yxc20098 commited on
Commit
e9d5097
·
1 Parent(s): 424da31

Curate: remove the auto-generated cat-* corpus

Browse files

Deliberate benchmark-design decision: drop all 68 procedurally-stamped
cat-* packs (low discrimination/rigor, memorization surface, near-dups
per SCENARIO_QUALITY.md) and keep only the 24 hand-authored, curated
decision problems (Art-of-War, economy/building planning, strategy,
rush-hour, adversarial, strict-spec, perception, reasoning). New packs
will be designed one at a time, deliberately.

- delete openra_bench/scenarios/packs/cat-c*.yaml (68 files)
- remove tests/test_catalog.py (tested the retired auto-catalog)
- rewrite tests/test_hygiene.py around the curated set + the
quarantine mechanism (harvest packs as the S0/S1-blocked exemplars)
- SCENARIO_CATALOG.md marked RETIRED (kept as historical notes)
- active set 48 → 24 packs / 72 levels; 273 passed, 1 skipped

This view is limited to 50 files because it contains too many changes.   See raw diff
Files changed (50) hide show
  1. SCENARIO_CATALOG.md +10 -0
  2. openra_bench/scenarios/packs/cat-c1-00.yaml +0 -63
  3. openra_bench/scenarios/packs/cat-c1-01.yaml +0 -63
  4. openra_bench/scenarios/packs/cat-c1-02.yaml +0 -65
  5. openra_bench/scenarios/packs/cat-c1-03.yaml +0 -65
  6. openra_bench/scenarios/packs/cat-c1-04.yaml +0 -65
  7. openra_bench/scenarios/packs/cat-c1-05.yaml +0 -65
  8. openra_bench/scenarios/packs/cat-c10-00.yaml +0 -80
  9. openra_bench/scenarios/packs/cat-c10-01.yaml +0 -80
  10. openra_bench/scenarios/packs/cat-c10-02.yaml +0 -82
  11. openra_bench/scenarios/packs/cat-c10-03.yaml +0 -82
  12. openra_bench/scenarios/packs/cat-c10-04.yaml +0 -82
  13. openra_bench/scenarios/packs/cat-c10-05.yaml +0 -82
  14. openra_bench/scenarios/packs/cat-c11-00.yaml +0 -87
  15. openra_bench/scenarios/packs/cat-c11-01.yaml +0 -87
  16. openra_bench/scenarios/packs/cat-c11-02.yaml +0 -88
  17. openra_bench/scenarios/packs/cat-c11-03.yaml +0 -88
  18. openra_bench/scenarios/packs/cat-c11-04.yaml +0 -88
  19. openra_bench/scenarios/packs/cat-c12-00.yaml +0 -86
  20. openra_bench/scenarios/packs/cat-c12-01.yaml +0 -86
  21. openra_bench/scenarios/packs/cat-c12-02.yaml +0 -87
  22. openra_bench/scenarios/packs/cat-c12-03.yaml +0 -87
  23. openra_bench/scenarios/packs/cat-c12-04.yaml +0 -87
  24. openra_bench/scenarios/packs/cat-c2-00.yaml +0 -94
  25. openra_bench/scenarios/packs/cat-c2-01.yaml +0 -94
  26. openra_bench/scenarios/packs/cat-c2-02.yaml +0 -95
  27. openra_bench/scenarios/packs/cat-c2-03.yaml +0 -95
  28. openra_bench/scenarios/packs/cat-c2-04.yaml +0 -95
  29. openra_bench/scenarios/packs/cat-c2-05.yaml +0 -95
  30. openra_bench/scenarios/packs/cat-c3-00.yaml +0 -80
  31. openra_bench/scenarios/packs/cat-c3-01.yaml +0 -80
  32. openra_bench/scenarios/packs/cat-c3-02.yaml +0 -82
  33. openra_bench/scenarios/packs/cat-c3-03.yaml +0 -82
  34. openra_bench/scenarios/packs/cat-c3-04.yaml +0 -82
  35. openra_bench/scenarios/packs/cat-c3-05.yaml +0 -82
  36. openra_bench/scenarios/packs/cat-c4-00.yaml +0 -77
  37. openra_bench/scenarios/packs/cat-c4-01.yaml +0 -77
  38. openra_bench/scenarios/packs/cat-c4-02.yaml +0 -79
  39. openra_bench/scenarios/packs/cat-c4-03.yaml +0 -79
  40. openra_bench/scenarios/packs/cat-c4-04.yaml +0 -79
  41. openra_bench/scenarios/packs/cat-c5-00.yaml +0 -89
  42. openra_bench/scenarios/packs/cat-c5-01.yaml +0 -89
  43. openra_bench/scenarios/packs/cat-c5-02.yaml +0 -90
  44. openra_bench/scenarios/packs/cat-c5-03.yaml +0 -90
  45. openra_bench/scenarios/packs/cat-c5-04.yaml +0 -90
  46. openra_bench/scenarios/packs/cat-c5-05.yaml +0 -90
  47. openra_bench/scenarios/packs/cat-c6-00.yaml +0 -84
  48. openra_bench/scenarios/packs/cat-c6-01.yaml +0 -84
  49. openra_bench/scenarios/packs/cat-c6-02.yaml +0 -85
  50. openra_bench/scenarios/packs/cat-c6-03.yaml +0 -85
SCENARIO_CATALOG.md CHANGED
@@ -1,3 +1,13 @@
 
 
 
 
 
 
 
 
 
 
1
  # OpenRA-Bench Scenario Catalog (research-grounded, 200 levels)
2
 
3
  Goal: a benchmark whose scenarios **generalize on reasoning** — we
 
1
+ > **RETIRED (curation decision).** The auto-generated 200-level `cat-*`
2
+ > corpus described below has been removed from the repo. It scored low
3
+ > on discrimination/rigor (see `SCENARIO_QUALITY.md`) and added
4
+ > memorization surface without distinct capability. The benchmark is now
5
+ > a small **curated, hand-authored** set (Art-of-War, economy/building
6
+ > planning, strategy, rush-hour, adversarial, strict-spec, perception,
7
+ > reasoning), each designed deliberately one at a time with a
8
+ > non-gameable win+fail and a named external-benchmark analogue. This
9
+ > file is kept only as historical design notes.
10
+
11
  # OpenRA-Bench Scenario Catalog (research-grounded, 200 levels)
12
 
13
  Goal: a benchmark whose scenarios **generalize on reasoning** — we
openra_bench/scenarios/packs/cat-c1-00.yaml DELETED
@@ -1,63 +0,0 @@
1
- meta:
2
- id: cat-c1-00
3
- title: 'Frontier Scouting #1'
4
- capability: perception
5
- real_world_meaning: Pathfinding is solved; deciding WHICH unexplored region to commit
6
- to under partial information and a deadline is the real search problem.
7
- robotics_analogue: 'UAV/UGV frontier selection: choose which unknown region to commit
8
- a time-limited searcher to.'
9
- author: catalog-generator
10
- base_map: rush-hour-arena
11
- base:
12
- agent:
13
- faction: allies
14
- enemy:
15
- faction: soviet
16
- tools:
17
- - move_units
18
- - attack_unit
19
- - stop
20
- planning: true
21
- termination:
22
- max_ticks: 30000
23
- actors:
24
- - type: jeep
25
- owner: agent
26
- position:
27
- - 6
28
- - 6
29
- count: 2
30
- - type: e1
31
- owner: agent
32
- position:
33
- - 7
34
- - 8
35
- count: 2
36
- levels:
37
- easy:
38
- description: "Frontier Scouting: easy \u2014 the decision is harder (less info\
39
- \ / tighter clock / committed choice)."
40
- win_condition:
41
- all_of:
42
- - explored_pct_gte: 38
43
- - within_ticks: 16000
44
- max_turns: 45
45
- medium:
46
- description: "Frontier Scouting: medium \u2014 the decision is harder (less info\
47
- \ / tighter clock / committed choice)."
48
- win_condition:
49
- all_of:
50
- - explored_pct_gte: 47
51
- - within_ticks: 12500
52
- max_turns: 53
53
- hard:
54
- description: "Frontier Scouting: hard \u2014 the decision is harder (less info\
55
- \ / tighter clock / committed choice)."
56
- win_condition:
57
- all_of:
58
- - explored_pct_gte: 56
59
- - units_lost_lte: 0
60
- - within_ticks: 9000
61
- max_turns: 61
62
- fail_condition:
63
- units_lost_lte: -1
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
openra_bench/scenarios/packs/cat-c1-01.yaml DELETED
@@ -1,63 +0,0 @@
1
- meta:
2
- id: cat-c1-01
3
- title: 'Frontier Scouting #2'
4
- capability: perception
5
- real_world_meaning: Pathfinding is solved; deciding WHICH unexplored region to commit
6
- to under partial information and a deadline is the real search problem.
7
- robotics_analogue: 'UAV/UGV frontier selection: choose which unknown region to commit
8
- a time-limited searcher to.'
9
- author: catalog-generator
10
- base_map: rush-hour-arena
11
- base:
12
- agent:
13
- faction: allies
14
- enemy:
15
- faction: soviet
16
- tools:
17
- - move_units
18
- - attack_unit
19
- - stop
20
- planning: true
21
- termination:
22
- max_ticks: 30000
23
- actors:
24
- - type: jeep
25
- owner: agent
26
- position:
27
- - 6
28
- - 6
29
- count: 2
30
- - type: e1
31
- owner: agent
32
- position:
33
- - 7
34
- - 8
35
- count: 2
36
- levels:
37
- easy:
38
- description: "Frontier Scouting: easy \u2014 the decision is harder (less info\
39
- \ / tighter clock / committed choice)."
40
- win_condition:
41
- all_of:
42
- - explored_pct_gte: 39
43
- - within_ticks: 15800
44
- max_turns: 45
45
- medium:
46
- description: "Frontier Scouting: medium \u2014 the decision is harder (less info\
47
- \ / tighter clock / committed choice)."
48
- win_condition:
49
- all_of:
50
- - explored_pct_gte: 48
51
- - within_ticks: 12300
52
- max_turns: 53
53
- hard:
54
- description: "Frontier Scouting: hard \u2014 the decision is harder (less info\
55
- \ / tighter clock / committed choice)."
56
- win_condition:
57
- all_of:
58
- - explored_pct_gte: 57
59
- - units_lost_lte: 0
60
- - within_ticks: 8800
61
- max_turns: 61
62
- fail_condition:
63
- units_lost_lte: -1
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
openra_bench/scenarios/packs/cat-c1-02.yaml DELETED
@@ -1,65 +0,0 @@
1
- meta:
2
- id: cat-c1-02
3
- title: 'Frontier Scouting #3'
4
- capability: perception
5
- real_world_meaning: Pathfinding is solved; deciding WHICH unexplored region to commit to under partial
6
- information and a deadline is the real search problem.
7
- robotics_analogue: 'UAV/UGV frontier selection: choose which unknown region to commit a time-limited
8
- searcher to.'
9
- author: catalog-generator
10
- status: quarantine
11
- quarantine_reason: 'redundant cat-* over-generation (audit: keep 2 representatives/category)'
12
- base_map: rush-hour-arena
13
- base:
14
- agent:
15
- faction: allies
16
- enemy:
17
- faction: soviet
18
- tools:
19
- - move_units
20
- - attack_unit
21
- - stop
22
- planning: true
23
- termination:
24
- max_ticks: 30000
25
- actors:
26
- - type: jeep
27
- owner: agent
28
- position:
29
- - 6
30
- - 6
31
- count: 2
32
- - type: e1
33
- owner: agent
34
- position:
35
- - 7
36
- - 8
37
- count: 2
38
- levels:
39
- easy:
40
- description: 'Frontier Scouting: easy — the decision is harder (less info / tighter clock / committed
41
- choice).'
42
- win_condition:
43
- all_of:
44
- - explored_pct_gte: 40
45
- - within_ticks: 15600
46
- max_turns: 45
47
- medium:
48
- description: 'Frontier Scouting: medium — the decision is harder (less info / tighter clock / committed
49
- choice).'
50
- win_condition:
51
- all_of:
52
- - explored_pct_gte: 49
53
- - within_ticks: 12100
54
- max_turns: 53
55
- hard:
56
- description: 'Frontier Scouting: hard — the decision is harder (less info / tighter clock / committed
57
- choice).'
58
- win_condition:
59
- all_of:
60
- - explored_pct_gte: 58
61
- - units_lost_lte: 0
62
- - within_ticks: 8600
63
- max_turns: 61
64
- fail_condition:
65
- units_lost_lte: -1
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
openra_bench/scenarios/packs/cat-c1-03.yaml DELETED
@@ -1,65 +0,0 @@
1
- meta:
2
- id: cat-c1-03
3
- title: 'Frontier Scouting #4'
4
- capability: perception
5
- real_world_meaning: Pathfinding is solved; deciding WHICH unexplored region to commit to under partial
6
- information and a deadline is the real search problem.
7
- robotics_analogue: 'UAV/UGV frontier selection: choose which unknown region to commit a time-limited
8
- searcher to.'
9
- author: catalog-generator
10
- status: quarantine
11
- quarantine_reason: 'redundant cat-* over-generation (audit: keep 2 representatives/category)'
12
- base_map: rush-hour-arena
13
- base:
14
- agent:
15
- faction: allies
16
- enemy:
17
- faction: soviet
18
- tools:
19
- - move_units
20
- - attack_unit
21
- - stop
22
- planning: true
23
- termination:
24
- max_ticks: 30000
25
- actors:
26
- - type: jeep
27
- owner: agent
28
- position:
29
- - 6
30
- - 6
31
- count: 2
32
- - type: e1
33
- owner: agent
34
- position:
35
- - 7
36
- - 8
37
- count: 2
38
- levels:
39
- easy:
40
- description: 'Frontier Scouting: easy — the decision is harder (less info / tighter clock / committed
41
- choice).'
42
- win_condition:
43
- all_of:
44
- - explored_pct_gte: 41
45
- - within_ticks: 15400
46
- max_turns: 45
47
- medium:
48
- description: 'Frontier Scouting: medium — the decision is harder (less info / tighter clock / committed
49
- choice).'
50
- win_condition:
51
- all_of:
52
- - explored_pct_gte: 50
53
- - within_ticks: 11900
54
- max_turns: 53
55
- hard:
56
- description: 'Frontier Scouting: hard — the decision is harder (less info / tighter clock / committed
57
- choice).'
58
- win_condition:
59
- all_of:
60
- - explored_pct_gte: 59
61
- - units_lost_lte: 0
62
- - within_ticks: 8400
63
- max_turns: 61
64
- fail_condition:
65
- units_lost_lte: -1
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
openra_bench/scenarios/packs/cat-c1-04.yaml DELETED
@@ -1,65 +0,0 @@
1
- meta:
2
- id: cat-c1-04
3
- title: 'Frontier Scouting #5'
4
- capability: perception
5
- real_world_meaning: Pathfinding is solved; deciding WHICH unexplored region to commit to under partial
6
- information and a deadline is the real search problem.
7
- robotics_analogue: 'UAV/UGV frontier selection: choose which unknown region to commit a time-limited
8
- searcher to.'
9
- author: catalog-generator
10
- status: quarantine
11
- quarantine_reason: 'redundant cat-* over-generation (audit: keep 2 representatives/category)'
12
- base_map: rush-hour-arena
13
- base:
14
- agent:
15
- faction: allies
16
- enemy:
17
- faction: soviet
18
- tools:
19
- - move_units
20
- - attack_unit
21
- - stop
22
- planning: true
23
- termination:
24
- max_ticks: 30000
25
- actors:
26
- - type: jeep
27
- owner: agent
28
- position:
29
- - 6
30
- - 6
31
- count: 2
32
- - type: e1
33
- owner: agent
34
- position:
35
- - 7
36
- - 8
37
- count: 2
38
- levels:
39
- easy:
40
- description: 'Frontier Scouting: easy — the decision is harder (less info / tighter clock / committed
41
- choice).'
42
- win_condition:
43
- all_of:
44
- - explored_pct_gte: 42
45
- - within_ticks: 15200
46
- max_turns: 45
47
- medium:
48
- description: 'Frontier Scouting: medium — the decision is harder (less info / tighter clock / committed
49
- choice).'
50
- win_condition:
51
- all_of:
52
- - explored_pct_gte: 51
53
- - within_ticks: 11700
54
- max_turns: 53
55
- hard:
56
- description: 'Frontier Scouting: hard — the decision is harder (less info / tighter clock / committed
57
- choice).'
58
- win_condition:
59
- all_of:
60
- - explored_pct_gte: 60
61
- - units_lost_lte: 0
62
- - within_ticks: 8200
63
- max_turns: 61
64
- fail_condition:
65
- units_lost_lte: -1
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
openra_bench/scenarios/packs/cat-c1-05.yaml DELETED
@@ -1,65 +0,0 @@
1
- meta:
2
- id: cat-c1-05
3
- title: 'Frontier Scouting #6'
4
- capability: perception
5
- real_world_meaning: Pathfinding is solved; deciding WHICH unexplored region to commit to under partial
6
- information and a deadline is the real search problem.
7
- robotics_analogue: 'UAV/UGV frontier selection: choose which unknown region to commit a time-limited
8
- searcher to.'
9
- author: catalog-generator
10
- status: quarantine
11
- quarantine_reason: 'redundant cat-* over-generation (audit: keep 2 representatives/category)'
12
- base_map: rush-hour-arena
13
- base:
14
- agent:
15
- faction: allies
16
- enemy:
17
- faction: soviet
18
- tools:
19
- - move_units
20
- - attack_unit
21
- - stop
22
- planning: true
23
- termination:
24
- max_ticks: 30000
25
- actors:
26
- - type: jeep
27
- owner: agent
28
- position:
29
- - 6
30
- - 6
31
- count: 2
32
- - type: e1
33
- owner: agent
34
- position:
35
- - 7
36
- - 8
37
- count: 2
38
- levels:
39
- easy:
40
- description: 'Frontier Scouting: easy — the decision is harder (less info / tighter clock / committed
41
- choice).'
42
- win_condition:
43
- all_of:
44
- - explored_pct_gte: 43
45
- - within_ticks: 15000
46
- max_turns: 45
47
- medium:
48
- description: 'Frontier Scouting: medium — the decision is harder (less info / tighter clock / committed
49
- choice).'
50
- win_condition:
51
- all_of:
52
- - explored_pct_gte: 52
53
- - within_ticks: 11500
54
- max_turns: 53
55
- hard:
56
- description: 'Frontier Scouting: hard — the decision is harder (less info / tighter clock / committed
57
- choice).'
58
- win_condition:
59
- all_of:
60
- - explored_pct_gte: 61
61
- - units_lost_lte: 0
62
- - within_ticks: 8000
63
- max_turns: 61
64
- fail_condition:
65
- units_lost_lte: -1
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
openra_bench/scenarios/packs/cat-c10-00.yaml DELETED
@@ -1,80 +0,0 @@
1
- meta:
2
- id: cat-c10-00
3
- title: 'Force Coordination #1'
4
- capability: action
5
- real_world_meaning: 'Coordinated multi-robot rendezvous: steer separated teams to
6
- a common staging point on time.'
7
- robotics_analogue: Drive split groups to converge simultaneously and survive.
8
- author: catalog-generator
9
- base_map: rush-hour-arena
10
- base:
11
- agent:
12
- faction: allies
13
- enemy:
14
- faction: soviet
15
- tools:
16
- - move_units
17
- - attack_unit
18
- - stop
19
- planning: true
20
- termination:
21
- max_ticks: 30000
22
- actors:
23
- - type: jeep
24
- owner: agent
25
- position:
26
- - 6
27
- - 6
28
- count: 2
29
- - type: jeep
30
- owner: agent
31
- position:
32
- - 6
33
- - 34
34
- count: 2
35
- - type: e1
36
- owner: enemy
37
- position:
38
- - 110
39
- - 20
40
- count: 2
41
- stance: 2
42
- levels:
43
- easy:
44
- description: "Force Coordination: easy \u2014 the decision is harder (less info\
45
- \ / tighter clock / committed choice)."
46
- win_condition:
47
- all_of:
48
- - all_units_in_region:
49
- x: 55
50
- y: 20
51
- radius: 7
52
- - units_lost_lte: 1
53
- - within_ticks: 13000
54
- max_turns: 45
55
- medium:
56
- description: "Force Coordination: medium \u2014 the decision is harder (less info\
57
- \ / tighter clock / committed choice)."
58
- win_condition:
59
- all_of:
60
- - all_units_in_region:
61
- x: 55
62
- y: 20
63
- radius: 6
64
- - units_lost_lte: 1
65
- - within_ticks: 10000
66
- max_turns: 53
67
- hard:
68
- description: "Force Coordination: hard \u2014 the decision is harder (less info\
69
- \ / tighter clock / committed choice)."
70
- win_condition:
71
- all_of:
72
- - all_units_in_region:
73
- x: 55
74
- y: 20
75
- radius: 5
76
- - units_lost_lte: 0
77
- - within_ticks: 7000
78
- max_turns: 61
79
- fail_condition:
80
- units_lost_lte: -1
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
openra_bench/scenarios/packs/cat-c10-01.yaml DELETED
@@ -1,80 +0,0 @@
1
- meta:
2
- id: cat-c10-01
3
- title: 'Force Coordination #2'
4
- capability: action
5
- real_world_meaning: 'Coordinated multi-robot rendezvous: steer separated teams to
6
- a common staging point on time.'
7
- robotics_analogue: Drive split groups to converge simultaneously and survive.
8
- author: catalog-generator
9
- base_map: rush-hour-arena
10
- base:
11
- agent:
12
- faction: allies
13
- enemy:
14
- faction: soviet
15
- tools:
16
- - move_units
17
- - attack_unit
18
- - stop
19
- planning: true
20
- termination:
21
- max_ticks: 30000
22
- actors:
23
- - type: jeep
24
- owner: agent
25
- position:
26
- - 6
27
- - 6
28
- count: 2
29
- - type: jeep
30
- owner: agent
31
- position:
32
- - 6
33
- - 34
34
- count: 2
35
- - type: e1
36
- owner: enemy
37
- position:
38
- - 110
39
- - 20
40
- count: 2
41
- stance: 2
42
- levels:
43
- easy:
44
- description: "Force Coordination: easy \u2014 the decision is harder (less info\
45
- \ / tighter clock / committed choice)."
46
- win_condition:
47
- all_of:
48
- - all_units_in_region:
49
- x: 58
50
- y: 20
51
- radius: 7
52
- - units_lost_lte: 1
53
- - within_ticks: 13000
54
- max_turns: 45
55
- medium:
56
- description: "Force Coordination: medium \u2014 the decision is harder (less info\
57
- \ / tighter clock / committed choice)."
58
- win_condition:
59
- all_of:
60
- - all_units_in_region:
61
- x: 58
62
- y: 20
63
- radius: 6
64
- - units_lost_lte: 1
65
- - within_ticks: 10000
66
- max_turns: 53
67
- hard:
68
- description: "Force Coordination: hard \u2014 the decision is harder (less info\
69
- \ / tighter clock / committed choice)."
70
- win_condition:
71
- all_of:
72
- - all_units_in_region:
73
- x: 58
74
- y: 20
75
- radius: 5
76
- - units_lost_lte: 0
77
- - within_ticks: 7000
78
- max_turns: 61
79
- fail_condition:
80
- units_lost_lte: -1
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
openra_bench/scenarios/packs/cat-c10-02.yaml DELETED
@@ -1,82 +0,0 @@
1
- meta:
2
- id: cat-c10-02
3
- title: 'Force Coordination #3'
4
- capability: action
5
- real_world_meaning: 'Coordinated multi-robot rendezvous: steer separated teams to a common staging point
6
- on time.'
7
- robotics_analogue: Drive split groups to converge simultaneously and survive.
8
- author: catalog-generator
9
- status: quarantine
10
- quarantine_reason: 'redundant cat-* over-generation (audit: keep 2 representatives/category)'
11
- base_map: rush-hour-arena
12
- base:
13
- agent:
14
- faction: allies
15
- enemy:
16
- faction: soviet
17
- tools:
18
- - move_units
19
- - attack_unit
20
- - stop
21
- planning: true
22
- termination:
23
- max_ticks: 30000
24
- actors:
25
- - type: jeep
26
- owner: agent
27
- position:
28
- - 6
29
- - 6
30
- count: 2
31
- - type: jeep
32
- owner: agent
33
- position:
34
- - 6
35
- - 34
36
- count: 2
37
- - type: e1
38
- owner: enemy
39
- position:
40
- - 110
41
- - 20
42
- count: 2
43
- stance: 2
44
- levels:
45
- easy:
46
- description: 'Force Coordination: easy — the decision is harder (less info / tighter clock / committed
47
- choice).'
48
- win_condition:
49
- all_of:
50
- - all_units_in_region:
51
- x: 61
52
- y: 20
53
- radius: 7
54
- - units_lost_lte: 1
55
- - within_ticks: 13000
56
- max_turns: 45
57
- medium:
58
- description: 'Force Coordination: medium — the decision is harder (less info / tighter clock / committed
59
- choice).'
60
- win_condition:
61
- all_of:
62
- - all_units_in_region:
63
- x: 61
64
- y: 20
65
- radius: 6
66
- - units_lost_lte: 1
67
- - within_ticks: 10000
68
- max_turns: 53
69
- hard:
70
- description: 'Force Coordination: hard — the decision is harder (less info / tighter clock / committed
71
- choice).'
72
- win_condition:
73
- all_of:
74
- - all_units_in_region:
75
- x: 61
76
- y: 20
77
- radius: 5
78
- - units_lost_lte: 0
79
- - within_ticks: 7000
80
- max_turns: 61
81
- fail_condition:
82
- units_lost_lte: -1
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
openra_bench/scenarios/packs/cat-c10-03.yaml DELETED
@@ -1,82 +0,0 @@
1
- meta:
2
- id: cat-c10-03
3
- title: 'Force Coordination #4'
4
- capability: action
5
- real_world_meaning: 'Coordinated multi-robot rendezvous: steer separated teams to a common staging point
6
- on time.'
7
- robotics_analogue: Drive split groups to converge simultaneously and survive.
8
- author: catalog-generator
9
- status: quarantine
10
- quarantine_reason: 'redundant cat-* over-generation (audit: keep 2 representatives/category)'
11
- base_map: rush-hour-arena
12
- base:
13
- agent:
14
- faction: allies
15
- enemy:
16
- faction: soviet
17
- tools:
18
- - move_units
19
- - attack_unit
20
- - stop
21
- planning: true
22
- termination:
23
- max_ticks: 30000
24
- actors:
25
- - type: jeep
26
- owner: agent
27
- position:
28
- - 6
29
- - 6
30
- count: 2
31
- - type: jeep
32
- owner: agent
33
- position:
34
- - 6
35
- - 34
36
- count: 2
37
- - type: e1
38
- owner: enemy
39
- position:
40
- - 110
41
- - 20
42
- count: 2
43
- stance: 2
44
- levels:
45
- easy:
46
- description: 'Force Coordination: easy — the decision is harder (less info / tighter clock / committed
47
- choice).'
48
- win_condition:
49
- all_of:
50
- - all_units_in_region:
51
- x: 64
52
- y: 20
53
- radius: 7
54
- - units_lost_lte: 1
55
- - within_ticks: 13000
56
- max_turns: 45
57
- medium:
58
- description: 'Force Coordination: medium — the decision is harder (less info / tighter clock / committed
59
- choice).'
60
- win_condition:
61
- all_of:
62
- - all_units_in_region:
63
- x: 64
64
- y: 20
65
- radius: 6
66
- - units_lost_lte: 1
67
- - within_ticks: 10000
68
- max_turns: 53
69
- hard:
70
- description: 'Force Coordination: hard — the decision is harder (less info / tighter clock / committed
71
- choice).'
72
- win_condition:
73
- all_of:
74
- - all_units_in_region:
75
- x: 64
76
- y: 20
77
- radius: 5
78
- - units_lost_lte: 0
79
- - within_ticks: 7000
80
- max_turns: 61
81
- fail_condition:
82
- units_lost_lte: -1
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
openra_bench/scenarios/packs/cat-c10-04.yaml DELETED
@@ -1,82 +0,0 @@
1
- meta:
2
- id: cat-c10-04
3
- title: 'Force Coordination #5'
4
- capability: action
5
- real_world_meaning: 'Coordinated multi-robot rendezvous: steer separated teams to a common staging point
6
- on time.'
7
- robotics_analogue: Drive split groups to converge simultaneously and survive.
8
- author: catalog-generator
9
- status: quarantine
10
- quarantine_reason: 'redundant cat-* over-generation (audit: keep 2 representatives/category)'
11
- base_map: rush-hour-arena
12
- base:
13
- agent:
14
- faction: allies
15
- enemy:
16
- faction: soviet
17
- tools:
18
- - move_units
19
- - attack_unit
20
- - stop
21
- planning: true
22
- termination:
23
- max_ticks: 30000
24
- actors:
25
- - type: jeep
26
- owner: agent
27
- position:
28
- - 6
29
- - 6
30
- count: 2
31
- - type: jeep
32
- owner: agent
33
- position:
34
- - 6
35
- - 34
36
- count: 2
37
- - type: e1
38
- owner: enemy
39
- position:
40
- - 110
41
- - 20
42
- count: 2
43
- stance: 2
44
- levels:
45
- easy:
46
- description: 'Force Coordination: easy — the decision is harder (less info / tighter clock / committed
47
- choice).'
48
- win_condition:
49
- all_of:
50
- - all_units_in_region:
51
- x: 67
52
- y: 20
53
- radius: 7
54
- - units_lost_lte: 1
55
- - within_ticks: 13000
56
- max_turns: 45
57
- medium:
58
- description: 'Force Coordination: medium — the decision is harder (less info / tighter clock / committed
59
- choice).'
60
- win_condition:
61
- all_of:
62
- - all_units_in_region:
63
- x: 67
64
- y: 20
65
- radius: 6
66
- - units_lost_lte: 1
67
- - within_ticks: 10000
68
- max_turns: 53
69
- hard:
70
- description: 'Force Coordination: hard — the decision is harder (less info / tighter clock / committed
71
- choice).'
72
- win_condition:
73
- all_of:
74
- - all_units_in_region:
75
- x: 67
76
- y: 20
77
- radius: 5
78
- - units_lost_lte: 0
79
- - within_ticks: 7000
80
- max_turns: 61
81
- fail_condition:
82
- units_lost_lte: -1
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
openra_bench/scenarios/packs/cat-c10-05.yaml DELETED
@@ -1,82 +0,0 @@
1
- meta:
2
- id: cat-c10-05
3
- title: 'Force Coordination #6'
4
- capability: action
5
- real_world_meaning: 'Coordinated multi-robot rendezvous: steer separated teams to a common staging point
6
- on time.'
7
- robotics_analogue: Drive split groups to converge simultaneously and survive.
8
- author: catalog-generator
9
- status: quarantine
10
- quarantine_reason: 'redundant cat-* over-generation (audit: keep 2 representatives/category)'
11
- base_map: rush-hour-arena
12
- base:
13
- agent:
14
- faction: allies
15
- enemy:
16
- faction: soviet
17
- tools:
18
- - move_units
19
- - attack_unit
20
- - stop
21
- planning: true
22
- termination:
23
- max_ticks: 30000
24
- actors:
25
- - type: jeep
26
- owner: agent
27
- position:
28
- - 6
29
- - 6
30
- count: 2
31
- - type: jeep
32
- owner: agent
33
- position:
34
- - 6
35
- - 34
36
- count: 2
37
- - type: e1
38
- owner: enemy
39
- position:
40
- - 110
41
- - 20
42
- count: 2
43
- stance: 2
44
- levels:
45
- easy:
46
- description: 'Force Coordination: easy — the decision is harder (less info / tighter clock / committed
47
- choice).'
48
- win_condition:
49
- all_of:
50
- - all_units_in_region:
51
- x: 70
52
- y: 20
53
- radius: 7
54
- - units_lost_lte: 1
55
- - within_ticks: 13000
56
- max_turns: 45
57
- medium:
58
- description: 'Force Coordination: medium — the decision is harder (less info / tighter clock / committed
59
- choice).'
60
- win_condition:
61
- all_of:
62
- - all_units_in_region:
63
- x: 70
64
- y: 20
65
- radius: 6
66
- - units_lost_lte: 1
67
- - within_ticks: 10000
68
- max_turns: 53
69
- hard:
70
- description: 'Force Coordination: hard — the decision is harder (less info / tighter clock / committed
71
- choice).'
72
- win_condition:
73
- all_of:
74
- - all_units_in_region:
75
- x: 70
76
- y: 20
77
- radius: 5
78
- - units_lost_lte: 0
79
- - within_ticks: 7000
80
- max_turns: 61
81
- fail_condition:
82
- units_lost_lte: -1
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
openra_bench/scenarios/packs/cat-c11-00.yaml DELETED
@@ -1,87 +0,0 @@
1
- meta:
2
- id: cat-c11-00
3
- title: 'Tempo / Timing Window #1'
4
- capability: reasoning
5
- real_world_meaning: 'Acting only inside a valid time window (takt-time): too early
6
- or too late both fail.'
7
- robotics_analogue: "Strike inside a window that opens late and closes \u2014 not\
8
- \ earliest, not biggest."
9
- author: catalog-generator
10
- base_map: rush-hour-arena
11
- base:
12
- agent:
13
- faction: allies
14
- enemy:
15
- faction: soviet
16
- tools:
17
- - move_units
18
- - attack_unit
19
- - attack_move
20
- - stop
21
- planning: true
22
- termination:
23
- max_ticks: 30000
24
- actors:
25
- - type: 2tnk
26
- owner: agent
27
- position:
28
- - 8
29
- - 18
30
- count: 2
31
- - type: e1
32
- owner: agent
33
- position:
34
- - 9
35
- - 20
36
- count: 3
37
- - type: e1
38
- owner: enemy
39
- position:
40
- - 100
41
- - 20
42
- count: 2
43
- stance: 2
44
- - type: e3
45
- owner: enemy
46
- position:
47
- - 104
48
- - 24
49
- count: 1
50
- stance: 2
51
- - type: e1
52
- owner: enemy
53
- position:
54
- - 40
55
- - 10
56
- count: 1
57
- stance: 2
58
- levels:
59
- easy:
60
- description: "Tempo / Timing Window: easy \u2014 the decision is harder (less\
61
- \ info / tighter clock / committed choice)."
62
- win_condition:
63
- all_of:
64
- - after_ticks: 1500
65
- - units_killed_gte: 2
66
- - within_ticks: 12000
67
- max_turns: 45
68
- medium:
69
- description: "Tempo / Timing Window: medium \u2014 the decision is harder (less\
70
- \ info / tighter clock / committed choice)."
71
- win_condition:
72
- all_of:
73
- - after_ticks: 2300
74
- - units_killed_gte: 3
75
- - within_ticks: 9500
76
- max_turns: 53
77
- hard:
78
- description: "Tempo / Timing Window: hard \u2014 the decision is harder (less\
79
- \ info / tighter clock / committed choice)."
80
- win_condition:
81
- all_of:
82
- - after_ticks: 3100
83
- - units_killed_gte: 4
84
- - within_ticks: 7000
85
- max_turns: 61
86
- fail_condition:
87
- units_lost_lte: -1
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
openra_bench/scenarios/packs/cat-c11-01.yaml DELETED
@@ -1,87 +0,0 @@
1
- meta:
2
- id: cat-c11-01
3
- title: 'Tempo / Timing Window #2'
4
- capability: reasoning
5
- real_world_meaning: 'Acting only inside a valid time window (takt-time): too early
6
- or too late both fail.'
7
- robotics_analogue: "Strike inside a window that opens late and closes \u2014 not\
8
- \ earliest, not biggest."
9
- author: catalog-generator
10
- base_map: rush-hour-arena
11
- base:
12
- agent:
13
- faction: allies
14
- enemy:
15
- faction: soviet
16
- tools:
17
- - move_units
18
- - attack_unit
19
- - attack_move
20
- - stop
21
- planning: true
22
- termination:
23
- max_ticks: 30000
24
- actors:
25
- - type: 2tnk
26
- owner: agent
27
- position:
28
- - 8
29
- - 18
30
- count: 2
31
- - type: e1
32
- owner: agent
33
- position:
34
- - 9
35
- - 20
36
- count: 3
37
- - type: e1
38
- owner: enemy
39
- position:
40
- - 100
41
- - 20
42
- count: 2
43
- stance: 2
44
- - type: e3
45
- owner: enemy
46
- position:
47
- - 104
48
- - 24
49
- count: 1
50
- stance: 2
51
- - type: e1
52
- owner: enemy
53
- position:
54
- - 40
55
- - 10
56
- count: 1
57
- stance: 2
58
- levels:
59
- easy:
60
- description: "Tempo / Timing Window: easy \u2014 the decision is harder (less\
61
- \ info / tighter clock / committed choice)."
62
- win_condition:
63
- all_of:
64
- - after_ticks: 1500
65
- - units_killed_gte: 2
66
- - within_ticks: 12000
67
- max_turns: 45
68
- medium:
69
- description: "Tempo / Timing Window: medium \u2014 the decision is harder (less\
70
- \ info / tighter clock / committed choice)."
71
- win_condition:
72
- all_of:
73
- - after_ticks: 2300
74
- - units_killed_gte: 3
75
- - within_ticks: 9500
76
- max_turns: 53
77
- hard:
78
- description: "Tempo / Timing Window: hard \u2014 the decision is harder (less\
79
- \ info / tighter clock / committed choice)."
80
- win_condition:
81
- all_of:
82
- - after_ticks: 3100
83
- - units_killed_gte: 4
84
- - within_ticks: 7000
85
- max_turns: 61
86
- fail_condition:
87
- units_lost_lte: -1
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
openra_bench/scenarios/packs/cat-c11-02.yaml DELETED
@@ -1,88 +0,0 @@
1
- meta:
2
- id: cat-c11-02
3
- title: 'Tempo / Timing Window #3'
4
- capability: reasoning
5
- real_world_meaning: 'Acting only inside a valid time window (takt-time): too early or too late both
6
- fail.'
7
- robotics_analogue: Strike inside a window that opens late and closes — not earliest, not biggest.
8
- author: catalog-generator
9
- status: quarantine
10
- quarantine_reason: 'redundant cat-* over-generation (audit: keep 2 representatives/category)'
11
- base_map: rush-hour-arena
12
- base:
13
- agent:
14
- faction: allies
15
- enemy:
16
- faction: soviet
17
- tools:
18
- - move_units
19
- - attack_unit
20
- - attack_move
21
- - stop
22
- planning: true
23
- termination:
24
- max_ticks: 30000
25
- actors:
26
- - type: 2tnk
27
- owner: agent
28
- position:
29
- - 8
30
- - 18
31
- count: 2
32
- - type: e1
33
- owner: agent
34
- position:
35
- - 9
36
- - 20
37
- count: 3
38
- - type: e1
39
- owner: enemy
40
- position:
41
- - 100
42
- - 20
43
- count: 2
44
- stance: 2
45
- - type: e3
46
- owner: enemy
47
- position:
48
- - 104
49
- - 24
50
- count: 1
51
- stance: 2
52
- - type: e1
53
- owner: enemy
54
- position:
55
- - 40
56
- - 10
57
- count: 1
58
- stance: 2
59
- levels:
60
- easy:
61
- description: 'Tempo / Timing Window: easy — the decision is harder (less info / tighter clock / committed
62
- choice).'
63
- win_condition:
64
- all_of:
65
- - after_ticks: 1500
66
- - units_killed_gte: 2
67
- - within_ticks: 12000
68
- max_turns: 45
69
- medium:
70
- description: 'Tempo / Timing Window: medium — the decision is harder (less info / tighter clock /
71
- committed choice).'
72
- win_condition:
73
- all_of:
74
- - after_ticks: 2300
75
- - units_killed_gte: 3
76
- - within_ticks: 9500
77
- max_turns: 53
78
- hard:
79
- description: 'Tempo / Timing Window: hard — the decision is harder (less info / tighter clock / committed
80
- choice).'
81
- win_condition:
82
- all_of:
83
- - after_ticks: 3100
84
- - units_killed_gte: 4
85
- - within_ticks: 7000
86
- max_turns: 61
87
- fail_condition:
88
- units_lost_lte: -1
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
openra_bench/scenarios/packs/cat-c11-03.yaml DELETED
@@ -1,88 +0,0 @@
1
- meta:
2
- id: cat-c11-03
3
- title: 'Tempo / Timing Window #4'
4
- capability: reasoning
5
- real_world_meaning: 'Acting only inside a valid time window (takt-time): too early or too late both
6
- fail.'
7
- robotics_analogue: Strike inside a window that opens late and closes — not earliest, not biggest.
8
- author: catalog-generator
9
- status: quarantine
10
- quarantine_reason: 'redundant cat-* over-generation (audit: keep 2 representatives/category)'
11
- base_map: rush-hour-arena
12
- base:
13
- agent:
14
- faction: allies
15
- enemy:
16
- faction: soviet
17
- tools:
18
- - move_units
19
- - attack_unit
20
- - attack_move
21
- - stop
22
- planning: true
23
- termination:
24
- max_ticks: 30000
25
- actors:
26
- - type: 2tnk
27
- owner: agent
28
- position:
29
- - 8
30
- - 18
31
- count: 2
32
- - type: e1
33
- owner: agent
34
- position:
35
- - 9
36
- - 20
37
- count: 3
38
- - type: e1
39
- owner: enemy
40
- position:
41
- - 100
42
- - 20
43
- count: 2
44
- stance: 2
45
- - type: e3
46
- owner: enemy
47
- position:
48
- - 104
49
- - 24
50
- count: 1
51
- stance: 2
52
- - type: e1
53
- owner: enemy
54
- position:
55
- - 40
56
- - 10
57
- count: 1
58
- stance: 2
59
- levels:
60
- easy:
61
- description: 'Tempo / Timing Window: easy — the decision is harder (less info / tighter clock / committed
62
- choice).'
63
- win_condition:
64
- all_of:
65
- - after_ticks: 1500
66
- - units_killed_gte: 2
67
- - within_ticks: 12000
68
- max_turns: 45
69
- medium:
70
- description: 'Tempo / Timing Window: medium — the decision is harder (less info / tighter clock /
71
- committed choice).'
72
- win_condition:
73
- all_of:
74
- - after_ticks: 2300
75
- - units_killed_gte: 3
76
- - within_ticks: 9500
77
- max_turns: 53
78
- hard:
79
- description: 'Tempo / Timing Window: hard — the decision is harder (less info / tighter clock / committed
80
- choice).'
81
- win_condition:
82
- all_of:
83
- - after_ticks: 3100
84
- - units_killed_gte: 4
85
- - within_ticks: 7000
86
- max_turns: 61
87
- fail_condition:
88
- units_lost_lte: -1
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
openra_bench/scenarios/packs/cat-c11-04.yaml DELETED
@@ -1,88 +0,0 @@
1
- meta:
2
- id: cat-c11-04
3
- title: 'Tempo / Timing Window #5'
4
- capability: reasoning
5
- real_world_meaning: 'Acting only inside a valid time window (takt-time): too early or too late both
6
- fail.'
7
- robotics_analogue: Strike inside a window that opens late and closes — not earliest, not biggest.
8
- author: catalog-generator
9
- status: quarantine
10
- quarantine_reason: 'redundant cat-* over-generation (audit: keep 2 representatives/category)'
11
- base_map: rush-hour-arena
12
- base:
13
- agent:
14
- faction: allies
15
- enemy:
16
- faction: soviet
17
- tools:
18
- - move_units
19
- - attack_unit
20
- - attack_move
21
- - stop
22
- planning: true
23
- termination:
24
- max_ticks: 30000
25
- actors:
26
- - type: 2tnk
27
- owner: agent
28
- position:
29
- - 8
30
- - 18
31
- count: 2
32
- - type: e1
33
- owner: agent
34
- position:
35
- - 9
36
- - 20
37
- count: 3
38
- - type: e1
39
- owner: enemy
40
- position:
41
- - 100
42
- - 20
43
- count: 2
44
- stance: 2
45
- - type: e3
46
- owner: enemy
47
- position:
48
- - 104
49
- - 24
50
- count: 1
51
- stance: 2
52
- - type: e1
53
- owner: enemy
54
- position:
55
- - 40
56
- - 10
57
- count: 1
58
- stance: 2
59
- levels:
60
- easy:
61
- description: 'Tempo / Timing Window: easy — the decision is harder (less info / tighter clock / committed
62
- choice).'
63
- win_condition:
64
- all_of:
65
- - after_ticks: 1500
66
- - units_killed_gte: 2
67
- - within_ticks: 12000
68
- max_turns: 45
69
- medium:
70
- description: 'Tempo / Timing Window: medium — the decision is harder (less info / tighter clock /
71
- committed choice).'
72
- win_condition:
73
- all_of:
74
- - after_ticks: 2300
75
- - units_killed_gte: 3
76
- - within_ticks: 9500
77
- max_turns: 53
78
- hard:
79
- description: 'Tempo / Timing Window: hard — the decision is harder (less info / tighter clock / committed
80
- choice).'
81
- win_condition:
82
- all_of:
83
- - after_ticks: 3100
84
- - units_killed_gte: 4
85
- - within_ticks: 7000
86
- max_turns: 61
87
- fail_condition:
88
- units_lost_lte: -1
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
openra_bench/scenarios/packs/cat-c12-00.yaml DELETED
@@ -1,86 +0,0 @@
1
- meta:
2
- id: cat-c12-00
3
- title: 'Error Recovery / Replan #1'
4
- capability: reasoning
5
- real_world_meaning: "Detecting plan failure and repairing it under reduced means\
6
- \ and a deadline \u2014 receding-horizon replanning."
7
- robotics_analogue: After a setback, rebuild/repair and re-establish capability under
8
- a hard clock.
9
- author: catalog-generator
10
- base_map: rush-hour-arena
11
- base:
12
- agent:
13
- faction: allies
14
- enemy:
15
- faction: soviet
16
- tools:
17
- - build
18
- - place_building
19
- - move_units
20
- - deploy
21
- - stop
22
- planning: true
23
- termination:
24
- max_ticks: 40000
25
- actors:
26
- - type: fact
27
- owner: agent
28
- position:
29
- - 10
30
- - 18
31
- - type: powr
32
- owner: agent
33
- position:
34
- - 14
35
- - 18
36
- - type: tent
37
- owner: agent
38
- position:
39
- - 12
40
- - 21
41
- - type: jeep
42
- owner: agent
43
- position:
44
- - 8
45
- - 16
46
- - type: e1
47
- owner: enemy
48
- position:
49
- - 114
50
- - 34
51
- count: 1
52
- stance: 2
53
- levels:
54
- easy:
55
- description: "Error Recovery / Replan: easy \u2014 the decision is harder (less\
56
- \ info / tighter clock / committed choice)."
57
- win_condition:
58
- all_of:
59
- - has_building: tent
60
- - building_total_gte: 4
61
- - after_ticks: 1200
62
- - within_ticks: 22000
63
- max_turns: 45
64
- medium:
65
- description: "Error Recovery / Replan: medium \u2014 the decision is harder (less\
66
- \ info / tighter clock / committed choice)."
67
- win_condition:
68
- all_of:
69
- - has_building: tent
70
- - building_total_gte: 5
71
- - after_ticks: 1200
72
- - within_ticks: 17000
73
- max_turns: 53
74
- hard:
75
- description: "Error Recovery / Replan: hard \u2014 the decision is harder (less\
76
- \ info / tighter clock / committed choice)."
77
- win_condition:
78
- all_of:
79
- - has_building: tent
80
- - building_total_gte: 6
81
- - after_ticks: 1200
82
- - within_ticks: 12000
83
- max_turns: 61
84
- fail_condition:
85
- units_lost_lte: -1
86
- starting_cash: 3000
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
openra_bench/scenarios/packs/cat-c12-01.yaml DELETED
@@ -1,86 +0,0 @@
1
- meta:
2
- id: cat-c12-01
3
- title: 'Error Recovery / Replan #2'
4
- capability: reasoning
5
- real_world_meaning: "Detecting plan failure and repairing it under reduced means\
6
- \ and a deadline \u2014 receding-horizon replanning."
7
- robotics_analogue: After a setback, rebuild/repair and re-establish capability under
8
- a hard clock.
9
- author: catalog-generator
10
- base_map: rush-hour-arena
11
- base:
12
- agent:
13
- faction: allies
14
- enemy:
15
- faction: soviet
16
- tools:
17
- - build
18
- - place_building
19
- - move_units
20
- - deploy
21
- - stop
22
- planning: true
23
- termination:
24
- max_ticks: 40000
25
- actors:
26
- - type: fact
27
- owner: agent
28
- position:
29
- - 10
30
- - 18
31
- - type: powr
32
- owner: agent
33
- position:
34
- - 14
35
- - 18
36
- - type: tent
37
- owner: agent
38
- position:
39
- - 12
40
- - 21
41
- - type: jeep
42
- owner: agent
43
- position:
44
- - 8
45
- - 16
46
- - type: e1
47
- owner: enemy
48
- position:
49
- - 114
50
- - 34
51
- count: 1
52
- stance: 2
53
- levels:
54
- easy:
55
- description: "Error Recovery / Replan: easy \u2014 the decision is harder (less\
56
- \ info / tighter clock / committed choice)."
57
- win_condition:
58
- all_of:
59
- - has_building: tent
60
- - building_total_gte: 4
61
- - after_ticks: 1200
62
- - within_ticks: 22000
63
- max_turns: 45
64
- medium:
65
- description: "Error Recovery / Replan: medium \u2014 the decision is harder (less\
66
- \ info / tighter clock / committed choice)."
67
- win_condition:
68
- all_of:
69
- - has_building: tent
70
- - building_total_gte: 5
71
- - after_ticks: 1200
72
- - within_ticks: 17000
73
- max_turns: 53
74
- hard:
75
- description: "Error Recovery / Replan: hard \u2014 the decision is harder (less\
76
- \ info / tighter clock / committed choice)."
77
- win_condition:
78
- all_of:
79
- - has_building: tent
80
- - building_total_gte: 6
81
- - after_ticks: 1200
82
- - within_ticks: 12000
83
- max_turns: 61
84
- fail_condition:
85
- units_lost_lte: -1
86
- starting_cash: 3000
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
openra_bench/scenarios/packs/cat-c12-02.yaml DELETED
@@ -1,87 +0,0 @@
1
- meta:
2
- id: cat-c12-02
3
- title: 'Error Recovery / Replan #3'
4
- capability: reasoning
5
- real_world_meaning: Detecting plan failure and repairing it under reduced means and a deadline — receding-horizon
6
- replanning.
7
- robotics_analogue: After a setback, rebuild/repair and re-establish capability under a hard clock.
8
- author: catalog-generator
9
- status: quarantine
10
- quarantine_reason: 'redundant cat-* over-generation (audit: keep 2 representatives/category)'
11
- base_map: rush-hour-arena
12
- base:
13
- agent:
14
- faction: allies
15
- enemy:
16
- faction: soviet
17
- tools:
18
- - build
19
- - place_building
20
- - move_units
21
- - deploy
22
- - stop
23
- planning: true
24
- termination:
25
- max_ticks: 40000
26
- actors:
27
- - type: fact
28
- owner: agent
29
- position:
30
- - 10
31
- - 18
32
- - type: powr
33
- owner: agent
34
- position:
35
- - 14
36
- - 18
37
- - type: tent
38
- owner: agent
39
- position:
40
- - 12
41
- - 21
42
- - type: jeep
43
- owner: agent
44
- position:
45
- - 8
46
- - 16
47
- - type: e1
48
- owner: enemy
49
- position:
50
- - 114
51
- - 34
52
- count: 1
53
- stance: 2
54
- levels:
55
- easy:
56
- description: 'Error Recovery / Replan: easy — the decision is harder (less info / tighter clock /
57
- committed choice).'
58
- win_condition:
59
- all_of:
60
- - has_building: tent
61
- - building_total_gte: 4
62
- - after_ticks: 1200
63
- - within_ticks: 22000
64
- max_turns: 45
65
- medium:
66
- description: 'Error Recovery / Replan: medium — the decision is harder (less info / tighter clock
67
- / committed choice).'
68
- win_condition:
69
- all_of:
70
- - has_building: tent
71
- - building_total_gte: 5
72
- - after_ticks: 1200
73
- - within_ticks: 17000
74
- max_turns: 53
75
- hard:
76
- description: 'Error Recovery / Replan: hard — the decision is harder (less info / tighter clock /
77
- committed choice).'
78
- win_condition:
79
- all_of:
80
- - has_building: tent
81
- - building_total_gte: 6
82
- - after_ticks: 1200
83
- - within_ticks: 12000
84
- max_turns: 61
85
- fail_condition:
86
- units_lost_lte: -1
87
- starting_cash: 3000
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
openra_bench/scenarios/packs/cat-c12-03.yaml DELETED
@@ -1,87 +0,0 @@
1
- meta:
2
- id: cat-c12-03
3
- title: 'Error Recovery / Replan #4'
4
- capability: reasoning
5
- real_world_meaning: Detecting plan failure and repairing it under reduced means and a deadline — receding-horizon
6
- replanning.
7
- robotics_analogue: After a setback, rebuild/repair and re-establish capability under a hard clock.
8
- author: catalog-generator
9
- status: quarantine
10
- quarantine_reason: 'redundant cat-* over-generation (audit: keep 2 representatives/category)'
11
- base_map: rush-hour-arena
12
- base:
13
- agent:
14
- faction: allies
15
- enemy:
16
- faction: soviet
17
- tools:
18
- - build
19
- - place_building
20
- - move_units
21
- - deploy
22
- - stop
23
- planning: true
24
- termination:
25
- max_ticks: 40000
26
- actors:
27
- - type: fact
28
- owner: agent
29
- position:
30
- - 10
31
- - 18
32
- - type: powr
33
- owner: agent
34
- position:
35
- - 14
36
- - 18
37
- - type: tent
38
- owner: agent
39
- position:
40
- - 12
41
- - 21
42
- - type: jeep
43
- owner: agent
44
- position:
45
- - 8
46
- - 16
47
- - type: e1
48
- owner: enemy
49
- position:
50
- - 114
51
- - 34
52
- count: 1
53
- stance: 2
54
- levels:
55
- easy:
56
- description: 'Error Recovery / Replan: easy — the decision is harder (less info / tighter clock /
57
- committed choice).'
58
- win_condition:
59
- all_of:
60
- - has_building: tent
61
- - building_total_gte: 4
62
- - after_ticks: 1200
63
- - within_ticks: 22000
64
- max_turns: 45
65
- medium:
66
- description: 'Error Recovery / Replan: medium — the decision is harder (less info / tighter clock
67
- / committed choice).'
68
- win_condition:
69
- all_of:
70
- - has_building: tent
71
- - building_total_gte: 5
72
- - after_ticks: 1200
73
- - within_ticks: 17000
74
- max_turns: 53
75
- hard:
76
- description: 'Error Recovery / Replan: hard — the decision is harder (less info / tighter clock /
77
- committed choice).'
78
- win_condition:
79
- all_of:
80
- - has_building: tent
81
- - building_total_gte: 6
82
- - after_ticks: 1200
83
- - within_ticks: 12000
84
- max_turns: 61
85
- fail_condition:
86
- units_lost_lte: -1
87
- starting_cash: 3000
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
openra_bench/scenarios/packs/cat-c12-04.yaml DELETED
@@ -1,87 +0,0 @@
1
- meta:
2
- id: cat-c12-04
3
- title: 'Error Recovery / Replan #5'
4
- capability: reasoning
5
- real_world_meaning: Detecting plan failure and repairing it under reduced means and a deadline — receding-horizon
6
- replanning.
7
- robotics_analogue: After a setback, rebuild/repair and re-establish capability under a hard clock.
8
- author: catalog-generator
9
- status: quarantine
10
- quarantine_reason: 'redundant cat-* over-generation (audit: keep 2 representatives/category)'
11
- base_map: rush-hour-arena
12
- base:
13
- agent:
14
- faction: allies
15
- enemy:
16
- faction: soviet
17
- tools:
18
- - build
19
- - place_building
20
- - move_units
21
- - deploy
22
- - stop
23
- planning: true
24
- termination:
25
- max_ticks: 40000
26
- actors:
27
- - type: fact
28
- owner: agent
29
- position:
30
- - 10
31
- - 18
32
- - type: powr
33
- owner: agent
34
- position:
35
- - 14
36
- - 18
37
- - type: tent
38
- owner: agent
39
- position:
40
- - 12
41
- - 21
42
- - type: jeep
43
- owner: agent
44
- position:
45
- - 8
46
- - 16
47
- - type: e1
48
- owner: enemy
49
- position:
50
- - 114
51
- - 34
52
- count: 1
53
- stance: 2
54
- levels:
55
- easy:
56
- description: 'Error Recovery / Replan: easy — the decision is harder (less info / tighter clock /
57
- committed choice).'
58
- win_condition:
59
- all_of:
60
- - has_building: tent
61
- - building_total_gte: 4
62
- - after_ticks: 1200
63
- - within_ticks: 22000
64
- max_turns: 45
65
- medium:
66
- description: 'Error Recovery / Replan: medium — the decision is harder (less info / tighter clock
67
- / committed choice).'
68
- win_condition:
69
- all_of:
70
- - has_building: tent
71
- - building_total_gte: 5
72
- - after_ticks: 1200
73
- - within_ticks: 17000
74
- max_turns: 53
75
- hard:
76
- description: 'Error Recovery / Replan: hard — the decision is harder (less info / tighter clock /
77
- committed choice).'
78
- win_condition:
79
- all_of:
80
- - has_building: tent
81
- - building_total_gte: 6
82
- - after_ticks: 1200
83
- - within_ticks: 12000
84
- max_turns: 61
85
- fail_condition:
86
- units_lost_lte: -1
87
- starting_cash: 3000
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
openra_bench/scenarios/packs/cat-c2-00.yaml DELETED
@@ -1,94 +0,0 @@
1
- meta:
2
- id: cat-c2-00
3
- title: 'Threat Enumeration #1'
4
- capability: perception
5
- real_world_meaning: "Estimating hidden adversary state from partial sightings before\
6
- \ acting \u2014 perception under occlusion and distractors."
7
- robotics_analogue: Scene state-estimation under fog and decoys before committing
8
- force.
9
- author: catalog-generator
10
- base_map: rush-hour-arena
11
- base:
12
- agent:
13
- faction: allies
14
- enemy:
15
- faction: soviet
16
- tools:
17
- - move_units
18
- - attack_unit
19
- - attack_move
20
- - stop
21
- planning: true
22
- termination:
23
- max_ticks: 30000
24
- actors:
25
- - type: 2tnk
26
- owner: agent
27
- position:
28
- - 8
29
- - 18
30
- count: 2
31
- - type: e1
32
- owner: agent
33
- position:
34
- - 9
35
- - 20
36
- count: 3
37
- - type: e1
38
- owner: enemy
39
- position:
40
- - 100
41
- - 20
42
- count: 2
43
- stance: 2
44
- - type: e3
45
- owner: enemy
46
- position:
47
- - 104
48
- - 24
49
- count: 1
50
- stance: 2
51
- - type: e1
52
- owner: enemy
53
- position:
54
- - 40
55
- - 10
56
- count: 1
57
- stance: 2
58
- - type: e1
59
- owner: enemy
60
- position:
61
- - 48
62
- - 10
63
- count: 1
64
- stance: 2
65
- levels:
66
- easy:
67
- description: "Threat Enumeration: easy \u2014 the decision is harder (less info\
68
- \ / tighter clock / committed choice)."
69
- win_condition:
70
- all_of:
71
- - enemies_discovered_gte: 2
72
- - units_lost_lte: 1
73
- - within_ticks: 14000
74
- max_turns: 45
75
- medium:
76
- description: "Threat Enumeration: medium \u2014 the decision is harder (less info\
77
- \ / tighter clock / committed choice)."
78
- win_condition:
79
- all_of:
80
- - enemies_discovered_gte: 3
81
- - units_lost_lte: 1
82
- - within_ticks: 11000
83
- max_turns: 53
84
- hard:
85
- description: "Threat Enumeration: hard \u2014 the decision is harder (less info\
86
- \ / tighter clock / committed choice)."
87
- win_condition:
88
- all_of:
89
- - enemies_discovered_gte: 4
90
- - units_lost_lte: 0
91
- - within_ticks: 8000
92
- max_turns: 61
93
- fail_condition:
94
- units_lost_lte: -1
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
openra_bench/scenarios/packs/cat-c2-01.yaml DELETED
@@ -1,94 +0,0 @@
1
- meta:
2
- id: cat-c2-01
3
- title: 'Threat Enumeration #2'
4
- capability: perception
5
- real_world_meaning: "Estimating hidden adversary state from partial sightings before\
6
- \ acting \u2014 perception under occlusion and distractors."
7
- robotics_analogue: Scene state-estimation under fog and decoys before committing
8
- force.
9
- author: catalog-generator
10
- base_map: rush-hour-arena
11
- base:
12
- agent:
13
- faction: allies
14
- enemy:
15
- faction: soviet
16
- tools:
17
- - move_units
18
- - attack_unit
19
- - attack_move
20
- - stop
21
- planning: true
22
- termination:
23
- max_ticks: 30000
24
- actors:
25
- - type: 2tnk
26
- owner: agent
27
- position:
28
- - 8
29
- - 18
30
- count: 2
31
- - type: e1
32
- owner: agent
33
- position:
34
- - 9
35
- - 20
36
- count: 3
37
- - type: e1
38
- owner: enemy
39
- position:
40
- - 100
41
- - 20
42
- count: 2
43
- stance: 2
44
- - type: e3
45
- owner: enemy
46
- position:
47
- - 104
48
- - 24
49
- count: 1
50
- stance: 2
51
- - type: e1
52
- owner: enemy
53
- position:
54
- - 40
55
- - 10
56
- count: 1
57
- stance: 2
58
- - type: e1
59
- owner: enemy
60
- position:
61
- - 48
62
- - 10
63
- count: 1
64
- stance: 2
65
- levels:
66
- easy:
67
- description: "Threat Enumeration: easy \u2014 the decision is harder (less info\
68
- \ / tighter clock / committed choice)."
69
- win_condition:
70
- all_of:
71
- - enemies_discovered_gte: 2
72
- - units_lost_lte: 1
73
- - within_ticks: 13850
74
- max_turns: 45
75
- medium:
76
- description: "Threat Enumeration: medium \u2014 the decision is harder (less info\
77
- \ / tighter clock / committed choice)."
78
- win_condition:
79
- all_of:
80
- - enemies_discovered_gte: 3
81
- - units_lost_lte: 1
82
- - within_ticks: 10850
83
- max_turns: 53
84
- hard:
85
- description: "Threat Enumeration: hard \u2014 the decision is harder (less info\
86
- \ / tighter clock / committed choice)."
87
- win_condition:
88
- all_of:
89
- - enemies_discovered_gte: 4
90
- - units_lost_lte: 0
91
- - within_ticks: 7850
92
- max_turns: 61
93
- fail_condition:
94
- units_lost_lte: -1
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
openra_bench/scenarios/packs/cat-c2-02.yaml DELETED
@@ -1,95 +0,0 @@
1
- meta:
2
- id: cat-c2-02
3
- title: 'Threat Enumeration #3'
4
- capability: perception
5
- real_world_meaning: Estimating hidden adversary state from partial sightings before acting — perception
6
- under occlusion and distractors.
7
- robotics_analogue: Scene state-estimation under fog and decoys before committing force.
8
- author: catalog-generator
9
- status: quarantine
10
- quarantine_reason: 'redundant cat-* over-generation (audit: keep 2 representatives/category)'
11
- base_map: rush-hour-arena
12
- base:
13
- agent:
14
- faction: allies
15
- enemy:
16
- faction: soviet
17
- tools:
18
- - move_units
19
- - attack_unit
20
- - attack_move
21
- - stop
22
- planning: true
23
- termination:
24
- max_ticks: 30000
25
- actors:
26
- - type: 2tnk
27
- owner: agent
28
- position:
29
- - 8
30
- - 18
31
- count: 2
32
- - type: e1
33
- owner: agent
34
- position:
35
- - 9
36
- - 20
37
- count: 3
38
- - type: e1
39
- owner: enemy
40
- position:
41
- - 100
42
- - 20
43
- count: 2
44
- stance: 2
45
- - type: e3
46
- owner: enemy
47
- position:
48
- - 104
49
- - 24
50
- count: 1
51
- stance: 2
52
- - type: e1
53
- owner: enemy
54
- position:
55
- - 40
56
- - 10
57
- count: 1
58
- stance: 2
59
- - type: e1
60
- owner: enemy
61
- position:
62
- - 48
63
- - 10
64
- count: 1
65
- stance: 2
66
- levels:
67
- easy:
68
- description: 'Threat Enumeration: easy — the decision is harder (less info / tighter clock / committed
69
- choice).'
70
- win_condition:
71
- all_of:
72
- - enemies_discovered_gte: 2
73
- - units_lost_lte: 1
74
- - within_ticks: 13700
75
- max_turns: 45
76
- medium:
77
- description: 'Threat Enumeration: medium — the decision is harder (less info / tighter clock / committed
78
- choice).'
79
- win_condition:
80
- all_of:
81
- - enemies_discovered_gte: 3
82
- - units_lost_lte: 1
83
- - within_ticks: 10700
84
- max_turns: 53
85
- hard:
86
- description: 'Threat Enumeration: hard — the decision is harder (less info / tighter clock / committed
87
- choice).'
88
- win_condition:
89
- all_of:
90
- - enemies_discovered_gte: 4
91
- - units_lost_lte: 0
92
- - within_ticks: 7700
93
- max_turns: 61
94
- fail_condition:
95
- units_lost_lte: -1
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
openra_bench/scenarios/packs/cat-c2-03.yaml DELETED
@@ -1,95 +0,0 @@
1
- meta:
2
- id: cat-c2-03
3
- title: 'Threat Enumeration #4'
4
- capability: perception
5
- real_world_meaning: Estimating hidden adversary state from partial sightings before acting — perception
6
- under occlusion and distractors.
7
- robotics_analogue: Scene state-estimation under fog and decoys before committing force.
8
- author: catalog-generator
9
- status: quarantine
10
- quarantine_reason: 'redundant cat-* over-generation (audit: keep 2 representatives/category)'
11
- base_map: rush-hour-arena
12
- base:
13
- agent:
14
- faction: allies
15
- enemy:
16
- faction: soviet
17
- tools:
18
- - move_units
19
- - attack_unit
20
- - attack_move
21
- - stop
22
- planning: true
23
- termination:
24
- max_ticks: 30000
25
- actors:
26
- - type: 2tnk
27
- owner: agent
28
- position:
29
- - 8
30
- - 18
31
- count: 2
32
- - type: e1
33
- owner: agent
34
- position:
35
- - 9
36
- - 20
37
- count: 3
38
- - type: e1
39
- owner: enemy
40
- position:
41
- - 100
42
- - 20
43
- count: 2
44
- stance: 2
45
- - type: e3
46
- owner: enemy
47
- position:
48
- - 104
49
- - 24
50
- count: 1
51
- stance: 2
52
- - type: e1
53
- owner: enemy
54
- position:
55
- - 40
56
- - 10
57
- count: 1
58
- stance: 2
59
- - type: e1
60
- owner: enemy
61
- position:
62
- - 48
63
- - 10
64
- count: 1
65
- stance: 2
66
- levels:
67
- easy:
68
- description: 'Threat Enumeration: easy — the decision is harder (less info / tighter clock / committed
69
- choice).'
70
- win_condition:
71
- all_of:
72
- - enemies_discovered_gte: 2
73
- - units_lost_lte: 1
74
- - within_ticks: 13550
75
- max_turns: 45
76
- medium:
77
- description: 'Threat Enumeration: medium — the decision is harder (less info / tighter clock / committed
78
- choice).'
79
- win_condition:
80
- all_of:
81
- - enemies_discovered_gte: 3
82
- - units_lost_lte: 1
83
- - within_ticks: 10550
84
- max_turns: 53
85
- hard:
86
- description: 'Threat Enumeration: hard — the decision is harder (less info / tighter clock / committed
87
- choice).'
88
- win_condition:
89
- all_of:
90
- - enemies_discovered_gte: 4
91
- - units_lost_lte: 0
92
- - within_ticks: 7550
93
- max_turns: 61
94
- fail_condition:
95
- units_lost_lte: -1
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
openra_bench/scenarios/packs/cat-c2-04.yaml DELETED
@@ -1,95 +0,0 @@
1
- meta:
2
- id: cat-c2-04
3
- title: 'Threat Enumeration #5'
4
- capability: perception
5
- real_world_meaning: Estimating hidden adversary state from partial sightings before acting — perception
6
- under occlusion and distractors.
7
- robotics_analogue: Scene state-estimation under fog and decoys before committing force.
8
- author: catalog-generator
9
- status: quarantine
10
- quarantine_reason: 'redundant cat-* over-generation (audit: keep 2 representatives/category)'
11
- base_map: rush-hour-arena
12
- base:
13
- agent:
14
- faction: allies
15
- enemy:
16
- faction: soviet
17
- tools:
18
- - move_units
19
- - attack_unit
20
- - attack_move
21
- - stop
22
- planning: true
23
- termination:
24
- max_ticks: 30000
25
- actors:
26
- - type: 2tnk
27
- owner: agent
28
- position:
29
- - 8
30
- - 18
31
- count: 2
32
- - type: e1
33
- owner: agent
34
- position:
35
- - 9
36
- - 20
37
- count: 3
38
- - type: e1
39
- owner: enemy
40
- position:
41
- - 100
42
- - 20
43
- count: 2
44
- stance: 2
45
- - type: e3
46
- owner: enemy
47
- position:
48
- - 104
49
- - 24
50
- count: 1
51
- stance: 2
52
- - type: e1
53
- owner: enemy
54
- position:
55
- - 40
56
- - 10
57
- count: 1
58
- stance: 2
59
- - type: e1
60
- owner: enemy
61
- position:
62
- - 48
63
- - 10
64
- count: 1
65
- stance: 2
66
- levels:
67
- easy:
68
- description: 'Threat Enumeration: easy — the decision is harder (less info / tighter clock / committed
69
- choice).'
70
- win_condition:
71
- all_of:
72
- - enemies_discovered_gte: 2
73
- - units_lost_lte: 1
74
- - within_ticks: 13400
75
- max_turns: 45
76
- medium:
77
- description: 'Threat Enumeration: medium — the decision is harder (less info / tighter clock / committed
78
- choice).'
79
- win_condition:
80
- all_of:
81
- - enemies_discovered_gte: 3
82
- - units_lost_lte: 1
83
- - within_ticks: 10400
84
- max_turns: 53
85
- hard:
86
- description: 'Threat Enumeration: hard — the decision is harder (less info / tighter clock / committed
87
- choice).'
88
- win_condition:
89
- all_of:
90
- - enemies_discovered_gte: 4
91
- - units_lost_lte: 0
92
- - within_ticks: 7400
93
- max_turns: 61
94
- fail_condition:
95
- units_lost_lte: -1
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
openra_bench/scenarios/packs/cat-c2-05.yaml DELETED
@@ -1,95 +0,0 @@
1
- meta:
2
- id: cat-c2-05
3
- title: 'Threat Enumeration #6'
4
- capability: perception
5
- real_world_meaning: Estimating hidden adversary state from partial sightings before acting — perception
6
- under occlusion and distractors.
7
- robotics_analogue: Scene state-estimation under fog and decoys before committing force.
8
- author: catalog-generator
9
- status: quarantine
10
- quarantine_reason: 'redundant cat-* over-generation (audit: keep 2 representatives/category)'
11
- base_map: rush-hour-arena
12
- base:
13
- agent:
14
- faction: allies
15
- enemy:
16
- faction: soviet
17
- tools:
18
- - move_units
19
- - attack_unit
20
- - attack_move
21
- - stop
22
- planning: true
23
- termination:
24
- max_ticks: 30000
25
- actors:
26
- - type: 2tnk
27
- owner: agent
28
- position:
29
- - 8
30
- - 18
31
- count: 2
32
- - type: e1
33
- owner: agent
34
- position:
35
- - 9
36
- - 20
37
- count: 3
38
- - type: e1
39
- owner: enemy
40
- position:
41
- - 100
42
- - 20
43
- count: 2
44
- stance: 2
45
- - type: e3
46
- owner: enemy
47
- position:
48
- - 104
49
- - 24
50
- count: 1
51
- stance: 2
52
- - type: e1
53
- owner: enemy
54
- position:
55
- - 40
56
- - 10
57
- count: 1
58
- stance: 2
59
- - type: e1
60
- owner: enemy
61
- position:
62
- - 48
63
- - 10
64
- count: 1
65
- stance: 2
66
- levels:
67
- easy:
68
- description: 'Threat Enumeration: easy — the decision is harder (less info / tighter clock / committed
69
- choice).'
70
- win_condition:
71
- all_of:
72
- - enemies_discovered_gte: 2
73
- - units_lost_lte: 1
74
- - within_ticks: 13250
75
- max_turns: 45
76
- medium:
77
- description: 'Threat Enumeration: medium — the decision is harder (less info / tighter clock / committed
78
- choice).'
79
- win_condition:
80
- all_of:
81
- - enemies_discovered_gte: 3
82
- - units_lost_lte: 1
83
- - within_ticks: 10250
84
- max_turns: 53
85
- hard:
86
- description: 'Threat Enumeration: hard — the decision is harder (less info / tighter clock / committed
87
- choice).'
88
- win_condition:
89
- all_of:
90
- - enemies_discovered_gte: 4
91
- - units_lost_lte: 0
92
- - within_ticks: 7250
93
- max_turns: 61
94
- fail_condition:
95
- units_lost_lte: -1
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
openra_bench/scenarios/packs/cat-c3-00.yaml DELETED
@@ -1,80 +0,0 @@
1
- meta:
2
- id: cat-c3-00
3
- title: 'Tech Critical Path #1'
4
- capability: reasoning
5
- real_world_meaning: "Precondition-ordered construction (B needs A) to hit a capability\
6
- \ deadline \u2014 topological planning under a cost+time budget."
7
- robotics_analogue: Task-graph scheduling with prerequisites under a makespan.
8
- author: catalog-generator
9
- base_map: rush-hour-arena
10
- base:
11
- agent:
12
- faction: allies
13
- enemy:
14
- faction: soviet
15
- tools:
16
- - build
17
- - place_building
18
- - move_units
19
- - deploy
20
- - stop
21
- planning: true
22
- termination:
23
- max_ticks: 40000
24
- actors:
25
- - type: fact
26
- owner: agent
27
- position:
28
- - 10
29
- - 18
30
- - type: powr
31
- owner: agent
32
- position:
33
- - 14
34
- - 18
35
- - type: jeep
36
- owner: agent
37
- position:
38
- - 8
39
- - 16
40
- - type: e1
41
- owner: enemy
42
- position:
43
- - 114
44
- - 34
45
- count: 1
46
- stance: 2
47
- levels:
48
- easy:
49
- description: "Tech Critical Path: easy \u2014 the decision is harder (less info\
50
- \ / tighter clock / committed choice)."
51
- win_condition:
52
- all_of:
53
- - has_building: tent
54
- - building_total_gte: 4
55
- - power_surplus_gte: 0
56
- - within_ticks: 22000
57
- max_turns: 45
58
- medium:
59
- description: "Tech Critical Path: medium \u2014 the decision is harder (less info\
60
- \ / tighter clock / committed choice)."
61
- win_condition:
62
- all_of:
63
- - has_building: tent
64
- - building_total_gte: 5
65
- - power_surplus_gte: 0
66
- - within_ticks: 17000
67
- max_turns: 53
68
- hard:
69
- description: "Tech Critical Path: hard \u2014 the decision is harder (less info\
70
- \ / tighter clock / committed choice)."
71
- win_condition:
72
- all_of:
73
- - has_building: tent
74
- - building_total_gte: 6
75
- - power_surplus_gte: 0
76
- - within_ticks: 12000
77
- max_turns: 61
78
- fail_condition:
79
- units_lost_lte: -1
80
- starting_cash: 6000
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
openra_bench/scenarios/packs/cat-c3-01.yaml DELETED
@@ -1,80 +0,0 @@
1
- meta:
2
- id: cat-c3-01
3
- title: 'Tech Critical Path #2'
4
- capability: reasoning
5
- real_world_meaning: "Precondition-ordered construction (B needs A) to hit a capability\
6
- \ deadline \u2014 topological planning under a cost+time budget."
7
- robotics_analogue: Task-graph scheduling with prerequisites under a makespan.
8
- author: catalog-generator
9
- base_map: rush-hour-arena
10
- base:
11
- agent:
12
- faction: allies
13
- enemy:
14
- faction: soviet
15
- tools:
16
- - build
17
- - place_building
18
- - move_units
19
- - deploy
20
- - stop
21
- planning: true
22
- termination:
23
- max_ticks: 40000
24
- actors:
25
- - type: fact
26
- owner: agent
27
- position:
28
- - 10
29
- - 18
30
- - type: powr
31
- owner: agent
32
- position:
33
- - 14
34
- - 18
35
- - type: jeep
36
- owner: agent
37
- position:
38
- - 8
39
- - 16
40
- - type: e1
41
- owner: enemy
42
- position:
43
- - 114
44
- - 34
45
- count: 1
46
- stance: 2
47
- levels:
48
- easy:
49
- description: "Tech Critical Path: easy \u2014 the decision is harder (less info\
50
- \ / tighter clock / committed choice)."
51
- win_condition:
52
- all_of:
53
- - has_building: tent
54
- - building_total_gte: 4
55
- - power_surplus_gte: 0
56
- - within_ticks: 21700
57
- max_turns: 45
58
- medium:
59
- description: "Tech Critical Path: medium \u2014 the decision is harder (less info\
60
- \ / tighter clock / committed choice)."
61
- win_condition:
62
- all_of:
63
- - has_building: tent
64
- - building_total_gte: 5
65
- - power_surplus_gte: 0
66
- - within_ticks: 16700
67
- max_turns: 53
68
- hard:
69
- description: "Tech Critical Path: hard \u2014 the decision is harder (less info\
70
- \ / tighter clock / committed choice)."
71
- win_condition:
72
- all_of:
73
- - has_building: tent
74
- - building_total_gte: 6
75
- - power_surplus_gte: 0
76
- - within_ticks: 11700
77
- max_turns: 61
78
- fail_condition:
79
- units_lost_lte: -1
80
- starting_cash: 6000
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
openra_bench/scenarios/packs/cat-c3-02.yaml DELETED
@@ -1,82 +0,0 @@
1
- meta:
2
- id: cat-c3-02
3
- title: 'Tech Critical Path #3'
4
- capability: reasoning
5
- real_world_meaning: Precondition-ordered construction (B needs A) to hit a capability deadline — topological
6
- planning under a cost+time budget.
7
- robotics_analogue: Task-graph scheduling with prerequisites under a makespan.
8
- author: catalog-generator
9
- status: quarantine
10
- quarantine_reason: 'redundant cat-* over-generation (audit: keep 2 representatives/category)'
11
- base_map: rush-hour-arena
12
- base:
13
- agent:
14
- faction: allies
15
- enemy:
16
- faction: soviet
17
- tools:
18
- - build
19
- - place_building
20
- - move_units
21
- - deploy
22
- - stop
23
- planning: true
24
- termination:
25
- max_ticks: 40000
26
- actors:
27
- - type: fact
28
- owner: agent
29
- position:
30
- - 10
31
- - 18
32
- - type: powr
33
- owner: agent
34
- position:
35
- - 14
36
- - 18
37
- - type: jeep
38
- owner: agent
39
- position:
40
- - 8
41
- - 16
42
- - type: e1
43
- owner: enemy
44
- position:
45
- - 114
46
- - 34
47
- count: 1
48
- stance: 2
49
- levels:
50
- easy:
51
- description: 'Tech Critical Path: easy — the decision is harder (less info / tighter clock / committed
52
- choice).'
53
- win_condition:
54
- all_of:
55
- - has_building: tent
56
- - building_total_gte: 4
57
- - power_surplus_gte: 0
58
- - within_ticks: 21400
59
- max_turns: 45
60
- medium:
61
- description: 'Tech Critical Path: medium — the decision is harder (less info / tighter clock / committed
62
- choice).'
63
- win_condition:
64
- all_of:
65
- - has_building: tent
66
- - building_total_gte: 5
67
- - power_surplus_gte: 0
68
- - within_ticks: 16400
69
- max_turns: 53
70
- hard:
71
- description: 'Tech Critical Path: hard — the decision is harder (less info / tighter clock / committed
72
- choice).'
73
- win_condition:
74
- all_of:
75
- - has_building: tent
76
- - building_total_gte: 6
77
- - power_surplus_gte: 0
78
- - within_ticks: 11400
79
- max_turns: 61
80
- fail_condition:
81
- units_lost_lte: -1
82
- starting_cash: 6000
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
openra_bench/scenarios/packs/cat-c3-03.yaml DELETED
@@ -1,82 +0,0 @@
1
- meta:
2
- id: cat-c3-03
3
- title: 'Tech Critical Path #4'
4
- capability: reasoning
5
- real_world_meaning: Precondition-ordered construction (B needs A) to hit a capability deadline — topological
6
- planning under a cost+time budget.
7
- robotics_analogue: Task-graph scheduling with prerequisites under a makespan.
8
- author: catalog-generator
9
- status: quarantine
10
- quarantine_reason: 'redundant cat-* over-generation (audit: keep 2 representatives/category)'
11
- base_map: rush-hour-arena
12
- base:
13
- agent:
14
- faction: allies
15
- enemy:
16
- faction: soviet
17
- tools:
18
- - build
19
- - place_building
20
- - move_units
21
- - deploy
22
- - stop
23
- planning: true
24
- termination:
25
- max_ticks: 40000
26
- actors:
27
- - type: fact
28
- owner: agent
29
- position:
30
- - 10
31
- - 18
32
- - type: powr
33
- owner: agent
34
- position:
35
- - 14
36
- - 18
37
- - type: jeep
38
- owner: agent
39
- position:
40
- - 8
41
- - 16
42
- - type: e1
43
- owner: enemy
44
- position:
45
- - 114
46
- - 34
47
- count: 1
48
- stance: 2
49
- levels:
50
- easy:
51
- description: 'Tech Critical Path: easy — the decision is harder (less info / tighter clock / committed
52
- choice).'
53
- win_condition:
54
- all_of:
55
- - has_building: tent
56
- - building_total_gte: 4
57
- - power_surplus_gte: 0
58
- - within_ticks: 21100
59
- max_turns: 45
60
- medium:
61
- description: 'Tech Critical Path: medium — the decision is harder (less info / tighter clock / committed
62
- choice).'
63
- win_condition:
64
- all_of:
65
- - has_building: tent
66
- - building_total_gte: 5
67
- - power_surplus_gte: 0
68
- - within_ticks: 16100
69
- max_turns: 53
70
- hard:
71
- description: 'Tech Critical Path: hard — the decision is harder (less info / tighter clock / committed
72
- choice).'
73
- win_condition:
74
- all_of:
75
- - has_building: tent
76
- - building_total_gte: 6
77
- - power_surplus_gte: 0
78
- - within_ticks: 11100
79
- max_turns: 61
80
- fail_condition:
81
- units_lost_lte: -1
82
- starting_cash: 6000
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
openra_bench/scenarios/packs/cat-c3-04.yaml DELETED
@@ -1,82 +0,0 @@
1
- meta:
2
- id: cat-c3-04
3
- title: 'Tech Critical Path #5'
4
- capability: reasoning
5
- real_world_meaning: Precondition-ordered construction (B needs A) to hit a capability deadline — topological
6
- planning under a cost+time budget.
7
- robotics_analogue: Task-graph scheduling with prerequisites under a makespan.
8
- author: catalog-generator
9
- status: quarantine
10
- quarantine_reason: 'redundant cat-* over-generation (audit: keep 2 representatives/category)'
11
- base_map: rush-hour-arena
12
- base:
13
- agent:
14
- faction: allies
15
- enemy:
16
- faction: soviet
17
- tools:
18
- - build
19
- - place_building
20
- - move_units
21
- - deploy
22
- - stop
23
- planning: true
24
- termination:
25
- max_ticks: 40000
26
- actors:
27
- - type: fact
28
- owner: agent
29
- position:
30
- - 10
31
- - 18
32
- - type: powr
33
- owner: agent
34
- position:
35
- - 14
36
- - 18
37
- - type: jeep
38
- owner: agent
39
- position:
40
- - 8
41
- - 16
42
- - type: e1
43
- owner: enemy
44
- position:
45
- - 114
46
- - 34
47
- count: 1
48
- stance: 2
49
- levels:
50
- easy:
51
- description: 'Tech Critical Path: easy — the decision is harder (less info / tighter clock / committed
52
- choice).'
53
- win_condition:
54
- all_of:
55
- - has_building: tent
56
- - building_total_gte: 4
57
- - power_surplus_gte: 0
58
- - within_ticks: 20800
59
- max_turns: 45
60
- medium:
61
- description: 'Tech Critical Path: medium — the decision is harder (less info / tighter clock / committed
62
- choice).'
63
- win_condition:
64
- all_of:
65
- - has_building: tent
66
- - building_total_gte: 5
67
- - power_surplus_gte: 0
68
- - within_ticks: 15800
69
- max_turns: 53
70
- hard:
71
- description: 'Tech Critical Path: hard — the decision is harder (less info / tighter clock / committed
72
- choice).'
73
- win_condition:
74
- all_of:
75
- - has_building: tent
76
- - building_total_gte: 6
77
- - power_surplus_gte: 0
78
- - within_ticks: 10800
79
- max_turns: 61
80
- fail_condition:
81
- units_lost_lte: -1
82
- starting_cash: 6000
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
openra_bench/scenarios/packs/cat-c3-05.yaml DELETED
@@ -1,82 +0,0 @@
1
- meta:
2
- id: cat-c3-05
3
- title: 'Tech Critical Path #6'
4
- capability: reasoning
5
- real_world_meaning: Precondition-ordered construction (B needs A) to hit a capability deadline — topological
6
- planning under a cost+time budget.
7
- robotics_analogue: Task-graph scheduling with prerequisites under a makespan.
8
- author: catalog-generator
9
- status: quarantine
10
- quarantine_reason: 'redundant cat-* over-generation (audit: keep 2 representatives/category)'
11
- base_map: rush-hour-arena
12
- base:
13
- agent:
14
- faction: allies
15
- enemy:
16
- faction: soviet
17
- tools:
18
- - build
19
- - place_building
20
- - move_units
21
- - deploy
22
- - stop
23
- planning: true
24
- termination:
25
- max_ticks: 40000
26
- actors:
27
- - type: fact
28
- owner: agent
29
- position:
30
- - 10
31
- - 18
32
- - type: powr
33
- owner: agent
34
- position:
35
- - 14
36
- - 18
37
- - type: jeep
38
- owner: agent
39
- position:
40
- - 8
41
- - 16
42
- - type: e1
43
- owner: enemy
44
- position:
45
- - 114
46
- - 34
47
- count: 1
48
- stance: 2
49
- levels:
50
- easy:
51
- description: 'Tech Critical Path: easy — the decision is harder (less info / tighter clock / committed
52
- choice).'
53
- win_condition:
54
- all_of:
55
- - has_building: tent
56
- - building_total_gte: 4
57
- - power_surplus_gte: 0
58
- - within_ticks: 20500
59
- max_turns: 45
60
- medium:
61
- description: 'Tech Critical Path: medium — the decision is harder (less info / tighter clock / committed
62
- choice).'
63
- win_condition:
64
- all_of:
65
- - has_building: tent
66
- - building_total_gte: 5
67
- - power_surplus_gte: 0
68
- - within_ticks: 15500
69
- max_turns: 53
70
- hard:
71
- description: 'Tech Critical Path: hard — the decision is harder (less info / tighter clock / committed
72
- choice).'
73
- win_condition:
74
- all_of:
75
- - has_building: tent
76
- - building_total_gte: 6
77
- - power_surplus_gte: 0
78
- - within_ticks: 10500
79
- max_turns: 61
80
- fail_condition:
81
- units_lost_lte: -1
82
- starting_cash: 6000
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
openra_bench/scenarios/packs/cat-c4-00.yaml DELETED
@@ -1,77 +0,0 @@
1
- meta:
2
- id: cat-c4-00
3
- title: 'Power-Budget Online #1'
4
- capability: reasoning
5
- real_world_meaning: "Sequencing subsystem bring-up so a shared online resource never\
6
- \ goes negative \u2014 constraint-respecting schedule."
7
- robotics_analogue: Keeping an online resource (power) non-negative while expanding.
8
- author: catalog-generator
9
- base_map: rush-hour-arena
10
- base:
11
- agent:
12
- faction: allies
13
- enemy:
14
- faction: soviet
15
- tools:
16
- - build
17
- - place_building
18
- - move_units
19
- - deploy
20
- - stop
21
- planning: true
22
- termination:
23
- max_ticks: 40000
24
- actors:
25
- - type: fact
26
- owner: agent
27
- position:
28
- - 10
29
- - 18
30
- - type: powr
31
- owner: agent
32
- position:
33
- - 14
34
- - 18
35
- - type: jeep
36
- owner: agent
37
- position:
38
- - 8
39
- - 16
40
- - type: e1
41
- owner: enemy
42
- position:
43
- - 114
44
- - 34
45
- count: 1
46
- stance: 2
47
- levels:
48
- easy:
49
- description: "Power-Budget Online: easy \u2014 the decision is harder (less info\
50
- \ / tighter clock / committed choice)."
51
- win_condition:
52
- all_of:
53
- - power_surplus_gte: 0
54
- - building_total_gte: 4
55
- - within_ticks: 20000
56
- max_turns: 45
57
- medium:
58
- description: "Power-Budget Online: medium \u2014 the decision is harder (less\
59
- \ info / tighter clock / committed choice)."
60
- win_condition:
61
- all_of:
62
- - power_surplus_gte: 0
63
- - building_total_gte: 5
64
- - within_ticks: 15500
65
- max_turns: 53
66
- hard:
67
- description: "Power-Budget Online: hard \u2014 the decision is harder (less info\
68
- \ / tighter clock / committed choice)."
69
- win_condition:
70
- all_of:
71
- - power_surplus_gte: 0
72
- - building_total_gte: 6
73
- - within_ticks: 11000
74
- max_turns: 61
75
- fail_condition:
76
- units_lost_lte: -1
77
- starting_cash: 5000
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
openra_bench/scenarios/packs/cat-c4-01.yaml DELETED
@@ -1,77 +0,0 @@
1
- meta:
2
- id: cat-c4-01
3
- title: 'Power-Budget Online #2'
4
- capability: reasoning
5
- real_world_meaning: "Sequencing subsystem bring-up so a shared online resource never\
6
- \ goes negative \u2014 constraint-respecting schedule."
7
- robotics_analogue: Keeping an online resource (power) non-negative while expanding.
8
- author: catalog-generator
9
- base_map: rush-hour-arena
10
- base:
11
- agent:
12
- faction: allies
13
- enemy:
14
- faction: soviet
15
- tools:
16
- - build
17
- - place_building
18
- - move_units
19
- - deploy
20
- - stop
21
- planning: true
22
- termination:
23
- max_ticks: 40000
24
- actors:
25
- - type: fact
26
- owner: agent
27
- position:
28
- - 10
29
- - 18
30
- - type: powr
31
- owner: agent
32
- position:
33
- - 14
34
- - 18
35
- - type: jeep
36
- owner: agent
37
- position:
38
- - 8
39
- - 16
40
- - type: e1
41
- owner: enemy
42
- position:
43
- - 114
44
- - 34
45
- count: 1
46
- stance: 2
47
- levels:
48
- easy:
49
- description: "Power-Budget Online: easy \u2014 the decision is harder (less info\
50
- \ / tighter clock / committed choice)."
51
- win_condition:
52
- all_of:
53
- - power_surplus_gte: 0
54
- - building_total_gte: 4
55
- - within_ticks: 19750
56
- max_turns: 45
57
- medium:
58
- description: "Power-Budget Online: medium \u2014 the decision is harder (less\
59
- \ info / tighter clock / committed choice)."
60
- win_condition:
61
- all_of:
62
- - power_surplus_gte: 0
63
- - building_total_gte: 5
64
- - within_ticks: 15250
65
- max_turns: 53
66
- hard:
67
- description: "Power-Budget Online: hard \u2014 the decision is harder (less info\
68
- \ / tighter clock / committed choice)."
69
- win_condition:
70
- all_of:
71
- - power_surplus_gte: 0
72
- - building_total_gte: 6
73
- - within_ticks: 10750
74
- max_turns: 61
75
- fail_condition:
76
- units_lost_lte: -1
77
- starting_cash: 5000
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
openra_bench/scenarios/packs/cat-c4-02.yaml DELETED
@@ -1,79 +0,0 @@
1
- meta:
2
- id: cat-c4-02
3
- title: 'Power-Budget Online #3'
4
- capability: reasoning
5
- real_world_meaning: Sequencing subsystem bring-up so a shared online resource never goes negative —
6
- constraint-respecting schedule.
7
- robotics_analogue: Keeping an online resource (power) non-negative while expanding.
8
- author: catalog-generator
9
- status: quarantine
10
- quarantine_reason: 'redundant cat-* over-generation (audit: keep 2 representatives/category)'
11
- base_map: rush-hour-arena
12
- base:
13
- agent:
14
- faction: allies
15
- enemy:
16
- faction: soviet
17
- tools:
18
- - build
19
- - place_building
20
- - move_units
21
- - deploy
22
- - stop
23
- planning: true
24
- termination:
25
- max_ticks: 40000
26
- actors:
27
- - type: fact
28
- owner: agent
29
- position:
30
- - 10
31
- - 18
32
- - type: powr
33
- owner: agent
34
- position:
35
- - 14
36
- - 18
37
- - type: jeep
38
- owner: agent
39
- position:
40
- - 8
41
- - 16
42
- - type: e1
43
- owner: enemy
44
- position:
45
- - 114
46
- - 34
47
- count: 1
48
- stance: 2
49
- levels:
50
- easy:
51
- description: 'Power-Budget Online: easy — the decision is harder (less info / tighter clock / committed
52
- choice).'
53
- win_condition:
54
- all_of:
55
- - power_surplus_gte: 0
56
- - building_total_gte: 4
57
- - within_ticks: 19500
58
- max_turns: 45
59
- medium:
60
- description: 'Power-Budget Online: medium — the decision is harder (less info / tighter clock / committed
61
- choice).'
62
- win_condition:
63
- all_of:
64
- - power_surplus_gte: 0
65
- - building_total_gte: 5
66
- - within_ticks: 15000
67
- max_turns: 53
68
- hard:
69
- description: 'Power-Budget Online: hard — the decision is harder (less info / tighter clock / committed
70
- choice).'
71
- win_condition:
72
- all_of:
73
- - power_surplus_gte: 0
74
- - building_total_gte: 6
75
- - within_ticks: 10500
76
- max_turns: 61
77
- fail_condition:
78
- units_lost_lte: -1
79
- starting_cash: 5000
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
openra_bench/scenarios/packs/cat-c4-03.yaml DELETED
@@ -1,79 +0,0 @@
1
- meta:
2
- id: cat-c4-03
3
- title: 'Power-Budget Online #4'
4
- capability: reasoning
5
- real_world_meaning: Sequencing subsystem bring-up so a shared online resource never goes negative —
6
- constraint-respecting schedule.
7
- robotics_analogue: Keeping an online resource (power) non-negative while expanding.
8
- author: catalog-generator
9
- status: quarantine
10
- quarantine_reason: 'redundant cat-* over-generation (audit: keep 2 representatives/category)'
11
- base_map: rush-hour-arena
12
- base:
13
- agent:
14
- faction: allies
15
- enemy:
16
- faction: soviet
17
- tools:
18
- - build
19
- - place_building
20
- - move_units
21
- - deploy
22
- - stop
23
- planning: true
24
- termination:
25
- max_ticks: 40000
26
- actors:
27
- - type: fact
28
- owner: agent
29
- position:
30
- - 10
31
- - 18
32
- - type: powr
33
- owner: agent
34
- position:
35
- - 14
36
- - 18
37
- - type: jeep
38
- owner: agent
39
- position:
40
- - 8
41
- - 16
42
- - type: e1
43
- owner: enemy
44
- position:
45
- - 114
46
- - 34
47
- count: 1
48
- stance: 2
49
- levels:
50
- easy:
51
- description: 'Power-Budget Online: easy — the decision is harder (less info / tighter clock / committed
52
- choice).'
53
- win_condition:
54
- all_of:
55
- - power_surplus_gte: 0
56
- - building_total_gte: 4
57
- - within_ticks: 19250
58
- max_turns: 45
59
- medium:
60
- description: 'Power-Budget Online: medium — the decision is harder (less info / tighter clock / committed
61
- choice).'
62
- win_condition:
63
- all_of:
64
- - power_surplus_gte: 0
65
- - building_total_gte: 5
66
- - within_ticks: 14750
67
- max_turns: 53
68
- hard:
69
- description: 'Power-Budget Online: hard — the decision is harder (less info / tighter clock / committed
70
- choice).'
71
- win_condition:
72
- all_of:
73
- - power_surplus_gte: 0
74
- - building_total_gte: 6
75
- - within_ticks: 10250
76
- max_turns: 61
77
- fail_condition:
78
- units_lost_lte: -1
79
- starting_cash: 5000
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
openra_bench/scenarios/packs/cat-c4-04.yaml DELETED
@@ -1,79 +0,0 @@
1
- meta:
2
- id: cat-c4-04
3
- title: 'Power-Budget Online #5'
4
- capability: reasoning
5
- real_world_meaning: Sequencing subsystem bring-up so a shared online resource never goes negative —
6
- constraint-respecting schedule.
7
- robotics_analogue: Keeping an online resource (power) non-negative while expanding.
8
- author: catalog-generator
9
- status: quarantine
10
- quarantine_reason: 'redundant cat-* over-generation (audit: keep 2 representatives/category)'
11
- base_map: rush-hour-arena
12
- base:
13
- agent:
14
- faction: allies
15
- enemy:
16
- faction: soviet
17
- tools:
18
- - build
19
- - place_building
20
- - move_units
21
- - deploy
22
- - stop
23
- planning: true
24
- termination:
25
- max_ticks: 40000
26
- actors:
27
- - type: fact
28
- owner: agent
29
- position:
30
- - 10
31
- - 18
32
- - type: powr
33
- owner: agent
34
- position:
35
- - 14
36
- - 18
37
- - type: jeep
38
- owner: agent
39
- position:
40
- - 8
41
- - 16
42
- - type: e1
43
- owner: enemy
44
- position:
45
- - 114
46
- - 34
47
- count: 1
48
- stance: 2
49
- levels:
50
- easy:
51
- description: 'Power-Budget Online: easy — the decision is harder (less info / tighter clock / committed
52
- choice).'
53
- win_condition:
54
- all_of:
55
- - power_surplus_gte: 0
56
- - building_total_gte: 4
57
- - within_ticks: 19000
58
- max_turns: 45
59
- medium:
60
- description: 'Power-Budget Online: medium — the decision is harder (less info / tighter clock / committed
61
- choice).'
62
- win_condition:
63
- all_of:
64
- - power_surplus_gte: 0
65
- - building_total_gte: 5
66
- - within_ticks: 14500
67
- max_turns: 53
68
- hard:
69
- description: 'Power-Budget Online: hard — the decision is harder (less info / tighter clock / committed
70
- choice).'
71
- win_condition:
72
- all_of:
73
- - power_surplus_gte: 0
74
- - building_total_gte: 6
75
- - within_ticks: 10000
76
- max_turns: 61
77
- fail_condition:
78
- units_lost_lte: -1
79
- starting_cash: 5000
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
openra_bench/scenarios/packs/cat-c5-00.yaml DELETED
@@ -1,89 +0,0 @@
1
- meta:
2
- id: cat-c5-00
3
- title: 'Budget Allocation (spend) #1'
4
- capability: reasoning
5
- real_world_meaning: "Allocating a finite budget across competing investments where\
6
- \ a half-measure on both fails \u2014 resource-allocation under a hard cap."
7
- robotics_analogue: 'Indivisible-budget allocation: wide vs deep, splitting clears
8
- nothing.'
9
- author: catalog-generator
10
- base_map: rush-hour-arena
11
- base:
12
- agent:
13
- faction: allies
14
- enemy:
15
- faction: soviet
16
- tools:
17
- - build
18
- - place_building
19
- - move_units
20
- - deploy
21
- - stop
22
- planning: true
23
- termination:
24
- max_ticks: 40000
25
- actors:
26
- - type: fact
27
- owner: agent
28
- position:
29
- - 10
30
- - 18
31
- - type: powr
32
- owner: agent
33
- position:
34
- - 14
35
- - 18
36
- - type: tent
37
- owner: agent
38
- position:
39
- - 12
40
- - 21
41
- - type: jeep
42
- owner: agent
43
- position:
44
- - 8
45
- - 16
46
- - type: e1
47
- owner: enemy
48
- position:
49
- - 114
50
- - 34
51
- count: 1
52
- stance: 2
53
- levels:
54
- easy:
55
- description: "Budget Allocation (spend): easy \u2014 the decision is harder (less\
56
- \ info / tighter clock / committed choice)."
57
- win_condition:
58
- all_of:
59
- - building_count_gte:
60
- type: powr
61
- n: 2
62
- - building_total_gte: 5
63
- - within_ticks: 20000
64
- max_turns: 45
65
- medium:
66
- description: "Budget Allocation (spend): medium \u2014 the decision is harder\
67
- \ (less info / tighter clock / committed choice)."
68
- win_condition:
69
- all_of:
70
- - building_count_gte:
71
- type: powr
72
- n: 2
73
- - building_total_gte: 6
74
- - within_ticks: 15500
75
- max_turns: 53
76
- hard:
77
- description: "Budget Allocation (spend): hard \u2014 the decision is harder (less\
78
- \ info / tighter clock / committed choice)."
79
- win_condition:
80
- all_of:
81
- - building_count_gte:
82
- type: powr
83
- n: 2
84
- - building_total_gte: 7
85
- - within_ticks: 11000
86
- max_turns: 61
87
- fail_condition:
88
- units_lost_lte: -1
89
- starting_cash: 2200
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
openra_bench/scenarios/packs/cat-c5-01.yaml DELETED
@@ -1,89 +0,0 @@
1
- meta:
2
- id: cat-c5-01
3
- title: 'Budget Allocation (spend) #2'
4
- capability: reasoning
5
- real_world_meaning: "Allocating a finite budget across competing investments where\
6
- \ a half-measure on both fails \u2014 resource-allocation under a hard cap."
7
- robotics_analogue: 'Indivisible-budget allocation: wide vs deep, splitting clears
8
- nothing.'
9
- author: catalog-generator
10
- base_map: rush-hour-arena
11
- base:
12
- agent:
13
- faction: allies
14
- enemy:
15
- faction: soviet
16
- tools:
17
- - build
18
- - place_building
19
- - move_units
20
- - deploy
21
- - stop
22
- planning: true
23
- termination:
24
- max_ticks: 40000
25
- actors:
26
- - type: fact
27
- owner: agent
28
- position:
29
- - 10
30
- - 18
31
- - type: powr
32
- owner: agent
33
- position:
34
- - 14
35
- - 18
36
- - type: tent
37
- owner: agent
38
- position:
39
- - 12
40
- - 21
41
- - type: jeep
42
- owner: agent
43
- position:
44
- - 8
45
- - 16
46
- - type: e1
47
- owner: enemy
48
- position:
49
- - 114
50
- - 34
51
- count: 1
52
- stance: 2
53
- levels:
54
- easy:
55
- description: "Budget Allocation (spend): easy \u2014 the decision is harder (less\
56
- \ info / tighter clock / committed choice)."
57
- win_condition:
58
- all_of:
59
- - building_count_gte:
60
- type: powr
61
- n: 2
62
- - building_total_gte: 5
63
- - within_ticks: 20000
64
- max_turns: 45
65
- medium:
66
- description: "Budget Allocation (spend): medium \u2014 the decision is harder\
67
- \ (less info / tighter clock / committed choice)."
68
- win_condition:
69
- all_of:
70
- - building_count_gte:
71
- type: powr
72
- n: 2
73
- - building_total_gte: 6
74
- - within_ticks: 15500
75
- max_turns: 53
76
- hard:
77
- description: "Budget Allocation (spend): hard \u2014 the decision is harder (less\
78
- \ info / tighter clock / committed choice)."
79
- win_condition:
80
- all_of:
81
- - building_count_gte:
82
- type: powr
83
- n: 2
84
- - building_total_gte: 7
85
- - within_ticks: 11000
86
- max_turns: 61
87
- fail_condition:
88
- units_lost_lte: -1
89
- starting_cash: 2080
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
openra_bench/scenarios/packs/cat-c5-02.yaml DELETED
@@ -1,90 +0,0 @@
1
- meta:
2
- id: cat-c5-02
3
- title: 'Budget Allocation (spend) #3'
4
- capability: reasoning
5
- real_world_meaning: Allocating a finite budget across competing investments where a half-measure on
6
- both fails — resource-allocation under a hard cap.
7
- robotics_analogue: 'Indivisible-budget allocation: wide vs deep, splitting clears nothing.'
8
- author: catalog-generator
9
- status: quarantine
10
- quarantine_reason: 'redundant cat-* over-generation (audit: keep 2 representatives/category)'
11
- base_map: rush-hour-arena
12
- base:
13
- agent:
14
- faction: allies
15
- enemy:
16
- faction: soviet
17
- tools:
18
- - build
19
- - place_building
20
- - move_units
21
- - deploy
22
- - stop
23
- planning: true
24
- termination:
25
- max_ticks: 40000
26
- actors:
27
- - type: fact
28
- owner: agent
29
- position:
30
- - 10
31
- - 18
32
- - type: powr
33
- owner: agent
34
- position:
35
- - 14
36
- - 18
37
- - type: tent
38
- owner: agent
39
- position:
40
- - 12
41
- - 21
42
- - type: jeep
43
- owner: agent
44
- position:
45
- - 8
46
- - 16
47
- - type: e1
48
- owner: enemy
49
- position:
50
- - 114
51
- - 34
52
- count: 1
53
- stance: 2
54
- levels:
55
- easy:
56
- description: 'Budget Allocation (spend): easy — the decision is harder (less info / tighter clock
57
- / committed choice).'
58
- win_condition:
59
- all_of:
60
- - building_count_gte:
61
- type: powr
62
- n: 2
63
- - building_total_gte: 5
64
- - within_ticks: 20000
65
- max_turns: 45
66
- medium:
67
- description: 'Budget Allocation (spend): medium — the decision is harder (less info / tighter clock
68
- / committed choice).'
69
- win_condition:
70
- all_of:
71
- - building_count_gte:
72
- type: powr
73
- n: 2
74
- - building_total_gte: 6
75
- - within_ticks: 15500
76
- max_turns: 53
77
- hard:
78
- description: 'Budget Allocation (spend): hard — the decision is harder (less info / tighter clock
79
- / committed choice).'
80
- win_condition:
81
- all_of:
82
- - building_count_gte:
83
- type: powr
84
- n: 2
85
- - building_total_gte: 7
86
- - within_ticks: 11000
87
- max_turns: 61
88
- fail_condition:
89
- units_lost_lte: -1
90
- starting_cash: 1960
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
openra_bench/scenarios/packs/cat-c5-03.yaml DELETED
@@ -1,90 +0,0 @@
1
- meta:
2
- id: cat-c5-03
3
- title: 'Budget Allocation (spend) #4'
4
- capability: reasoning
5
- real_world_meaning: Allocating a finite budget across competing investments where a half-measure on
6
- both fails — resource-allocation under a hard cap.
7
- robotics_analogue: 'Indivisible-budget allocation: wide vs deep, splitting clears nothing.'
8
- author: catalog-generator
9
- status: quarantine
10
- quarantine_reason: 'redundant cat-* over-generation (audit: keep 2 representatives/category)'
11
- base_map: rush-hour-arena
12
- base:
13
- agent:
14
- faction: allies
15
- enemy:
16
- faction: soviet
17
- tools:
18
- - build
19
- - place_building
20
- - move_units
21
- - deploy
22
- - stop
23
- planning: true
24
- termination:
25
- max_ticks: 40000
26
- actors:
27
- - type: fact
28
- owner: agent
29
- position:
30
- - 10
31
- - 18
32
- - type: powr
33
- owner: agent
34
- position:
35
- - 14
36
- - 18
37
- - type: tent
38
- owner: agent
39
- position:
40
- - 12
41
- - 21
42
- - type: jeep
43
- owner: agent
44
- position:
45
- - 8
46
- - 16
47
- - type: e1
48
- owner: enemy
49
- position:
50
- - 114
51
- - 34
52
- count: 1
53
- stance: 2
54
- levels:
55
- easy:
56
- description: 'Budget Allocation (spend): easy — the decision is harder (less info / tighter clock
57
- / committed choice).'
58
- win_condition:
59
- all_of:
60
- - building_count_gte:
61
- type: powr
62
- n: 2
63
- - building_total_gte: 5
64
- - within_ticks: 20000
65
- max_turns: 45
66
- medium:
67
- description: 'Budget Allocation (spend): medium — the decision is harder (less info / tighter clock
68
- / committed choice).'
69
- win_condition:
70
- all_of:
71
- - building_count_gte:
72
- type: powr
73
- n: 2
74
- - building_total_gte: 6
75
- - within_ticks: 15500
76
- max_turns: 53
77
- hard:
78
- description: 'Budget Allocation (spend): hard — the decision is harder (less info / tighter clock
79
- / committed choice).'
80
- win_condition:
81
- all_of:
82
- - building_count_gte:
83
- type: powr
84
- n: 2
85
- - building_total_gte: 7
86
- - within_ticks: 11000
87
- max_turns: 61
88
- fail_condition:
89
- units_lost_lte: -1
90
- starting_cash: 1840
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
openra_bench/scenarios/packs/cat-c5-04.yaml DELETED
@@ -1,90 +0,0 @@
1
- meta:
2
- id: cat-c5-04
3
- title: 'Budget Allocation (spend) #5'
4
- capability: reasoning
5
- real_world_meaning: Allocating a finite budget across competing investments where a half-measure on
6
- both fails — resource-allocation under a hard cap.
7
- robotics_analogue: 'Indivisible-budget allocation: wide vs deep, splitting clears nothing.'
8
- author: catalog-generator
9
- status: quarantine
10
- quarantine_reason: 'redundant cat-* over-generation (audit: keep 2 representatives/category)'
11
- base_map: rush-hour-arena
12
- base:
13
- agent:
14
- faction: allies
15
- enemy:
16
- faction: soviet
17
- tools:
18
- - build
19
- - place_building
20
- - move_units
21
- - deploy
22
- - stop
23
- planning: true
24
- termination:
25
- max_ticks: 40000
26
- actors:
27
- - type: fact
28
- owner: agent
29
- position:
30
- - 10
31
- - 18
32
- - type: powr
33
- owner: agent
34
- position:
35
- - 14
36
- - 18
37
- - type: tent
38
- owner: agent
39
- position:
40
- - 12
41
- - 21
42
- - type: jeep
43
- owner: agent
44
- position:
45
- - 8
46
- - 16
47
- - type: e1
48
- owner: enemy
49
- position:
50
- - 114
51
- - 34
52
- count: 1
53
- stance: 2
54
- levels:
55
- easy:
56
- description: 'Budget Allocation (spend): easy — the decision is harder (less info / tighter clock
57
- / committed choice).'
58
- win_condition:
59
- all_of:
60
- - building_count_gte:
61
- type: powr
62
- n: 2
63
- - building_total_gte: 5
64
- - within_ticks: 20000
65
- max_turns: 45
66
- medium:
67
- description: 'Budget Allocation (spend): medium — the decision is harder (less info / tighter clock
68
- / committed choice).'
69
- win_condition:
70
- all_of:
71
- - building_count_gte:
72
- type: powr
73
- n: 2
74
- - building_total_gte: 6
75
- - within_ticks: 15500
76
- max_turns: 53
77
- hard:
78
- description: 'Budget Allocation (spend): hard — the decision is harder (less info / tighter clock
79
- / committed choice).'
80
- win_condition:
81
- all_of:
82
- - building_count_gte:
83
- type: powr
84
- n: 2
85
- - building_total_gte: 7
86
- - within_ticks: 11000
87
- max_turns: 61
88
- fail_condition:
89
- units_lost_lte: -1
90
- starting_cash: 1720
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
openra_bench/scenarios/packs/cat-c5-05.yaml DELETED
@@ -1,90 +0,0 @@
1
- meta:
2
- id: cat-c5-05
3
- title: 'Budget Allocation (spend) #6'
4
- capability: reasoning
5
- real_world_meaning: Allocating a finite budget across competing investments where a half-measure on
6
- both fails — resource-allocation under a hard cap.
7
- robotics_analogue: 'Indivisible-budget allocation: wide vs deep, splitting clears nothing.'
8
- author: catalog-generator
9
- status: quarantine
10
- quarantine_reason: 'redundant cat-* over-generation (audit: keep 2 representatives/category)'
11
- base_map: rush-hour-arena
12
- base:
13
- agent:
14
- faction: allies
15
- enemy:
16
- faction: soviet
17
- tools:
18
- - build
19
- - place_building
20
- - move_units
21
- - deploy
22
- - stop
23
- planning: true
24
- termination:
25
- max_ticks: 40000
26
- actors:
27
- - type: fact
28
- owner: agent
29
- position:
30
- - 10
31
- - 18
32
- - type: powr
33
- owner: agent
34
- position:
35
- - 14
36
- - 18
37
- - type: tent
38
- owner: agent
39
- position:
40
- - 12
41
- - 21
42
- - type: jeep
43
- owner: agent
44
- position:
45
- - 8
46
- - 16
47
- - type: e1
48
- owner: enemy
49
- position:
50
- - 114
51
- - 34
52
- count: 1
53
- stance: 2
54
- levels:
55
- easy:
56
- description: 'Budget Allocation (spend): easy — the decision is harder (less info / tighter clock
57
- / committed choice).'
58
- win_condition:
59
- all_of:
60
- - building_count_gte:
61
- type: powr
62
- n: 2
63
- - building_total_gte: 5
64
- - within_ticks: 20000
65
- max_turns: 45
66
- medium:
67
- description: 'Budget Allocation (spend): medium — the decision is harder (less info / tighter clock
68
- / committed choice).'
69
- win_condition:
70
- all_of:
71
- - building_count_gte:
72
- type: powr
73
- n: 2
74
- - building_total_gte: 6
75
- - within_ticks: 15500
76
- max_turns: 53
77
- hard:
78
- description: 'Budget Allocation (spend): hard — the decision is harder (less info / tighter clock
79
- / committed choice).'
80
- win_condition:
81
- all_of:
82
- - building_count_gte:
83
- type: powr
84
- n: 2
85
- - building_total_gte: 7
86
- - within_ticks: 11000
87
- max_turns: 61
88
- fail_condition:
89
- units_lost_lte: -1
90
- starting_cash: 1600
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
openra_bench/scenarios/packs/cat-c6-00.yaml DELETED
@@ -1,84 +0,0 @@
1
- meta:
2
- id: cat-c6-00
3
- title: 'Time-Boxed Capital Deploy #1'
4
- capability: reasoning
5
- real_world_meaning: "Provisioning under an energy/money budget and a mission deadline\
6
- \ \u2014 how much to commit and when."
7
- robotics_analogue: Convert a budget into fielded force AND standing infra before
8
- a clock.
9
- author: catalog-generator
10
- base_map: rush-hour-arena
11
- base:
12
- agent:
13
- faction: allies
14
- enemy:
15
- faction: soviet
16
- tools:
17
- - build
18
- - place_building
19
- - move_units
20
- - deploy
21
- - stop
22
- planning: true
23
- termination:
24
- max_ticks: 40000
25
- actors:
26
- - type: fact
27
- owner: agent
28
- position:
29
- - 10
30
- - 18
31
- - type: powr
32
- owner: agent
33
- position:
34
- - 14
35
- - 18
36
- - type: tent
37
- owner: agent
38
- position:
39
- - 12
40
- - 21
41
- - type: jeep
42
- owner: agent
43
- position:
44
- - 8
45
- - 16
46
- - type: e1
47
- owner: enemy
48
- position:
49
- - 114
50
- - 34
51
- count: 1
52
- stance: 2
53
- levels:
54
- easy:
55
- description: "Time-Boxed Capital Deploy: easy \u2014 the decision is harder (less\
56
- \ info / tighter clock / committed choice)."
57
- win_condition:
58
- all_of:
59
- - own_units_gte: 4
60
- - building_total_gte: 4
61
- - within_ticks: 16000
62
- max_turns: 45
63
- medium:
64
- description: "Time-Boxed Capital Deploy: medium \u2014 the decision is harder\
65
- \ (less info / tighter clock / committed choice)."
66
- win_condition:
67
- all_of:
68
- - own_units_gte: 5
69
- - building_total_gte: 5
70
- - within_ticks: 12500
71
- max_turns: 53
72
- hard:
73
- description: "Time-Boxed Capital Deploy: hard \u2014 the decision is harder (less\
74
- \ info / tighter clock / committed choice)."
75
- win_condition:
76
- all_of:
77
- - own_units_gte: 6
78
- - building_total_gte: 6
79
- - units_lost_lte: 0
80
- - within_ticks: 9000
81
- max_turns: 61
82
- fail_condition:
83
- units_lost_lte: -1
84
- starting_cash: 2000
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
openra_bench/scenarios/packs/cat-c6-01.yaml DELETED
@@ -1,84 +0,0 @@
1
- meta:
2
- id: cat-c6-01
3
- title: 'Time-Boxed Capital Deploy #2'
4
- capability: reasoning
5
- real_world_meaning: "Provisioning under an energy/money budget and a mission deadline\
6
- \ \u2014 how much to commit and when."
7
- robotics_analogue: Convert a budget into fielded force AND standing infra before
8
- a clock.
9
- author: catalog-generator
10
- base_map: rush-hour-arena
11
- base:
12
- agent:
13
- faction: allies
14
- enemy:
15
- faction: soviet
16
- tools:
17
- - build
18
- - place_building
19
- - move_units
20
- - deploy
21
- - stop
22
- planning: true
23
- termination:
24
- max_ticks: 40000
25
- actors:
26
- - type: fact
27
- owner: agent
28
- position:
29
- - 10
30
- - 18
31
- - type: powr
32
- owner: agent
33
- position:
34
- - 14
35
- - 18
36
- - type: tent
37
- owner: agent
38
- position:
39
- - 12
40
- - 21
41
- - type: jeep
42
- owner: agent
43
- position:
44
- - 8
45
- - 16
46
- - type: e1
47
- owner: enemy
48
- position:
49
- - 114
50
- - 34
51
- count: 1
52
- stance: 2
53
- levels:
54
- easy:
55
- description: "Time-Boxed Capital Deploy: easy \u2014 the decision is harder (less\
56
- \ info / tighter clock / committed choice)."
57
- win_condition:
58
- all_of:
59
- - own_units_gte: 4
60
- - building_total_gte: 4
61
- - within_ticks: 16000
62
- max_turns: 45
63
- medium:
64
- description: "Time-Boxed Capital Deploy: medium \u2014 the decision is harder\
65
- \ (less info / tighter clock / committed choice)."
66
- win_condition:
67
- all_of:
68
- - own_units_gte: 5
69
- - building_total_gte: 5
70
- - within_ticks: 12500
71
- max_turns: 53
72
- hard:
73
- description: "Time-Boxed Capital Deploy: hard \u2014 the decision is harder (less\
74
- \ info / tighter clock / committed choice)."
75
- win_condition:
76
- all_of:
77
- - own_units_gte: 6
78
- - building_total_gte: 6
79
- - units_lost_lte: 0
80
- - within_ticks: 9000
81
- max_turns: 61
82
- fail_condition:
83
- units_lost_lte: -1
84
- starting_cash: 1900
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
openra_bench/scenarios/packs/cat-c6-02.yaml DELETED
@@ -1,85 +0,0 @@
1
- meta:
2
- id: cat-c6-02
3
- title: 'Time-Boxed Capital Deploy #3'
4
- capability: reasoning
5
- real_world_meaning: Provisioning under an energy/money budget and a mission deadline — how much to commit
6
- and when.
7
- robotics_analogue: Convert a budget into fielded force AND standing infra before a clock.
8
- author: catalog-generator
9
- status: quarantine
10
- quarantine_reason: 'redundant cat-* over-generation (audit: keep 2 representatives/category)'
11
- base_map: rush-hour-arena
12
- base:
13
- agent:
14
- faction: allies
15
- enemy:
16
- faction: soviet
17
- tools:
18
- - build
19
- - place_building
20
- - move_units
21
- - deploy
22
- - stop
23
- planning: true
24
- termination:
25
- max_ticks: 40000
26
- actors:
27
- - type: fact
28
- owner: agent
29
- position:
30
- - 10
31
- - 18
32
- - type: powr
33
- owner: agent
34
- position:
35
- - 14
36
- - 18
37
- - type: tent
38
- owner: agent
39
- position:
40
- - 12
41
- - 21
42
- - type: jeep
43
- owner: agent
44
- position:
45
- - 8
46
- - 16
47
- - type: e1
48
- owner: enemy
49
- position:
50
- - 114
51
- - 34
52
- count: 1
53
- stance: 2
54
- levels:
55
- easy:
56
- description: 'Time-Boxed Capital Deploy: easy — the decision is harder (less info / tighter clock
57
- / committed choice).'
58
- win_condition:
59
- all_of:
60
- - own_units_gte: 4
61
- - building_total_gte: 4
62
- - within_ticks: 16000
63
- max_turns: 45
64
- medium:
65
- description: 'Time-Boxed Capital Deploy: medium — the decision is harder (less info / tighter clock
66
- / committed choice).'
67
- win_condition:
68
- all_of:
69
- - own_units_gte: 5
70
- - building_total_gte: 5
71
- - within_ticks: 12500
72
- max_turns: 53
73
- hard:
74
- description: 'Time-Boxed Capital Deploy: hard — the decision is harder (less info / tighter clock
75
- / committed choice).'
76
- win_condition:
77
- all_of:
78
- - own_units_gte: 6
79
- - building_total_gte: 6
80
- - units_lost_lte: 0
81
- - within_ticks: 9000
82
- max_turns: 61
83
- fail_condition:
84
- units_lost_lte: -1
85
- starting_cash: 1800
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
openra_bench/scenarios/packs/cat-c6-03.yaml DELETED
@@ -1,85 +0,0 @@
1
- meta:
2
- id: cat-c6-03
3
- title: 'Time-Boxed Capital Deploy #4'
4
- capability: reasoning
5
- real_world_meaning: Provisioning under an energy/money budget and a mission deadline — how much to commit
6
- and when.
7
- robotics_analogue: Convert a budget into fielded force AND standing infra before a clock.
8
- author: catalog-generator
9
- status: quarantine
10
- quarantine_reason: 'redundant cat-* over-generation (audit: keep 2 representatives/category)'
11
- base_map: rush-hour-arena
12
- base:
13
- agent:
14
- faction: allies
15
- enemy:
16
- faction: soviet
17
- tools:
18
- - build
19
- - place_building
20
- - move_units
21
- - deploy
22
- - stop
23
- planning: true
24
- termination:
25
- max_ticks: 40000
26
- actors:
27
- - type: fact
28
- owner: agent
29
- position:
30
- - 10
31
- - 18
32
- - type: powr
33
- owner: agent
34
- position:
35
- - 14
36
- - 18
37
- - type: tent
38
- owner: agent
39
- position:
40
- - 12
41
- - 21
42
- - type: jeep
43
- owner: agent
44
- position:
45
- - 8
46
- - 16
47
- - type: e1
48
- owner: enemy
49
- position:
50
- - 114
51
- - 34
52
- count: 1
53
- stance: 2
54
- levels:
55
- easy:
56
- description: 'Time-Boxed Capital Deploy: easy — the decision is harder (less info / tighter clock
57
- / committed choice).'
58
- win_condition:
59
- all_of:
60
- - own_units_gte: 4
61
- - building_total_gte: 4
62
- - within_ticks: 16000
63
- max_turns: 45
64
- medium:
65
- description: 'Time-Boxed Capital Deploy: medium — the decision is harder (less info / tighter clock
66
- / committed choice).'
67
- win_condition:
68
- all_of:
69
- - own_units_gte: 5
70
- - building_total_gte: 5
71
- - within_ticks: 12500
72
- max_turns: 53
73
- hard:
74
- description: 'Time-Boxed Capital Deploy: hard — the decision is harder (less info / tighter clock
75
- / committed choice).'
76
- win_condition:
77
- all_of:
78
- - own_units_gte: 6
79
- - building_total_gte: 6
80
- - units_lost_lte: 0
81
- - within_ticks: 9000
82
- max_turns: 61
83
- fail_condition:
84
- units_lost_lte: -1
85
- starting_cash: 1700