faction: allies - enemy: - faction: soviet - tools: - - build - - place_building - - move_units - - deploy - - stop - planning: true - termination: - max_ticks: 40000 - actors: - - type: fact - owner: agent - position: - - 10 - - 18 - - type: powr - owner: agent - position: - - 14 - - 18 - - type: jeep - owner: agent - position: - - 8 - - 16 - - type: e1 - owner: enemy - position: - - 114 - - 34 - count: 1 - stance: 2 -levels: - easy: - description: 'Tech Critical Path: easy — the decision is harder (less info / tighter clock / committed - choice).' - win_condition: - all_of: - - has_building: tent - - building_total_gte: 4 - - power_surplus_gte: 0 - - within_ticks: 20500 - max_turns: 45 - medium: - description: 'Tech Critical Path: medium — the decision is harder (less info / tighter clock / committed - choice).' - win_condition: - all_of: - - has_building: tent - - building_total_gte: 5 - - power_surplus_gte: 0 - - within_ticks: 15500 - max_turns: 53 - hard: - description: 'Tech Critical Path: hard — the decision is harder (less info / tighter clock / committed - choice).' - win_condition: - all_of: - - has_building: tent - - building_total_gte: 6 - - power_surplus_gte: 0 - - within_ticks: 10500 - max_turns: 61 - fail_condition: - units_lost_lte: -1 -starting_cash: 6000 diff --git a/openra_bench/scenarios/packs/cat-c4-00.yaml b/openra_bench/scenarios/packs/cat-c4-00.yaml deleted file mode 100644 index 78cd7a0ebaa6b08e79676aa1a846bf08a67bfd69..0000000000000000000000000000000000000000 --- a/openra_bench/scenarios/packs/cat-c4-00.yaml +++ /dev/null @@ -1,77 +0,0 @@ -meta: - id: cat-c4-00 - title: 'Power-Budget Online #1' - capability: reasoning - real_world_meaning: "Sequencing subsystem bring-up so a shared online resource never\ - \ goes negative \u2014 constraint-respecting schedule." - robotics_analogue: Keeping an online resource (power) non-negative while expanding. - author: catalog-generator -base_map: rush-hour-arena -base: - agent: - faction: allies - enemy: - faction: soviet - tools: - - build - - place_building - - move_units - - deploy - - stop - planning: true - termination: - max_ticks: 40000 - actors: - - type: fact - owner: agent - position: - - 10 - - 18 - - type: powr - owner: agent - position: - - 14 - - 18 - - type: jeep - owner: agent - position: - - 8 - - 16 - - type: e1 - owner: enemy - position: - - 114 - - 34 - count: 1 - stance: 2 -levels: - easy: - description: "Power-Budget Online: easy \u2014 the decision is harder (less info\ - \ / tighter clock / committed choice)." - win_condition: - all_of: - - power_surplus_gte: 0 - - building_total_gte: 4 - - within_ticks: 20000 - max_turns: 45 - medium: - description: "Power-Budget Online: medium \u2014 the decision is harder (less\ - \ info / tighter clock / committed choice)." - win_condition: - all_of: - - power_surplus_gte: 0 - - building_total_gte: 5 - - within_ticks: 15500 - max_turns: 53 - hard: - description: "Power-Budget Online: hard \u2014 the decision is harder (less info\ - \ / tighter clock / committed choice)." - win_condition: - all_of: - - power_surplus_gte: 0 - - building_total_gte: 6 - - within_ticks: 11000 - max_turns: 61 - fail_condition: - units_lost_lte: -1 -starting_cash: 5000 diff --git a/openra_bench/scenarios/packs/cat-c4-01.yaml b/openra_bench/scenarios/packs/cat-c4-01.yaml deleted file mode 100644 index ef358ec7c033ac5ca8f95e896b09f3d7ce07cb15..0000000000000000000000000000000000000000 --- a/openra_bench/scenarios/packs/cat-c4-01.yaml +++ /dev/null @@ -1,77 +0,0 @@ -meta: - id: cat-c4-01 - title: 'Power-Budget Online #2' - capability: reasoning - real_world_meaning: "Sequencing subsystem bring-up so a shared online resource never\ - \ goes negative \u2014 constraint-respecting schedule." - robotics_analogue: Keeping an online resource (power) non-negative while expanding. - author: catalog-generator -base_map: rush-hour-arena -base: - agent: - faction: allies - enemy: - faction: soviet - tools: - - build - - place_building - - move_units - - deploy - - stop - planning: true - termination: - max_ticks: 40000 - actors: - - type: fact - owner: agent - position: - - 10 - - 18 - - type: powr - owner: agent - position: - - 14 - - 18 - - type: jeep - owner: agent - position: - - 8 - - 16 - - type: e1 - owner: enemy - position: - - 114 - - 34 - count: 1 - stance: 2 -levels: - easy: - description: "Power-Budget Online: easy \u2014 the decision is harder (less info\ - \ / tighter clock / committed choice)." - win_condition: - all_of: - - power_surplus_gte: 0 - - building_total_gte: 4 - - within_ticks: 19750 - max_turns: 45 - medium: - description: "Power-Budget Online: medium \u2014 the decision is harder (less\ - \ info / tighter clock / committed choice)." - win_condition: - all_of: - - power_surplus_gte: 0 - - building_total_gte: 5 - - within_ticks: 15250 - max_turns: 53 - hard: - description: "Power-Budget Online: hard \u2014 the decision is harder (less info\ - \ / tighter clock / committed choice)." - win_condition: - all_of: - - power_surplus_gte: 0 - - building_total_gte: 6 - - within_ticks: 10750 - max_turns: 61 - fail_condition: - units_lost_lte: -1 -starting_cash: 5000 diff --git a/openra_bench/scenarios/packs/cat-c4-02.yaml b/openra_bench/scenarios/packs/cat-c4-02.yaml deleted file mode 100644 index ad0f3c2e301b462df2a187caf2bf1500332d75dc..0000000000000000000000000000000000000000 --- a/openra_bench/scenarios/packs/cat-c4-02.yaml +++ /dev/null @@ -1,79 +0,0 @@ -meta: - id: cat-c4-02 - title: 'Power-Budget Online #3' - capability: reasoning - real_world_meaning: Sequencing subsystem bring-up so a shared online resource never goes negative — - constraint-respecting schedule. - robotics_analogue: Keeping an online resource (power) non-negative while expanding. - author: catalog-generator - status: quarantine - quarantine_reason: 'redundant cat-* over-generation (audit: keep 2 representatives/category)' -base_map: rush-hour-arena -base: - agent: - faction: allies - enemy: - faction: soviet - tools: - - build - - place_building - - move_units - - deploy - - stop - planning: true - termination: - max_ticks: 40000 - actors: - - type: fact - owner: agent - position: - - 10 - - 18 - - type: powr - owner: agent - position: - - 14 - - 18 - - type: jeep - owner: agent - position: - - 8 - - 16 - - type: e1 - owner: enemy - position: - - 114 - - 34 - count: 1 - stance: 2 -levels: - easy: - description: 'Power-Budget Online: easy — the decision is harder (less info / tighter clock / committed - choice).' - win_condition: - all_of: - - power_surplus_gte: 0 - - building_total_gte: 4 - - within_ticks: 19500 - max_turns: 45 - medium: - description: 'Power-Budget Online: medium — the decision is harder (less info / tighter clock / committed - choice).' - win_condition: - all_of: - - power_surplus_gte: 0 - - building_total_gte: 5 - - within_ticks: 15000 - max_turns: 53 - hard: - description: 'Power-Budget Online: hard — the decision is harder (less info / tighter clock / committed - choice).' - win_condition: - all_of: - - power_surplus_gte: 0 - - building_total_gte: 6 - - within_ticks: 10500 - max_turns: 61 - fail_condition: - units_lost_lte: -1 -starting_cash: 5000 diff --git a/openra_bench/scenarios/packs/cat-c4-03.yaml b/openra_bench/scenarios/packs/cat-c4-03.yaml deleted file mode 100644 index f101f3e83f05820a0942c67d96078571ffebe5d0..0000000000000000000000000000000000000000 --- a/openra_bench/scenarios/packs/cat-c4-03.yaml +++ /dev/null @@ -1,79 +0,0 @@ -meta: - id: cat-c4-03 - title: 'Power-Budget Online #4' - capability: reasoning - real_world_meaning: Sequencing subsystem bring-up so a shared online resource never goes negative — - constraint-respecting schedule. - robotics_analogue: Keeping an online resource (power) non-negative while expanding. - author: catalog-generator - status: quarantine - quarantine_reason: 'redundant cat-* over-generation (audit: keep 2 representatives/category)' -base_map: rush-hour-arena -base: - agent: - faction: allies - enemy: - faction: soviet - tools: - - build - - place_building - - move_units - - deploy - - stop - planning: true - termination: - max_ticks: 40000 - actors: - - type: fact - owner: agent - position: - - 10 - - 18 - - type: powr - owner: agent - position: - - 14 - - 18 - - type: jeep - owner: agent - position: - - 8 - - 16 - - type: e1 - owner: enemy - position: - - 114 - - 34 - count: 1 - stance: 2 -levels: - easy: - description: 'Power-Budget Online: easy — the decision is harder (less info / tighter clock / committed - choice).' - win_condition: - all_of: - - power_surplus_gte: 0 - - building_total_gte: 4 - - within_ticks: 19250 - max_turns: 45 - medium: - description: 'Power-Budget Online: medium — the decision is harder (less info / tighter clock / committed - choice).' - win_condition: - all_of: - - power_surplus_gte: 0 - - building_total_gte: 5 - - within_ticks: 14750 - max_turns: 53 - hard: - description: 'Power-Budget Online: hard — the decision is harder (less info / tighter clock / committed - choice).' - win_condition: - all_of: - - power_surplus_gte: 0 - - building_total_gte: 6 - - within_ticks: 10250 - max_turns: 61 - fail_condition: - units_lost_lte: -1 -starting_cash: 5000 diff --git a/openra_bench/scenarios/packs/cat-c4-04.yaml b/openra_bench/scenarios/packs/cat-c4-04.yaml deleted file mode 100644 index 9c745cd0c06f19ce01c770a62a9bda17a60d69ad..0000000000000000000000000000000000000000 --- a/openra_bench/scenarios/packs/cat-c4-04.yaml +++ /dev/null @@ -1,79 +0,0 @@ -meta: - id: cat-c4-04 - title: 'Power-Budget Online #5' - capability: reasoning - real_world_meaning: Sequencing subsystem bring-up so a shared online resource never goes negative — - constraint-respecting schedule. - robotics_analogue: Keeping an online resource (power) non-negative while expanding. - author: catalog-generator - status: quarantine - quarantine_reason: 'redundant cat-* over-generation (audit: keep 2 representatives/category)' -base_map: rush-hour-arena -base: - agent: - faction: allies - enemy: - faction: soviet - tools: - - build - - place_building - - move_units - - deploy - - stop - planning: true - termination: - max_ticks: 40000 - actors: - - type: fact - owner: agent - position: - - 10 - - 18 - - type: powr - owner: agent - position: - - 14 - - 18 - - type: jeep - owner: agent - position: - - 8 - - 16 - - type: e1 - owner: enemy - position: - - 114 - - 34 - count: 1 - stance: 2 -levels: - easy: - description: 'Power-Budget Online: easy — the decision is harder (less info / tighter clock / committed - choice).' - win_condition: - all_of: - - power_surplus_gte: 0 - - building_total_gte: 4 - - within_ticks: 19000 - max_turns: 45 - medium: - description: 'Power-Budget Online: medium — the decision is harder (less info / tighter clock / committed - choice).' - win_condition: - all_of: - - power_surplus_gte: 0 - - building_total_gte: 5 - - within_ticks: 14500 - max_turns: 53 - hard: - description: 'Power-Budget Online: hard — the decision is harder (less info / tighter clock / committed - choice).' - win_condition: - all_of: - - power_surplus_gte: 0 - - building_total_gte: 6 - - within_ticks: 10000 - max_turns: 61 - fail_condition: - units_lost_lte: -1 -starting_cash: 5000 diff --git a/openra_bench/scenarios/packs/cat-c5-00.yaml b/openra_bench/scenarios/packs/cat-c5-00.yaml deleted file mode 100644 index 6443c4425a65fb5f24f0d40cfb4ee0c1a45e786b..0000000000000000000000000000000000000000 --- a/openra_bench/scenarios/packs/cat-c5-00.yaml +++ /dev/null @@ -1,89 +0,0 @@ -meta: - id: cat-c5-00 - title: 'Budget Allocation (spend) #1' - capability: reasoning - real_world_meaning: "Allocating a finite budget across competing investments where\ - \ a half-measure on both fails \u2014 resource-allocation under a hard cap." - robotics_analogue: 'Indivisible-budget allocation: wide vs deep, splitting clears - nothing.' - author: catalog-generator -base_map: rush-hour-arena -base: - agent: - faction: allies - enemy: - faction: soviet - tools: - - build - - place_building - - move_units - - deploy - - stop - planning: true - termination: - max_ticks: 40000 - actors: - - type: fact - owner: agent - position: - - 10 - - 18 - - type: powr - owner: agent - position: - - 14 - - 18 - - type: tent - owner: agent - position: - - 12 - - 21 - - type: jeep - owner: agent - position: - - 8 - - 16 - - type: e1 - owner: enemy - position: - - 114 - - 34 - count: 1 - stance: 2 -levels: - easy: - description: "Budget Allocation (spend): easy \u2014 the decision is harder (less\ - \ info / tighter clock / committed choice)." - win_condition: - all_of: - - building_count_gte: - type: powr - n: 2 - - building_total_gte: 5 - - within_ticks: 20000 - max_turns: 45 - medium: - description: "Budget Allocation (spend): medium \u2014 the decision is harder\ - \ (less info / tighter clock / committed choice)." - win_condition: - all_of: - - building_count_gte: - type: powr - n: 2 - - building_total_gte: 6 - - within_ticks: 15500 - max_turns: 53 - hard: - description: "Budget Allocation (spend): hard \u2014 the decision is harder (less\ - \ info / tighter clock / committed choice)." - win_condition: - all_of: - - building_count_gte: - type: powr - n: 2 - - building_total_gte: 7 - - within_ticks: 11000 - max_turns: 61 - fail_condition: - units_lost_lte: -1 -starting_cash: 2200 diff --git a/openra_bench/scenarios/packs/cat-c5-01.yaml b/openra_bench/scenarios/packs/cat-c5-01.yaml deleted file mode 100644 index ef5b337bcfd6ec70d41ddabd856d829b9597f5e1..0000000000000000000000000000000000000000 --- a/openra_bench/scenarios/packs/cat-c5-01.yaml +++ /dev/null @@ -1,89 +0,0 @@ -meta: - id: cat-c5-01 - title: 'Budget Allocation (spend) #2' - capability: reasoning - real_world_meaning: "Allocating a finite budget across competing investments where\ - \ a half-measure on both fails \u2014 resource-allocation under a hard cap." - robotics_analogue: 'Indivisible-budget allocation: wide vs deep, splitting clears - nothing.' - author: catalog-generator -base_map: rush-hour-arena -base: - agent: - faction: allies - enemy: - faction: soviet - tools: - - build - - place_building - - move_units - - deploy - - stop - planning: true - termination: - max_ticks: 40000 - actors: - - type: fact - owner: agent - position: - - 10 - - 18 - - type: powr - owner: agent - position: - - 14 - - 18 - - type: tent - owner: agent - position: - - 12 - - 21 - - type: jeep - owner: agent - position: - - 8 - - 16 - - type: e1 - owner: enemy - position: - - 114 - - 34 - count: 1 - stance: 2 -levels: - easy: - description: "Budget Allocation (spend): easy \u2014 the decision is harder (less\ - \ info / tighter clock / committed choice)." - win_condition: - all_of: - - building_count_gte: - type: powr - n: 2 - - building_total_gte: 5 - - within_ticks: 20000 - max_turns: 45 - medium: - description: "Budget Allocation (spend): medium \u2014 the decision is harder\ - \ (less info / tighter clock / committed choice)." - win_condition: - all_of: - - building_count_gte: - type: powr - n: 2 - - building_total_gte: 6 - - within_ticks: 15500 - max_turns: 53 - hard: - description: "Budget Allocation (spend): hard \u2014 the decision is harder (less\ - \ info / tighter clock / committed choice)." - win_condition: - all_of: - - building_count_gte: - type: powr - n: 2 - - building_total_gte: 7 - - within_ticks: 11000 - max_turns: 61 - fail_condition: - units_lost_lte: -1 -starting_cash: 2080 diff --git a/openra_bench/scenarios/packs/cat-c5-02.yaml b/openra_bench/scenarios/packs/cat-c5-02.yaml deleted file mode 100644 index ad2c291b2c6f75edf41616dcb3b3c171bb4a1c4b..0000000000000000000000000000000000000000 --- a/openra_bench/scenarios/packs/cat-c5-02.yaml +++ /dev/null @@ -1,90 +0,0 @@ -meta: - id: cat-c5-02 - title: 'Budget Allocation (spend) #3' - capability: reasoning - real_world_meaning: Allocating a finite budget across competing investments where a half-measure on - both fails — resource-allocation under a hard cap. - robotics_analogue: 'Indivisible-budget allocation: wide vs deep, splitting clears nothing.' - author: catalog-generator - status: quarantine - quarantine_reason: 'redundant cat-* over-generation (audit: keep 2 representatives/category)' -base_map: rush-hour-arena -base: - agent: - faction: allies - enemy: - faction: soviet - tools: - - build - - place_building - - move_units - - deploy - - stop - planning: true - termination: - max_ticks: 40000 - actors: - - type: fact - owner: agent - position: - - 10 - - 18 - - type: powr - owner: agent - position: - - 14 - - 18 - - type: tent - owner: agent - position: - - 12 - - 21 - - type: jeep - owner: agent - position: - - 8 - - 16 - - type: e1 - owner: enemy - position: - - 114 - - 34 - count: 1 - stance: 2 -levels: - easy: - description: 'Budget Allocation (spend): easy — the decision is harder (less info / tighter clock - / committed choice).' - win_condition: - all_of: - - building_count_gte: - type: powr - n: 2 - - building_total_gte: 5 - - within_ticks: 20000 - max_turns: 45 - medium: - description: 'Budget Allocation (spend): medium — the decision is harder (less info / tighter clock - / committed choice).' - win_condition: - all_of: - - building_count_gte: - type: powr - n: 2 - - building_total_gte: 6 - - within_ticks: 15500 - max_turns: 53 - hard: - description: 'Budget Allocation (spend): hard — the decision is harder (less info / tighter clock - / committed choice).' - win_condition: - all_of: - - building_count_gte: - type: powr - n: 2 - - building_total_gte: 7 - - within_ticks: 11000 - max_turns: 61 - fail_condition: - units_lost_lte: -1 -starting_cash: 1960 diff --git a/openra_bench/scenarios/packs/cat-c5-03.yaml b/openra_bench/scenarios/packs/cat-c5-03.yaml deleted file mode 100644 index 30d24bb7f8da47aa641aa0d2547b79c1dc0be74e..0000000000000000000000000000000000000000 --- a/openra_bench/scenarios/packs/cat-c5-03.yaml +++ /dev/null @@ -1,90 +0,0 @@ -meta: - id: cat-c5-03 - title: 'Budget Allocation (spend) #4' - capability: reasoning - real_world_meaning: Allocating a finite budget across competing investments where a half-measure on - both fails — resource-allocation under a hard cap. - robotics_analogue: 'Indivisible-budget allocation: wide vs deep, splitting clears nothing.' - author: catalog-generator - status: quarantine - quarantine_reason: 'redundant cat-* over-generation (audit: keep 2 representatives/category)' -base_map: rush-hour-arena -base: - agent: - faction: allies - enemy: - faction: soviet - tools: - - build - - place_building - - move_units - - deploy - - stop - planning: true - termination: - max_ticks: 40000 - actors: - - type: fact - owner: agent - position: - - 10 - - 18 - - type: powr - owner: agent - position: - - 14 - - 18 - - type: tent - owner: agent - position: - - 12 - - 21 - - type: jeep - owner: agent - position: - - 8 - - 16 - - type: e1 - owner: enemy - position: - - 114 - - 34 - count: 1 - stance: 2 -levels: - easy: - description: 'Budget Allocation (spend): easy — the decision is harder (less info / tighter clock - / committed choice).' - win_condition: - all_of: - - building_count_gte: - type: powr - n: 2 - - building_total_gte: 5 - - within_ticks: 20000 - max_turns: 45 - medium: - description: 'Budget Allocation (spend): medium — the decision is harder (less info / tighter clock - / committed choice).' - win_condition: - all_of: - - building_count_gte: - type: powr - n: 2 - - building_total_gte: 6 - - within_ticks: 15500 - max_turns: 53 - hard: - description: 'Budget Allocation (spend): hard — the decision is harder (less info / tighter clock - / committed choice).' - win_condition: - all_of: - - building_count_gte: - type: powr - n: 2 - - building_total_gte: 7 - - within_ticks: 11000 - max_turns: 61 - fail_condition: - units_lost_lte: -1 -starting_cash: 1840 diff --git a/openra_bench/scenarios/packs/cat-c5-04.yaml b/openra_bench/scenarios/packs/cat-c5-04.yaml deleted file mode 100644 index c6ada1b36aeb724f460720965be3d0ee2a65da6a..0000000000000000000000000000000000000000 --- a/openra_bench/scenarios/packs/cat-c5-04.yaml +++ /dev/null @@ -1,90 +0,0 @@ -meta: - id: cat-c5-04 - title: 'Budget Allocation (spend) #5' - capability: reasoning - real_world_meaning: Allocating a finite budget across competing investments where a half-measure on - both fails — resource-allocation under a hard cap. - robotics_analogue: 'Indivisible-budget allocation: wide vs deep, splitting clears nothing.' - author: catalog-generator - status: quarantine - quarantine_reason: 'redundant cat-* over-generation (audit: keep 2 representatives/category)' -base_map: rush-hour-arena -base: - agent: - faction: allies - enemy: - faction: soviet - tools: - - build - - place_building - - move_units - - deploy - - stop - planning: true - termination: - max_ticks: 40000 - actors: - - type: fact - owner: agent - position: - - 10 - - 18 - - type: powr - owner: agent - position: - - 14 - - 18 - - type: tent - owner: agent - position: - - 12 - - 21 - - type: jeep - owner: agent - position: - - 8 - - 16 - - type: e1 - owner: enemy - position: - - 114 - - 34 - count: 1 - stance: 2 -levels: - easy: - description: 'Budget Allocation (spend): easy — the decision is harder (less info / tighter clock - / committed choice).' - win_condition: - all_of: - - building_count_gte: - type: powr - n: 2 - - building_total_gte: 5 - - within_ticks: 20000 - max_turns: 45 - medium: - description: 'Budget Allocation (spend): medium — the decision is harder (less info / tighter clock - / committed choice).' - win_condition: - all_of: - - building_count_gte: - type: powr - n: 2 - - building_total_gte: 6 - - within_ticks: 15500 - max_turns: 53 - hard: - description: 'Budget Allocation (spend): hard — the decision is harder (less info / tighter clock - / committed choice).' - win_condition: - all_of: - - building_count_gte: - type: powr - n: 2 - - building_total_gte: 7 - - within_ticks: 11000 - max_turns: 61 - fail_condition: - units_lost_lte: -1 -starting_cash: 1720 diff --git a/openra_bench/scenarios/packs/cat-c5-05.yaml b/openra_bench/scenarios/packs/cat-c5-05.yaml deleted file mode 100644 index fb59876462fe8bc74ccc8e17400dde9bd1f86328..0000000000000000000000000000000000000000 --- a/openra_bench/scenarios/packs/cat-c5-05.yaml +++ /dev/null @@ -1,90 +0,0 @@ -meta: - id: cat-c5-05 - title: 'Budget Allocation (spend) #6' - capability: reasoning - real_world_meaning: Allocating a finite budget across competing investments where a half-measure on - both fails — resource-allocation under a hard cap. - robotics_analogue: 'Indivisible-budget allocation: wide vs deep, splitting clears nothing.' - author: catalog-generator - status: quarantine - quarantine_reason: 'redundant cat-* over-generation (audit: keep 2 representatives/category)' -base_map: rush-hour-arena -base: - agent: - faction: allies - enemy: - faction: soviet - tools: - - build - - place_building - - move_units - - deploy - - stop - planning: true - termination: - max_ticks: 40000 - actors: - - type: fact - owner: agent - position: - - 10 - - 18 - - type: powr - owner: agent - position: - - 14 - - 18 - - type: tent - owner: agent - position: - - 12 - - 21 - - type: jeep - owner: agent - position: - - 8 - - 16 - - type: e1 - owner: enemy - position: - - 114 - - 34 - count: 1 - stance: 2 -levels: - easy: - description: 'Budget Allocation (spend): easy — the decision is harder (less info / tighter clock - / committed choice).' - win_condition: - all_of: - - building_count_gte: - type: powr - n: 2 - - building_total_gte: 5 - - within_ticks: 20000 - max_turns: 45 - medium: - description: 'Budget Allocation (spend): medium — the decision is harder (less info / tighter clock - / committed choice).' - win_condition: - all_of: - - building_count_gte: - type: powr - n: 2 - - building_total_gte: 6 - - within_ticks: 15500 - max_turns: 53 - hard: - description: 'Budget Allocation (spend): hard — the decision is harder (less info / tighter clock - / committed choice).' - win_condition: - all_of: - - building_count_gte: - type: powr - n: 2 - - building_total_gte: 7 - - within_ticks: 11000 - max_turns: 61 - fail_condition: - units_lost_lte: -1 -starting_cash: 1600 diff --git a/openra_bench/scenarios/packs/cat-c6-00.yaml b/openra_bench/scenarios/packs/cat-c6-00.yaml deleted file mode 100644 index e9e9483ced607962b1751578bb8e4f171647d9e9..0000000000000000000000000000000000000000 --- a/openra_bench/scenarios/packs/cat-c6-00.yaml +++ /dev/null @@ -1,84 +0,0 @@ -meta: - id: cat-c6-00 - title: 'Time-Boxed Capital Deploy #1' - capability: reasoning - real_world_meaning: "Provisioning under an energy/money budget and a mission deadline\ - \ \u2014 how much to commit and when." - robotics_analogue: Convert a budget into fielded force AND standing infra before - a clock. - author: catalog-generator -base_map: rush-hour-arena -base: - agent: - faction: allies - enemy: - faction: soviet - tools: - - build - - place_building - - move_units - - deploy - - stop - planning: true - termination: - max_ticks: 40000 - actors: - - type: fact - owner: agent - position: - - 10 - - 18 - - type: powr - owner: agent - position: - - 14 - - 18 - - type: tent - owner: agent - position: - - 12 - - 21 - - type: jeep - owner: agent - position: - - 8 - - 16 - - type: e1 - owner: enemy - position: - - 114 - - 34 - count: 1 - stance: 2 -levels: - easy: - description: "Time-Boxed Capital Deploy: easy \u2014 the decision is harder (less\ - \ info / tighter clock / committed choice)." - win_condition: - all_of: - - own_units_gte: 4 - - building_total_gte: 4 - - within_ticks: 16000 - max_turns: 45 - medium: - description: "Time-Boxed Capital Deploy: medium \u2014 the decision is harder\ - \ (less info / tighter clock / committed choice)." - win_condition: - all_of: - - own_units_gte: 5 - - building_total_gte: 5 - - within_ticks: 12500 - max_turns: 53 - hard: - description: "Time-Boxed Capital Deploy: hard \u2014 the decision is harder (less\ - \ info / tighter clock / committed choice)." - win_condition: - all_of: - - own_units_gte: 6 - - building_total_gte: 6 - - units_lost_lte: 0 - - within_ticks: 9000 - max_turns: 61 - fail_condition: - units_lost_lte: -1 -starting_cash: 2000 diff --git a/openra_bench/scenarios/packs/cat-c6-01.yaml b/openra_bench/scenarios/packs/cat-c6-01.yaml deleted file mode 100644 index 0e209b317aeb75cacbeed9d22467146b4a6cad2a..0000000000000000000000000000000000000000 --- a/openra_bench/scenarios/packs/cat-c6-01.yaml +++ /dev/null @@ -1,84 +0,0 @@ -meta: - id: cat-c6-01 - title: 'Time-Boxed Capital Deploy #2' - capability: reasoning - real_world_meaning: "Provisioning under an energy/money budget and a mission deadline\ - \ \u2014 how much to commit and when." - robotics_analogue: Convert a budget into fielded force AND standing infra before - a clock. - author: catalog-generator -base_map: rush-hour-arena -base: - agent: - faction: allies - enemy: - faction: soviet - tools: - - build - - place_building - - move_units - - deploy - - stop - planning: true - termination: - max_ticks: 40000 - actors: - - type: fact - owner: agent - position: - - 10 - - 18 - - type: powr - owner: agent - position: - - 14 - - 18 - - type: tent - owner: agent - position: - - 12 - - 21 - - type: jeep - owner: agent - position: - - 8 - - 16 - - type: e1 - owner: enemy - position: - - 114 - - 34 - count: 1 - stance: 2 -levels: - easy: - description: "Time-Boxed Capital Deploy: easy \u2014 the decision is harder (less\ - \ info / tighter clock / committed choice)." - win_condition: - all_of: - - own_units_gte: 4 - - building_total_gte: 4 - - within_ticks: 16000 - max_turns: 45 - medium: - description: "Time-Boxed Capital Deploy: medium \u2014 the decision is harder\ - \ (less info / tighter clock / committed choice)." - win_condition: - all_of: - - own_units_gte: 5 - - building_total_gte: 5 - - within_ticks: 12500 - max_turns: 53 - hard: - description: "Time-Boxed Capital Deploy: hard \u2014 the decision is harder (less\ - \ info / tighter clock / committed choice)." - win_condition: - all_of: - - own_units_gte: 6 - - building_total_gte: 6 - - units_lost_lte: 0 - - within_ticks: 9000 - max_turns: 61 - fail_condition: - units_lost_lte: -1 -starting_cash: 1900 diff --git a/openra_bench/scenarios/packs/cat-c6-02.yaml b/openra_bench/scenarios/packs/cat-c6-02.yaml deleted file mode 100644 index 83f22f07c0eeb18b84b5941d7c228e3d7a7fe67c..0000000000000000000000000000000000000000 --- a/openra_bench/scenarios/packs/cat-c6-02.yaml +++ /dev/null @@ -1,85 +0,0 @@ -meta: - id: cat-c6-02 - title: 'Time-Boxed Capital Deploy #3' - capability: reasoning - real_world_meaning: Provisioning under an energy/money budget and a mission deadline — how much to commit - and when. - robotics_analogue: Convert a budget into fielded force AND standing infra before a clock. - author: catalog-generator - status: quarantine - quarantine_reason: 'redundant cat-* over-generation (audit: keep 2 representatives/category)' -base_map: rush-hour-arena -base: - agent: - faction: allies - enemy: - faction: soviet - tools: - - build - - place_building - - move_units - - deploy - - stop - planning: true - termination: - max_ticks: 40000 - actors: - - type: fact - owner: agent - position: - - 10 - - 18 - - type: powr - owner: agent - position: - - 14 - - 18 - - type: tent - owner: agent - position: - - 12 - - 21 - - type: jeep - owner: agent - position: - - 8 - - 16 - - type: e1 - owner: enemy - position: - - 114 - - 34 - count: 1 - stance: 2 -levels: - easy: - description: 'Time-Boxed Capital Deploy: easy — the decision is harder (less info / tighter clock - / committed choice).' - win_condition: - all_of: - - own_units_gte: 4 - - building_total_gte: 4 - - within_ticks: 16000 - max_turns: 45 - medium: - description: 'Time-Boxed Capital Deploy: medium — the decision is harder (less info / tighter clock - / committed choice).' - win_condition: - all_of: - - own_units_gte: 5 - - building_total_gte: 5 - - within_ticks: 12500 - max_turns: 53 - hard: - description: 'Time-Boxed Capital Deploy: hard — the decision is harder (less info / tighter clock - / committed choice).' - win_condition: - all_of: - - own_units_gte: 6 - - building_total_gte: 6 - - units_lost_lte: 0 - - within_ticks: 9000 - max_turns: 61 - fail_condition: - units_lost_lte: -1 -starting_cash: 1800 diff --git a/openra_bench/scenarios/packs/cat-c6-03.yaml b/openra_bench/scenarios/packs/cat-c6-03.yaml deleted file mode 100644 index 73e3f7bc1fffbe9070488ce73efe8b650c3adfb6..0000000000000000000000000000000000000000 --- a/openra_bench/scenarios/packs/cat-c6-03.yaml +++ /dev/null @@ -1,85 +0,0 @@ -meta: - id: cat-c6-03 - title: 'Time-Boxed Capital Deploy #4' - capability: reasoning - real_world_meaning: Provisioning under an energy/money budget and a mission deadline — how much to commit - and when. - robotics_analogue: Convert a budget into fielded force AND standing infra before a clock. - author: catalog-generator - status: quarantine - quarantine_reason: 'redundant cat-* over-generation (audit: keep 2 representatives/category)' -base_map: rush-hour-arena -base: - agent: - faction: allies - enemy: - faction: soviet - tools: - - build - - place_building - - move_units - - deploy - - stop - planning: true - termination: - max_ticks: 40000 - actors: - - type: fact - owner: agent - position: - - 10 - - 18 - - type: powr - owner: agent - position: - - 14 - - 18 - - type: tent - owner: agent - position: - - 12 - - 21 - - type: jeep - owner: agent - position: - - 8 - - 16 - - type: e1 - owner: enemy - position: - - 114 - - 34 - count: 1 - stance: 2 -levels: - easy: - description: 'Time-Boxed Capital Deploy: easy — the decision is harder (less info / tighter clock - / committed choice).' - win_condition: - all_of: - - own_units_gte: 4 - - building_total_gte: 4 - - within_ticks: 16000 - max_turns: 45 - medium: - description: 'Time-Boxed Capital Deploy: medium — the decision is harder (less info / tighter clock - / committed choice).' - win_condition: - all_of: - - own_units_gte: 5 - - building_total_gte: 5 - - within_ticks: 12500 - max_turns: 53 - hard: - description: 'Time-Boxed Capital Deploy: hard — the decision is harder (less info / tighter clock - / committed choice).' - win_condition: - all_of: - - own_units_gte: 6 - - building_total_gte: 6 - - units_lost_lte: 0 - - within_ticks: 9000 - max_turns: 61 - fail_condition: - units_lost_lte: -1 -starting_cash: 1700 diff --git a/openra_bench/scenarios/packs/cat-c6-04.yaml b/openra_bench/scenarios/packs/cat-c6-04.yaml deleted file mode 100644 index abe5b823c1dd715db06812cbfa0d79ab18204149..0000000000000000000000000000000000000000 --- a/openra_bench/scenarios/packs/cat-c6-04.yaml +++ /dev/null @@ -1,85 +0,0 @@ -meta: - id: cat-c6-04 - title: 'Time-Boxed Capital Deploy #5' - capability: reasoning - real_world_meaning: Provisioning under an energy/money budget and a mission deadline — how much to commit - and when. - robotics_analogue: Convert a budget into fielded force AND standing infra before a clock. - author: catalog-generator - status: quarantine - quarantine_reason: 'redundant cat-* over-generation (audit: keep 2 representatives/category)' -base_map: rush-hour-arena -base: - agent: - faction: allies - enemy: - faction: soviet - tools: - - build - - place_building - - move_units - - deploy - - stop - planning: true - termination: - max_ticks: 40000 - actors: - - type: fact - owner: agent - position: - - 10 - - 18 - - type: powr - owner: agent - position: - - 14 - - 18 - - type: tent - owner: agent - position: - - 12 - - 21 - - type: jeep - owner: agent - position: - - 8 - - 16 - - type: e1 - owner: enemy - position: - - 114 - - 34 - count: 1 - stance: 2 -levels: - easy: - description: 'Time-Boxed Capital Deploy: easy — the decision is harder (less info / tighter clock - / committed choice).' - win_condition: - all_of: - - own_units_gte: 4 - - building_total_gte: 4 - - within_ticks: 16000 - max_turns: 45 - medium: - description: 'Time-Boxed Capital Deploy: medium — the decision is harder (less info / tighter clock - / committed choice).' - win_condition: - all_of: - - own_units_gte: 5 - - building_total_gte: 5 - - within_ticks: 12500 - max_turns: 53 - hard: - description: 'Time-Boxed Capital Deploy: hard — the decision is harder (less info / tighter clock - / committed choice).' - win_condition: - all_of: - - own_units_gte: 6 - - building_total_gte: 6 - - units_lost_lte: 0 - - within_ticks: 9000 - max_turns: 61 - fail_condition: - units_lost_lte: -1 -starting_cash: 1600 diff --git a/openra_bench/scenarios/packs/cat-c6-05.yaml b/openra_bench/scenarios/packs/cat-c6-05.yaml deleted file mode 100644 index a6c9339f6b408f8af32aab115421e0f19380289e..0000000000000000000000000000000000000000 --- a/openra_bench/scenarios/packs/cat-c6-05.yaml +++ /dev/null @@ -1,85 +0,0 @@ -meta: - id: cat-c6-05 - title: 'Time-Boxed Capital Deploy #6' - capability: reasoning - real_world_meaning: Provisioning under an energy/money budget and a mission deadline — how much to commit - and when. - robotics_analogue: Convert a budget into fielded force AND standing infra before a clock. - author: catalog-generator - status: quarantine - quarantine_reason: 'redundant cat-* over-generation (audit: keep 2 representatives/category)' -base_map: rush-hour-arena -base: - agent: - faction: allies - enemy: - faction: soviet - tools: - - build - - place_building - - move_units - - deploy - - stop - planning: true - termination: - max_ticks: 40000 - actors: - - type: fact - owner: agent - position: - - 10 - - 18 - - type: powr - owner: agent - position: - - 14 - - 18 - - type: tent - owner: agent - position: - - 12 - - 21 - - type: jeep - owner: agent - position: - - 8 - - 16 - - type: e1 - owner: enemy - position: - - 114 - - 34 - count: 1 - stance: 2 -levels: - easy: - description: 'Time-Boxed Capital Deploy: easy — the decision is harder (less info / tighter clock - / committed choice).' - win_condition: - all_of: - - own_units_gte: 4 - - building_total_gte: 4 - - within_ticks: 16000 - max_turns: 45 - medium: - description: 'Time-Boxed Capital Deploy: medium — the decision is harder (less info / tighter clock - / committed choice).' - win_condition: - all_of: - - own_units_gte: 5 - - building_total_gte: 5 - - within_ticks: 12500 - max_turns: 53 - hard: - description: 'Time-Boxed Capital Deploy: hard — the decision is harder (less info / tighter clock - / committed choice).' - win_condition: - all_of: - - own_units_gte: 6 - - building_total_gte: 6 - - units_lost_lte: 0 - - within_ticks: 9000 - max_turns: 61 - fail_condition: - units_lost_lte: -1 -starting_cash: 1500 diff --git a/openra_bench/scenarios/packs/cat-c7-00.yaml b/openra_bench/scenarios/packs/cat-c7-00.yaml deleted file mode 100644 index 2f120e6937a4ed3087bade0756cb72c7cbbf305e..0000000000000000000000000000000000000000 --- a/openra_bench/scenarios/packs/cat-c7-00.yaml +++ /dev/null @@ -1,95 +0,0 @@ -meta: - id: cat-c7-00 - title: 'Defensive-Direction Commit #1' - capability: perception - real_world_meaning: "Committing scarce defensive assets to an inferred threat axis\ - \ \u2014 spatial decision under directional ambiguity." - robotics_analogue: Place limited defenses toward the most likely approach under - directional uncertainty. - author: catalog-generator -base_map: rush-hour-arena -base: - agent: - faction: allies - enemy: - faction: soviet - tools: - - build - - place_building - - move_units - - deploy - - stop - planning: true - termination: - max_ticks: 40000 - actors: - - type: fact - owner: agent - position: - - 10 - - 18 - - type: powr - owner: agent - position: - - 14 - - 18 - - type: tent - owner: agent - position: - - 12 - - 21 - - type: jeep - owner: agent - position: - - 8 - - 16 - - type: e1 - owner: enemy - position: - - 114 - - 34 - count: 1 - stance: 2 -levels: - easy: - description: "Defensive-Direction Commit: easy \u2014 the decision is harder (less\ - \ info / tighter clock / committed choice)." - win_condition: - all_of: - - building_in_region: - type: pbox - x: 40 - y: 20 - radius: 9 - count: 2 - - within_ticks: 24000 - max_turns: 45 - medium: - description: "Defensive-Direction Commit: medium \u2014 the decision is harder\ - \ (less info / tighter clock / committed choice)." - win_condition: - all_of: - - building_in_region: - type: pbox - x: 40 - y: 20 - radius: 8 - count: 2 - - within_ticks: 19000 - max_turns: 53 - hard: - description: "Defensive-Direction Commit: hard \u2014 the decision is harder (less\ - \ info / tighter clock / committed choice)." - win_condition: - all_of: - - building_in_region: - type: pbox - x: 40 - y: 20 - radius: 7 - count: 3 - - within_ticks: 14000 - max_turns: 61 - fail_condition: - units_lost_lte: -1 -starting_cash: 5000 diff --git a/openra_bench/scenarios/packs/cat-c7-01.yaml b/openra_bench/scenarios/packs/cat-c7-01.yaml deleted file mode 100644 index 57ea2e85e2deaed001d7893cf5d32839722aa6f1..0000000000000000000000000000000000000000 --- a/openra_bench/scenarios/packs/cat-c7-01.yaml +++ /dev/null @@ -1,95 +0,0 @@ -meta: - id: cat-c7-01 - title: 'Defensive-Direction Commit #2' - capability: perception - real_world_meaning: "Committing scarce defensive assets to an inferred threat axis\ - \ \u2014 spatial decision under directional ambiguity." - robotics_analogue: Place limited defenses toward the most likely approach under - directional uncertainty. - author: catalog-generator -base_map: rush-hour-arena -base: - agent: - faction: allies - enemy: - faction: soviet - tools: - - build - - place_building - - move_units - - deploy - - stop - planning: true - termination: - max_ticks: 40000 - actors: - - type: fact - owner: agent - position: - - 10 - - 18 - - type: powr - owner: agent - position: - - 14 - - 18 - - type: tent - owner: agent - position: - - 12 - - 21 - - type: jeep - owner: agent - position: - - 8 - - 16 - - type: e1 - owner: enemy - position: - - 114 - - 34 - count: 1 - stance: 2 -levels: - easy: - description: "Defensive-Direction Commit: easy \u2014 the decision is harder (less\ - \ info / tighter clock / committed choice)." - win_condition: - all_of: - - building_in_region: - type: pbox - x: 45 - y: 20 - radius: 9 - count: 2 - - within_ticks: 24000 - max_turns: 45 - medium: - description: "Defensive-Direction Commit: medium \u2014 the decision is harder\ - \ (less info / tighter clock / committed choice)." - win_condition: - all_of: - - building_in_region: - type: pbox - x: 45 - y: 20 - radius: 8 - count: 2 - - within_ticks: 19000 - max_turns: 53 - hard: - description: "Defensive-Direction Commit: hard \u2014 the decision is harder (less\ - \ info / tighter clock / committed choice)." - win_condition: - all_of: - - building_in_region: - type: pbox - x: 45 - y: 20 - radius: 7 - count: 3 - - within_ticks: 14000 - max_turns: 61 - fail_condition: - units_lost_lte: -1 -starting_cash: 5000 diff --git a/openra_bench/scenarios/packs/cat-c7-02.yaml b/openra_bench/scenarios/packs/cat-c7-02.yaml deleted file mode 100644 index 59be00686b18212a1051c2c34331c0bcc635c5de..0000000000000000000000000000000000000000 --- a/openra_bench/scenarios/packs/cat-c7-02.yaml +++ /dev/null @@ -1,96 +0,0 @@ -meta: - id: cat-c7-02 - title: 'Defensive-Direction Commit #3' - capability: perception - real_world_meaning: Committing scarce defensive assets to an inferred threat axis — spatial decision - under directional ambiguity. - robotics_analogue: Place limited defenses toward the most likely approach under directional uncertainty. - author: catalog-generator - status: quarantine - quarantine_reason: 'redundant cat-* over-generation (audit: keep 2 representatives/category)' -base_map: rush-hour-arena -base: - agent: - faction: allies - enemy: - faction: soviet - tools: - - build - - place_building - - move_units - - deploy - - stop - planning: true - termination: - max_ticks: 40000 - actors: - - type: fact - owner: agent - position: - - 10 - - 18 - - type: powr - owner: agent - position: - - 14 - - 18 - - type: tent - owner: agent - position: - - 12 - - 21 - - type: jeep - owner: agent - position: - - 8 - - 16 - - type: e1 - owner: enemy - position: - - 114 - - 34 - count: 1 - stance: 2 -levels: - easy: - description: 'Defensive-Direction Commit: easy — the decision is harder (less info / tighter clock - / committed choice).' - win_condition: - all_of: - - building_in_region: - type: pbox - x: 50 - y: 20 - radius: 9 - count: 2 - - within_ticks: 24000 - max_turns: 45 - medium: - description: 'Defensive-Direction Commit: medium — the decision is harder (less info / tighter clock - / committed choice).' - win_condition: - all_of: - - building_in_region: - type: pbox - x: 50 - y: 20 - radius: 8 - count: 2 - - within_ticks: 19000 - max_turns: 53 - hard: - description: 'Defensive-Direction Commit: hard — the decision is harder (less info / tighter clock - / committed choice).' - win_condition: - all_of: - - building_in_region: - type: pbox - x: 50 - y: 20 - radius: 7 - count: 3 - - within_ticks: 14000 - max_turns: 61 - fail_condition: - units_lost_lte: -1 -starting_cash: 5000 diff --git a/openra_bench/scenarios/packs/cat-c7-03.yaml b/openra_bench/scenarios/packs/cat-c7-03.yaml deleted file mode 100644 index 684132776a701f7d78e3d087442056627fd6b590..0000000000000000000000000000000000000000 --- a/openra_bench/scenarios/packs/cat-c7-03.yaml +++ /dev/null @@ -1,96 +0,0 @@ -meta: - id: cat-c7-03 - title: 'Defensive-Direction Commit #4' - capability: perception - real_world_meaning: Committing scarce defensive assets to an inferred threat axis — spatial decision - under directional ambiguity. - robotics_analogue: Place limited defenses toward the most likely approach under directional uncertainty. - author: catalog-generator - status: quarantine - quarantine_reason: 'redundant cat-* over-generation (audit: keep 2 representatives/category)' -base_map: rush-hour-arena -base: - agent: - faction: allies - enemy: - faction: soviet - tools: - - build - - place_building - - move_units - - deploy - - stop - planning: true - termination: - max_ticks: 40000 - actors: - - type: fact - owner: agent - position: - - 10 - - 18 - - type: powr - owner: agent - position: - - 14 - - 18 - - type: tent - owner: agent - position: - - 12 - - 21 - - type: jeep - owner: agent - position: - - 8 - - 16 - - type: e1 - owner: enemy - position: - - 114 - - 34 - count: 1 - stance: 2 -levels: - easy: - description: 'Defensive-Direction Commit: easy — the decision is harder (less info / tighter clock - / committed choice).' - win_condition: - all_of: - - building_in_region: - type: pbox - x: 55 - y: 20 - radius: 9 - count: 2 - - within_ticks: 24000 - max_turns: 45 - medium: - description: 'Defensive-Direction Commit: medium — the decision is harder (less info / tighter clock - / committed choice).' - win_condition: - all_of: - - building_in_region: - type: pbox - x: 55 - y: 20 - radius: 8 - count: 2 - - within_ticks: 19000 - max_turns: 53 - hard: - description: 'Defensive-Direction Commit: hard — the decision is harder (less info / tighter clock - / committed choice).' - win_condition: - all_of: - - building_in_region: - type: pbox - x: 55 - y: 20 - radius: 7 - count: 3 - - within_ticks: 14000 - max_turns: 61 - fail_condition: - units_lost_lte: -1 -starting_cash: 5000 diff --git a/openra_bench/scenarios/packs/cat-c7-04.yaml b/openra_bench/scenarios/packs/cat-c7-04.yaml deleted file mode 100644 index 1f8da7db498274e45c4948c3d20a38b2f4ba08d1..0000000000000000000000000000000000000000 --- a/openra_bench/scenarios/packs/cat-c7-04.yaml +++ /dev/null @@ -1,96 +0,0 @@ -meta: - id: cat-c7-04 - title: 'Defensive-Direction Commit #5' - capability: perception - real_world_meaning: Committing scarce defensive assets to an inferred threat axis — spatial decision - under directional ambiguity. - robotics_analogue: Place limited defenses toward the most likely approach under directional uncertainty. - author: catalog-generator - status: quarantine - quarantine_reason: 'redundant cat-* over-generation (audit: keep 2 representatives/category)' -base_map: rush-hour-arena -base: - agent: - faction: allies - enemy: - faction: soviet - tools: - - build - - place_building - - move_units - - deploy - - stop - planning: true - termination: - max_ticks: 40000 - actors: - - type: fact - owner: agent - position: - - 10 - - 18 - - type: powr - owner: agent - position: - - 14 - - 18 - - type: tent - owner: agent - position: - - 12 - - 21 - - type: jeep - owner: agent - position: - - 8 - - 16 - - type: e1 - owner: enemy - position: - - 114 - - 34 - count: 1 - stance: 2 -levels: - easy: - description: 'Defensive-Direction Commit: easy — the decision is harder (less info / tighter clock - / committed choice).' - win_condition: - all_of: - - building_in_region: - type: pbox - x: 60 - y: 20 - radius: 9 - count: 2 - - within_ticks: 24000 - max_turns: 45 - medium: - description: 'Defensive-Direction Commit: medium — the decision is harder (less info / tighter clock - / committed choice).' - win_condition: - all_of: - - building_in_region: - type: pbox - x: 60 - y: 20 - radius: 8 - count: 2 - - within_ticks: 19000 - max_turns: 53 - hard: - description: 'Defensive-Direction Commit: hard — the decision is harder (less info / tighter clock - / committed choice).' - win_condition: - all_of: - - building_in_region: - type: pbox - x: 60 - y: 20 - radius: 7 - count: 3 - - within_ticks: 14000 - max_turns: 61 - fail_condition: - units_lost_lte: -1 -starting_cash: 5000 diff --git a/openra_bench/scenarios/packs/cat-c7-05.yaml b/openra_bench/scenarios/packs/cat-c7-05.yaml deleted file mode 100644 index 155ac5ceabdc86d811d4cea46f9be320184a6362..0000000000000000000000000000000000000000 --- a/openra_bench/scenarios/packs/cat-c7-05.yaml +++ /dev/null @@ -1,96 +0,0 @@ -meta: - id: cat-c7-05 - title: 'Defensive-Direction Commit #6' - capability: perception - real_world_meaning: Committing scarce defensive assets to an inferred threat axis — spatial decision - under directional ambiguity. - robotics_analogue: Place limited defenses toward the most likely approach under directional uncertainty. - author: catalog-generator - status: quarantine - quarantine_reason: 'redundant cat-* over-generation (audit: keep 2 representatives/category)' -base_map: rush-hour-arena -base: - agent: - faction: allies - enemy: - faction: soviet - tools: - - build - - place_building - - move_units - - deploy - - stop - planning: true - termination: - max_ticks: 40000 - actors: - - type: fact - owner: agent - position: - - 10 - - 18 - - type: powr - owner: agent - position: - - 14 - - 18 - - type: tent - owner: agent - position: - - 12 - - 21 - - type: jeep - owner: agent - position: - - 8 - - 16 - - type: e1 - owner: enemy - position: - - 114 - - 34 - count: 1 - stance: 2 -levels: - easy: - description: 'Defensive-Direction Commit: easy — the decision is harder (less info / tighter clock - / committed choice).' - win_condition: - all_of: - - building_in_region: - type: pbox - x: 65 - y: 20 - radius: 9 - count: 2 - - within_ticks: 24000 - max_turns: 45 - medium: - description: 'Defensive-Direction Commit: medium — the decision is harder (less info / tighter clock - / committed choice).' - win_condition: - all_of: - - building_in_region: - type: pbox - x: 65 - y: 20 - radius: 8 - count: 2 - - within_ticks: 19000 - max_turns: 53 - hard: - description: 'Defensive-Direction Commit: hard — the decision is harder (less info / tighter clock - / committed choice).' - win_condition: - all_of: - - building_in_region: - type: pbox - x: 65 - y: 20 - radius: 7 - count: 3 - - within_ticks: 14000 - max_turns: 61 - fail_condition: - units_lost_lte: -1 -starting_cash: 5000 diff --git a/openra_bench/scenarios/packs/cat-c8-00.yaml b/openra_bench/scenarios/packs/cat-c8-00.yaml deleted file mode 100644 index 4fb3ebc5dd35e63d7bfee6fb0d5d8b751e54a9cd..0000000000000000000000000000000000000000 --- a/openra_bench/scenarios/packs/cat-c8-00.yaml +++ /dev/null @@ -1,93 +0,0 @@ -meta: - id: cat-c8-00 - title: 'Base-Placement & Staging #1' - capability: perception - real_world_meaning: "Choosing a base site that satisfies spatial constraints, then\ - \ pre-positioning \u2014 facility-layout + staging." - robotics_analogue: Found a base in a designated region, then stage toward a threat - axis. - author: catalog-generator -base_map: rush-hour-arena -base: - agent: - faction: allies - enemy: - faction: soviet - tools: - - build - - place_building - - move_units - - deploy - - stop - planning: true - termination: - max_ticks: 40000 - actors: - - type: fact - owner: agent - position: - - 10 - - 18 - - type: powr - owner: agent - position: - - 14 - - 18 - - type: jeep - owner: agent - position: - - 8 - - 16 - - type: e1 - owner: enemy - position: - - 114 - - 34 - count: 1 - stance: 2 -levels: - easy: - description: "Base-Placement & Staging: easy \u2014 the decision is harder (less\ - \ info / tighter clock / committed choice)." - win_condition: - all_of: - - building_in_region: - type: powr - x: 30 - y: 18 - radius: 10 - count: 1 - - building_total_gte: 3 - - within_ticks: 22000 - max_turns: 45 - medium: - description: "Base-Placement & Staging: medium \u2014 the decision is harder (less\ - \ info / tighter clock / committed choice)." - win_condition: - all_of: - - building_in_region: - type: powr - x: 30 - y: 18 - radius: 8 - count: 1 - - building_total_gte: 4 - - within_ticks: 17000 - max_turns: 53 - hard: - description: "Base-Placement & Staging: hard \u2014 the decision is harder (less\ - \ info / tighter clock / committed choice)." - win_condition: - all_of: - - building_in_region: - type: powr - x: 30 - y: 18 - radius: 6 - count: 1 - - building_total_gte: 5 - - within_ticks: 12000 - max_turns: 61 - fail_condition: - units_lost_lte: -1 -starting_cash: 4000 diff --git a/openra_bench/scenarios/packs/cat-c8-01.yaml b/openra_bench/scenarios/packs/cat-c8-01.yaml deleted file mode 100644 index 6ed74bc9c0e0f9de7fff59b7aa6cb1fbe30d2301..0000000000000000000000000000000000000000 --- a/openra_bench/scenarios/packs/cat-c8-01.yaml +++ /dev/null @@ -1,93 +0,0 @@ -meta: - id: cat-c8-01 - title: 'Base-Placement & Staging #2' - capability: perception - real_world_meaning: "Choosing a base site that satisfies spatial constraints, then\ - \ pre-positioning \u2014 facility-layout + staging." - robotics_analogue: Found a base in a designated region, then stage toward a threat - axis. - author: catalog-generator -base_map: rush-hour-arena -base: - agent: - faction: allies - enemy: - faction: soviet - tools: - - build - - place_building - - move_units - - deploy - - stop - planning: true - termination: - max_ticks: 40000 - actors: - - type: fact - owner: agent - position: - - 10 - - 18 - - type: powr - owner: agent - position: - - 14 - - 18 - - type: jeep - owner: agent - position: - - 8 - - 16 - - type: e1 - owner: enemy - position: - - 114 - - 34 - count: 1 - stance: 2 -levels: - easy: - description: "Base-Placement & Staging: easy \u2014 the decision is harder (less\ - \ info / tighter clock / committed choice)." - win_condition: - all_of: - - building_in_region: - type: powr - x: 34 - y: 18 - radius: 10 - count: 1 - - building_total_gte: 3 - - within_ticks: 22000 - max_turns: 45 - medium: - description: "Base-Placement & Staging: medium \u2014 the decision is harder (less\ - \ info / tighter clock / committed choice)." - win_condition: - all_of: - - building_in_region: - type: powr - x: 34 - y: 18 - radius: 8 - count: 1 - - building_total_gte: 4 - - within_ticks: 17000 - max_turns: 53 - hard: - description: "Base-Placement & Staging: hard \u2014 the decision is harder (less\ - \ info / tighter clock / committed choice)." - win_condition: - all_of: - - building_in_region: - type: powr - x: 34 - y: 18 - radius: 6 - count: 1 - - building_total_gte: 5 - - within_ticks: 12000 - max_turns: 61 - fail_condition: - units_lost_lte: -1 -starting_cash: 4000 diff --git a/openra_bench/scenarios/packs/cat-c8-02.yaml b/openra_bench/scenarios/packs/cat-c8-02.yaml deleted file mode 100644 index 8c4d52662be5130242a5c947da23b51b762c2a77..0000000000000000000000000000000000000000 --- a/openra_bench/scenarios/packs/cat-c8-02.yaml +++ /dev/null @@ -1,94 +0,0 @@ -meta: - id: cat-c8-02 - title: 'Base-Placement & Staging #3' - capability: perception - real_world_meaning: Choosing a base site that satisfies spatial constraints, then pre-positioning — - facility-layout + staging. - robotics_analogue: Found a base in a designated region, then stage toward a threat axis. - author: catalog-generator - status: quarantine - quarantine_reason: 'redundant cat-* over-generation (audit: keep 2 representatives/category)' -base_map: rush-hour-arena -base: - agent: - faction: allies - enemy: - faction: soviet - tools: - - build - - place_building - - move_units - - deploy - - stop - planning: true - termination: - max_ticks: 40000 - actors: - - type: fact - owner: agent - position: - - 10 - - 18 - - type: powr - owner: agent - position: - - 14 - - 18 - - type: jeep - owner: agent - position: - - 8 - - 16 - - type: e1 - owner: enemy - position: - - 114 - - 34 - count: 1 - stance: 2 -levels: - easy: - description: 'Base-Placement & Staging: easy — the decision is harder (less info / tighter clock / - committed choice).' - win_condition: - all_of: - - building_in_region: - type: powr - x: 38 - y: 18 - radius: 10 - count: 1 - - building_total_gte: 3 - - within_ticks: 22000 - max_turns: 45 - medium: - description: 'Base-Placement & Staging: medium — the decision is harder (less info / tighter clock - / committed choice).' - win_condition: - all_of: - - building_in_region: - type: powr - x: 38 - y: 18 - radius: 8 - count: 1 - - building_total_gte: 4 - - within_ticks: 17000 - max_turns: 53 - hard: - description: 'Base-Placement & Staging: hard — the decision is harder (less info / tighter clock / - committed choice).' - win_condition: - all_of: - - building_in_region: - type: powr - x: 38 - y: 18 - radius: 6 - count: 1 - - building_total_gte: 5 - - within_ticks: 12000 - max_turns: 61 - fail_condition: - units_lost_lte: -1 -starting_cash: 4000 diff --git a/openra_bench/scenarios/packs/cat-c8-03.yaml b/openra_bench/scenarios/packs/cat-c8-03.yaml deleted file mode 100644 index 8af42436a7f5275aea3dca840d3cadbbef78b94a..0000000000000000000000000000000000000000 --- a/openra_bench/scenarios/packs/cat-c8-03.yaml +++ /dev/null @@ -1,94 +0,0 @@ -meta: - id: cat-c8-03 - title: 'Base-Placement & Staging #4' - capability: perception - real_world_meaning: Choosing a base site that satisfies spatial constraints, then pre-positioning — - facility-layout + staging. - robotics_analogue: Found a base in a designated region, then stage toward a threat axis. - author: catalog-generator - status: quarantine - quarantine_reason: 'redundant cat-* over-generation (audit: keep 2 representatives/category)' -base_map: rush-hour-arena -base: - agent: - faction: allies - enemy: - faction: soviet - tools: - - build - - place_building - - move_units - - deploy - - stop - planning: true - termination: - max_ticks: 40000 - actors: - - type: fact - owner: agent - position: - - 10 - - 18 - - type: powr - owner: agent - position: - - 14 - - 18 - - type: jeep - owner: agent - position: - - 8 - - 16 - - type: e1 - owner: enemy - position: - - 114 - - 34 - count: 1 - stance: 2 -levels: - easy: - description: 'Base-Placement & Staging: easy — the decision is harder (less info / tighter clock / - committed choice).' - win_condition: - all_of: - - building_in_region: - type: powr - x: 42 - y: 18 - radius: 10 - count: 1 - - building_total_gte: 3 - - within_ticks: 22000 - max_turns: 45 - medium: - description: 'Base-Placement & Staging: medium — the decision is harder (less info / tighter clock - / committed choice).' - win_condition: - all_of: - - building_in_region: - type: powr - x: 42 - y: 18 - radius: 8 - count: 1 - - building_total_gte: 4 - - within_ticks: 17000 - max_turns: 53 - hard: - description: 'Base-Placement & Staging: hard — the decision is harder (less info / tighter clock / - committed choice).' - win_condition: - all_of: - - building_in_region: - type: powr - x: 42 - y: 18 - radius: 6 - count: 1 - - building_total_gte: 5 - - within_ticks: 12000 - max_turns: 61 - fail_condition: - units_lost_lte: -1 -starting_cash: 4000 diff --git a/openra_bench/scenarios/packs/cat-c8-04.yaml b/openra_bench/scenarios/packs/cat-c8-04.yaml deleted file mode 100644 index 5c2cc7adbdbd44b70f305817f3f4d8007b7da90c..0000000000000000000000000000000000000000 --- a/openra_bench/scenarios/packs/cat-c8-04.yaml +++ /dev/null @@ -1,94 +0,0 @@ -meta: - id: cat-c8-04 - title: 'Base-Placement & Staging #5' - capability: perception - real_world_meaning: Choosing a base site that satisfies spatial constraints, then pre-positioning — - facility-layout + staging. - robotics_analogue: Found a base in a designated region, then stage toward a threat axis. - author: catalog-generator - status: quarantine - quarantine_reason: 'redundant cat-* over-generation (audit: keep 2 representatives/category)' -base_map: rush-hour-arena -base: - agent: - faction: allies - enemy: - faction: soviet - tools: - - build - - place_building - - move_units - - deploy - - stop - planning: true - termination: - max_ticks: 40000 - actors: - - type: fact - owner: agent - position: - - 10 - - 18 - - type: powr - owner: agent - position: - - 14 - - 18 - - type: jeep - owner: agent - position: - - 8 - - 16 - - type: e1 - owner: enemy - position: - - 114 - - 34 - count: 1 - stance: 2 -levels: - easy: - description: 'Base-Placement & Staging: easy — the decision is harder (less info / tighter clock / - committed choice).' - win_condition: - all_of: - - building_in_region: - type: powr - x: 46 - y: 18 - radius: 10 - count: 1 - - building_total_gte: 3 - - within_ticks: 22000 - max_turns: 45 - medium: - description: 'Base-Placement & Staging: medium — the decision is harder (less info / tighter clock - / committed choice).' - win_condition: - all_of: - - building_in_region: - type: powr - x: 46 - y: 18 - radius: 8 - count: 1 - - building_total_gte: 4 - - within_ticks: 17000 - max_turns: 53 - hard: - description: 'Base-Placement & Staging: hard — the decision is harder (less info / tighter clock / - committed choice).' - win_condition: - all_of: - - building_in_region: - type: powr - x: 46 - y: 18 - radius: 6 - count: 1 - - building_total_gte: 5 - - within_ticks: 12000 - max_turns: 61 - fail_condition: - units_lost_lte: -1 -starting_cash: 4000 diff --git a/openra_bench/scenarios/packs/cat-c9-00.yaml b/openra_bench/scenarios/packs/cat-c9-00.yaml deleted file mode 100644 index d7c9acca0be65ce3723282ae69342923b4ae8319..0000000000000000000000000000000000000000 --- a/openra_bench/scenarios/packs/cat-c9-00.yaml +++ /dev/null @@ -1,86 +0,0 @@ -meta: - id: cat-c9-00 - title: 'Commit-vs-Retreat #1' - capability: reasoning - real_world_meaning: "The core risk call: commit force on visible info vs wait for\ - \ more \u2014 decision under uncertainty with a loss budget." - robotics_analogue: Engage vs hold under partial info with an attrition cap. - author: catalog-generator -base_map: rush-hour-arena -base: - agent: - faction: allies - enemy: - faction: soviet - tools: - - move_units - - attack_unit - - attack_move - - stop - planning: true - termination: - max_ticks: 30000 - actors: - - type: 2tnk - owner: agent - position: - - 8 - - 18 - count: 2 - - type: e1 - owner: agent - position: - - 9 - - 20 - count: 3 - - type: e1 - owner: enemy - position: - - 100 - - 20 - count: 2 - stance: 2 - - type: e3 - owner: enemy - position: - - 104 - - 24 - count: 1 - stance: 2 - - type: e1 - owner: enemy - position: - - 40 - - 10 - count: 1 - stance: 2 -levels: - easy: - description: "Commit-vs-Retreat: easy \u2014 the decision is harder (less info\ - \ / tighter clock / committed choice)." - win_condition: - all_of: - - units_killed_gte: 2 - - units_lost_lte: 3 - - within_ticks: 14000 - max_turns: 45 - medium: - description: "Commit-vs-Retreat: medium \u2014 the decision is harder (less info\ - \ / tighter clock / committed choice)." - win_condition: - all_of: - - units_killed_gte: 3 - - units_lost_lte: 2 - - within_ticks: 11000 - max_turns: 53 - hard: - description: "Commit-vs-Retreat: hard \u2014 the decision is harder (less info\ - \ / tighter clock / committed choice)." - win_condition: - all_of: - - units_killed_gte: 4 - - units_lost_lte: 1 - - within_ticks: 8000 - max_turns: 61 - fail_condition: - units_lost_lte: -1 diff --git a/openra_bench/scenarios/packs/cat-c9-01.yaml b/openra_bench/scenarios/packs/cat-c9-01.yaml deleted file mode 100644 index c72346083e5c7e70eeed5d1ed2e8e76b65d01ab6..0000000000000000000000000000000000000000 --- a/openra_bench/scenarios/packs/cat-c9-01.yaml +++ /dev/null @@ -1,93 +0,0 @@ -meta: - id: cat-c9-01 - title: 'Commit-vs-Retreat #2' - capability: reasoning - real_world_meaning: "The core risk call: commit force on visible info vs wait for\ - \ more \u2014 decision under uncertainty with a loss budget." - robotics_analogue: Engage vs hold under partial info with an attrition cap. - author: catalog-generator -base_map: rush-hour-arena -base: - agent: - faction: allies - enemy: - faction: soviet - tools: - - move_units - - attack_unit - - attack_move - - stop - planning: true - termination: - max_ticks: 30000 - actors: - - type: 2tnk - owner: agent - position: - - 8 - - 18 - count: 2 - - type: e1 - owner: agent - position: - - 9 - - 20 - count: 3 - - type: e1 - owner: enemy - position: - - 100 - - 20 - count: 2 - stance: 2 - - type: e3 - owner: enemy - position: - - 104 - - 24 - count: 1 - stance: 2 - - type: e1 - owner: enemy - position: - - 40 - - 10 - count: 1 - stance: 2 - - type: e1 - owner: enemy - position: - - 48 - - 10 - count: 1 - stance: 2 -levels: - easy: - description: "Commit-vs-Retreat: easy \u2014 the decision is harder (less info\ - \ / tighter clock / committed choice)." - win_condition: - all_of: - - units_killed_gte: 2 - - units_lost_lte: 3 - - within_ticks: 14000 - max_turns: 45 - medium: - description: "Commit-vs-Retreat: medium \u2014 the decision is harder (less info\ - \ / tighter clock / committed choice)." - win_condition: - all_of: - - units_killed_gte: 3 - - units_lost_lte: 2 - - within_ticks: 11000 - max_turns: 53 - hard: - description: "Commit-vs-Retreat: hard \u2014 the decision is harder (less info\ - \ / tighter clock / committed choice)." - win_condition: - all_of: - - units_killed_gte: 4 - - units_lost_lte: 1 - - within_ticks: 8000 - max_turns: 61 - fail_condition: - units_lost_lte: -1 diff --git a/openra_bench/scenarios/packs/cat-c9-02.yaml b/openra_bench/scenarios/packs/cat-c9-02.yaml deleted file mode 100644 index 777dd00e5efade6667993b910c1956fe9478141d..0000000000000000000000000000000000000000 --- a/openra_bench/scenarios/packs/cat-c9-02.yaml +++ /dev/null @@ -1,102 +0,0 @@ -meta: - id: cat-c9-02 - title: 'Commit-vs-Retreat #3' - capability: reasoning - real_world_meaning: 'The core risk call: commit force on visible info vs wait for more — decision under - uncertainty with a loss budget.' - robotics_analogue: Engage vs hold under partial info with an attrition cap. - author: catalog-generator - status: quarantine - quarantine_reason: 'redundant cat-* over-generation (audit: keep 2 representatives/category)' -base_map: rush-hour-arena -base: - agent: - faction: allies - enemy: - faction: soviet - tools: - - move_units - - attack_unit - - attack_move - - stop - planning: true - termination: - max_ticks: 30000 - actors: - - type: 2tnk - owner: agent - position: - - 8 - - 18 - count: 2 - - type: e1 - owner: agent - position: - - 9 - - 20 - count: 3 - - type: e1 - owner: enemy - position: - - 100 - - 20 - count: 2 - stance: 2 - - type: e3 - owner: enemy - position: - - 104 - - 24 - count: 1 - stance: 2 - - type: e1 - owner: enemy - position: - - 40 - - 10 - count: 1 - stance: 2 - - type: e1 - owner: enemy - position: - - 48 - - 10 - count: 1 - stance: 2 - - type: e1 - owner: enemy - position: - - 56 - - 10 - count: 1 - stance: 2 -levels: - easy: - description: 'Commit-vs-Retreat: easy — the decision is harder (less info / tighter clock / committed - choice).' - win_condition: - all_of: - - units_killed_gte: 2 - - units_lost_lte: 3 - - within_ticks: 14000 - max_turns: 45 - medium: - description: 'Commit-vs-Retreat: medium — the decision is harder (less info / tighter clock / committed - choice).' - win_condition: - all_of: - - units_killed_gte: 3 - - units_lost_lte: 2 - - within_ticks: 11000 - max_turns: 53 - hard: - description: 'Commit-vs-Retreat: hard — the decision is harder (less info / tighter clock / committed - choice).' - win_condition: - all_of: - - units_killed_gte: 4 - - units_lost_lte: 1 - - within_ticks: 8000 - max_turns: 61 - fail_condition: - units_lost_lte: -1 diff --git a/openra_bench/scenarios/packs/cat-c9-03.yaml b/openra_bench/scenarios/packs/cat-c9-03.yaml deleted file mode 100644 index 54b907f592c6c27ee885094838228535354f05ac..0000000000000000000000000000000000000000 --- a/openra_bench/scenarios/packs/cat-c9-03.yaml +++ /dev/null @@ -1,88 +0,0 @@ -meta: - id: cat-c9-03 - title: 'Commit-vs-Retreat #4' - capability: reasoning - real_world_meaning: 'The core risk call: commit force on visible info vs wait for more — decision under - uncertainty with a loss budget.' - robotics_analogue: Engage vs hold under partial info with an attrition cap. - author: catalog-generator - status: quarantine - quarantine_reason: 'redundant cat-* over-generation (audit: keep 2 representatives/category)' -base_map: rush-hour-arena -base: - agent: - faction: allies - enemy: - faction: soviet - tools: - - move_units - - attack_unit - - attack_move - - stop - planning: true - termination: - max_ticks: 30000 - actors: - - type: 2tnk - owner: agent - position: - - 8 - - 18 - count: 2 - - type: e1 - owner: agent - position: - - 9 - - 20 - count: 3 - - type: e1 - owner: enemy - position: - - 100 - - 20 - count: 2 - stance: 2 - - type: e3 - owner: enemy - position: - - 104 - - 24 - count: 1 - stance: 2 - - type: e1 - owner: enemy - position: - - 40 - - 10 - count: 1 - stance: 2 -levels: - easy: - description: 'Commit-vs-Retreat: easy — the decision is harder (less info / tighter clock / committed - choice).' - win_condition: - all_of: - - units_killed_gte: 2 - - units_lost_lte: 3 - - within_ticks: 14000 - max_turns: 45 - medium: - description: 'Commit-vs-Retreat: medium — the decision is harder (less info / tighter clock / committed - choice).' - win_condition: - all_of: - - units_killed_gte: 3 - - units_lost_lte: 2 - - within_ticks: 11000 - max_turns: 53 - hard: - description: 'Commit-vs-Retreat: hard — the decision is harder (less info / tighter clock / committed - choice).' - win_condition: - all_of: - - units_killed_gte: 4 - - units_lost_lte: 1 - - within_ticks: 8000 - max_turns: 61 - fail_condition: - units_lost_lte: -1 diff --git a/openra_bench/scenarios/packs/cat-c9-04.yaml b/openra_bench/scenarios/packs/cat-c9-04.yaml deleted file mode 100644 index f5ab2c1e28322c4351872dc6782fe893dc07015f..0000000000000000000000000000000000000000 --- a/openra_bench/scenarios/packs/cat-c9-04.yaml +++ /dev/null @@ -1,95 +0,0 @@ -meta: - id: cat-c9-04 - title: 'Commit-vs-Retreat #5' - capability: reasoning - real_world_meaning: 'The core risk call: commit force on visible info vs wait for more — decision under - uncertainty with a loss budget.' - robotics_analogue: Engage vs hold under partial info with an attrition cap. - author: catalog-generator - status: quarantine - quarantine_reason: 'redundant cat-* over-generation (audit: keep 2 representatives/category)' -base_map: rush-hour-arena -base: - agent: - faction: allies - enemy: - faction: soviet - tools: - - move_units - - attack_unit - - attack_move - - stop - planning: true - termination: - max_ticks: 30000 - actors: - - type: 2tnk - owner: agent - position: - - 8 - - 18 - count: 2 - - type: e1 - owner: agent - position: - - 9 - - 20 - count: 3 - - type: e1 - owner: enemy - position: - - 100 - - 20 - count: 2 - stance: 2 - - type: e3 - owner: enemy - position: - - 104 - - 24 - count: 1 - stance: 2 - - type: e1 - owner: enemy - position: - - 40 - - 10 - count: 1 - stance: 2 - - type: e1 - owner: enemy - position: - - 48 - - 10 - count: 1 - stance: 2 -levels: - easy: - description: 'Commit-vs-Retreat: easy — the decision is harder (less info / tighter clock / committed - choice).' - win_condition: - all_of: - - units_killed_gte: 2 - - units_lost_lte: 3 - - within_ticks: 14000 - max_turns: 45 - medium: - description: 'Commit-vs-Retreat: medium — the decision is harder (less info / tighter clock / committed - choice).' - win_condition: - all_of: - - units_killed_gte: 3 - - units_lost_lte: 2 - - within_ticks: 11000 - max_turns: 53 - hard: - description: 'Commit-vs-Retreat: hard — the decision is harder (less info / tighter clock / committed - choice).' - win_condition: - all_of: - - units_killed_gte: 4 - - units_lost_lte: 1 - - within_ticks: 8000 - max_turns: 61 - fail_condition: - units_lost_lte: -1 diff --git a/openra_bench/scenarios/packs/cat-c9-05.yaml b/openra_bench/scenarios/packs/cat-c9-05.yaml deleted file mode 100644 index 8585ed5abefbb3d8bb76719294012b5460488d20..0000000000000000000000000000000000000000 --- a/openra_bench/scenarios/packs/cat-c9-05.yaml +++ /dev/null @@ -1,102 +0,0 @@ -meta: - id: cat-c9-05 - title: 'Commit-vs-Retreat #6' - capability: reasoning - real_world_meaning: 'The core risk call: commit force on visible info vs wait for more — decision under - uncertainty with a loss budget.' - robotics_analogue: Engage vs hold under partial info with an attrition cap. - author: catalog-generator - status: quarantine - quarantine_reason: 'redundant cat-* over-generation (audit: keep 2 representatives/category)' -base_map: rush-hour-arena -base: - agent: - faction: allies - enemy: - faction: soviet - tools: - - move_units - - attack_unit - - attack_move - - stop - planning: true - termination: - max_ticks: 30000 - actors: - - type: 2tnk - owner: agent - position: - - 8 - - 18 - count: 2 - - type: e1 - owner: agent - position: - - 9 - - 20 - count: 3 - - type: e1 - owner: enemy - position: - - 100 - - 20 - count: 2 - stance: 2 - - type: e3 - owner: enemy - position: - - 104 - - 24 - count: 1 - stance: 2 - - type: e1 - owner: enemy - position: - - 40 - - 10 - count: 1 - stance: 2 - - type: e1 - owner: enemy - position: - - 48 - - 10 - count: 1 - stance: 2 - - type: e1 - owner: enemy - position: - - 56 - - 10 - count: 1 - stance: 2 -levels: - easy: - description: 'Commit-vs-Retreat: easy — the decision is harder (less info / tighter clock / committed - choice).' - win_condition: - all_of: - - units_killed_gte: 2 - - units_lost_lte: 3 - - within_ticks: 14000 - max_turns: 45 - medium: - description: 'Commit-vs-Retreat: medium — the decision is harder (less info / tighter clock / committed - choice).' - win_condition: - all_of: - - units_killed_gte: 3 - - units_lost_lte: 2 - - within_ticks: 11000 - max_turns: 53 - hard: - description: 'Commit-vs-Retreat: hard — the decision is harder (less info / tighter clock / committed - choice).' - win_condition: - all_of: - - units_killed_gte: 4 - - units_lost_lte: 1 - - within_ticks: 8000 - max_turns: 61 - fail_condition: - units_lost_lte: -1 diff --git a/tests/test_catalog.py b/tests/test_catalog.py deleted file mode 100644 index 113653fb9d94aa736647f303fa17cc9c979d8dd5..0000000000000000000000000000000000000000 --- a/tests/test_catalog.py +++ /dev/null @@ -1,63 +0,0 @@ -"""Catalog integrity: the research-grounded 200-level suite is complete, -categorized, capability-tagged, and difficulty-monotone (the controlled --ladder requirement from Procgen/SMACv2/SmartPlay). Pure/schema only — -the engine-run gate lives in test_robustness.""" - -from __future__ import annotations - -import re -from collections import Counter - -from openra_bench.scenarios import load_pack -from openra_bench.scenarios.loader import PACKS_DIR - -CATEGORIES = [f"c{i}" for i in range(1, 13)] - - -def _catalog_files(): - return sorted(PACKS_DIR.glob("cat-*.yaml")) - - -def test_twelve_categories_and_about_200_levels(): - files = _catalog_files() - cats = Counter(re.match(r"cat-(c\d+)-", f.name).group(1) for f in files) - assert set(cats) == set(CATEGORIES), f"missing/extra categories: {sorted(cats)}" - for c in CATEGORIES: - assert cats[c] >= 5, f"category {c} has too few packs: {cats[c]}" - total_levels = len(files) * 3 - assert 195 <= total_levels <= 215, f"expected ~200 levels, got {total_levels}" - - -def test_every_catalog_pack_tagged_and_meaningful(): - for f in _catalog_files(): - pk = load_pack(f) - assert pk.meta.capability in {"perception", "reasoning", "action"} - # Research rule: every scenario must carry a genuine - # real-world / robotics meaning tying it to a capability. - assert len(pk.meta.real_world_meaning) >= 20 - assert len(pk.meta.robotics_analogue) >= 10 - assert set(pk.levels) == {"easy", "medium", "hard"} - - -def _deadline(level) -> int | None: - """Pull the within_ticks bound from an all_of win condition.""" - node = dict(level.win_condition.__pydantic_extra__ or {}) - for clause in node.get("all_of", []): - if "within_ticks" in clause: - return int(clause["within_ticks"]) - return None - - -def test_difficulty_is_monotone(): - """easy→medium→hard must get *harder*: the time budget must not - increase (decision-hardness ladder, not raw scaling).""" - for f in _catalog_files(): - pk = load_pack(f) - ds = [_deadline(pk.levels[l]) for l in ("easy", "medium", "hard")] - if all(d is not None for d in ds): - assert ds[0] >= ds[1] >= ds[2], f"{f.name}: non-monotone clock {ds}" - - -def test_capability_coverage_spans_all_three_links(): - caps = {load_pack(f).meta.capability for f in _catalog_files()} - assert caps == {"perception", "reasoning", "action"}, caps diff --git a/tests/test_hygiene.py b/tests/test_hygiene.py index b71ac71cd9224369f132df0645fbd4a62feca951..a32f3bd74ecfb652d336b3e7c61f6952f3b00458 100644 --- a/tests/test_hygiene.py +++ b/tests/test_hygiene.py @@ -1,18 +1,20 @@ -"""Scenario hygiene: quarantine excludes from the default sweep. - -Quarantined packs (redundant cat-* over-generation; harvest packs -blocked on engine S0/S1) must (a) still parse, (b) drop out of the -default evaluate set into `skipped`, and (c) remain runnable when -named explicitly. Two representatives per cat-* category stay active. +"""Scenario hygiene: the curated set + the quarantine mechanism. + +The auto-generated cat-* corpus was retired (deliberate curation — +keep only hand-authored, rigor-checked decision problems). What +remains under test: every active pack is hand-authored and meaningful, +and the quarantine mechanism (files on disk, excluded from the default +sweep, still runnable by explicit --packs) still holds via the harvest +packs blocked on the engine S0/S1 prerequisite. """ from __future__ import annotations import glob import os -from collections import Counter from pathlib import Path +from openra_bench import run_eval from openra_bench.scenarios import load_pack PACKS = Path(__file__).parent.parent / "openra_bench" / "scenarios" / "packs" @@ -27,9 +29,19 @@ def _all_metas(): return out -def test_all_packs_still_parse_after_retag(): +def test_no_autogenerated_cat_packs_remain(): + assert not list(PACKS.glob("cat-*.yaml")), "cat-* corpus must be gone" + + +def test_curated_set_is_small_handauthored_and_meaningful(): metas = _all_metas() - assert len(metas) > 80 # nothing lost on disk + active = [(n, m) for n, m in metas if m.status == "active"] + # Deliberately curated: a couple dozen, not hundreds. + assert 15 <= len(active) <= 40, f"active set size {len(active)}" + for name, m in active: + assert not name.startswith("cat-") + assert len(m.real_world_meaning) >= 20 # genuine transfer story + assert len(m.robotics_analogue) >= 10 for name, m in metas: assert m.status in ("active", "quarantine") if m.status == "quarantine": @@ -43,54 +55,32 @@ def test_harvest_quarantined_with_engine_reason(): assert "S0/S1" in m.quarantine_reason -def test_two_representatives_kept_active_per_cat_category(): - active = Counter() - quar = Counter() - for name, m in _all_metas(): - if not name.startswith("cat-c"): - continue - cat = name.split("-")[1] # c1..c12 - (active if m.status == "active" else quar)[cat] += 1 - assert active, "cat-* categories present" - for cat, n in active.items(): - assert n == 2, f"{cat}: expected 2 active reps, got {n}" - assert quar[cat] >= 1, f"{cat}: redundant ones must be quarantined" - - def test_default_evaluate_excludes_quarantine_keeps_active(monkeypatch): - # Use the real resolver but a tiny fake eval to avoid the engine: - # assert quarantined ids land in `skipped`, active ones in tasks. - from openra_bench import run_eval - - captured = {} + # Real resolver + a tiny fake compile (no engine): quarantined ids + # land in `skipped` and are never compiled; active ones are. + captured: dict = {} def fake_compile(pack, level): captured.setdefault("compiled", []).append(pack.meta.id) class C: - map_supported = False # short-circuit into skipped, no engine + map_supported = False # short-circuit into skipped c = C() c.meta = pack.meta return c monkeypatch.setattr(run_eval, "compile_level", fake_compile) - packs = run_eval._resolve_packs(None) # default bundled set + packs = run_eval._resolve_packs(None) out = run_eval.evaluate(packs=packs, levels=["easy"], seeds=[1]) sk = " ".join(out["skipped"]) assert "quarantine:" in sk - # a known-quarantined pack is in skipped and never compiled - assert any("cat-c1-05" in s for s in out["skipped"]) - assert "cat-c1-05" not in captured.get("compiled", []) - # a kept representative WAS compiled (reached the level loop) - assert "cat-c1-00" in captured.get("compiled", []) + assert any("economy-harvest-timebox" in s for s in out["skipped"]) + assert "economy-harvest-timebox" not in captured.get("compiled", []) + # a curated active pack reached the (faked) compile step + assert "rush-hour" in captured.get("compiled", []) def test_quarantined_pack_still_loads_when_named_explicitly(): - # explicit selection bypasses the default-set filter entirely - p = run_eval_resolve_one("cat-c1-05.yaml") - assert p.meta.id == "cat-c1-05" and p.meta.status == "quarantine" - - -def run_eval_resolve_one(fname: str): - return load_pack(PACKS / fname) + p = load_pack(PACKS / "economy-harvest-timebox.yaml") + assert p.meta.status == "quarantine" and p.meta.id == "economy-harvest-timebox"