Hugging Face's logo

Datasets:
nosuchstudios
/
fruit-fly-brain-runtime

Dataset card Files
xet
 Community
1
nosuchstudios commited on
Commit
debd0a4
·
verified ·
1 Parent(s): ba7fd5f

Publish cx-planner-2026-09-15 (Fruit Fly Brain, 2026-09-15)

Browse files
Files changed (4) hide show
  1. artifacts/cx-planner-2026-09-15/CX_PLANNER.md +149 -0
  2. artifacts/cx-planner-2026-09-15/SHA256SUMS +3 -0
  3. artifacts/cx-planner-2026-09-15/manifest.json +33 -0
  4. artifacts/cx-planner-2026-09-15/report.json +1030 -0
artifacts/cx-planner-2026-09-15/CX_PLANNER.md ADDED
@@ -0,0 +1,149 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ # Central-complex-inspired planner versus exact planners
2
+
3
+ A bounded, preregistered CPU study. The protocol is frozen in
4
+ `hf_space/assets/cx_planner_protocol.json` (`cx-planner-v1`) and is validated
5
+ before any arm runs. The implementation is `hf_space/ffb_playground/cx_planner.py`
6
+ and the runner is `scripts/run_cx_planner_bench.py`.
7
+
8
+ The study asks two questions. First, does a central-complex-inspired
9
+ path-integration planner (a ring-attractor heading module plus an accumulated
10
+ home vector that steers a reactive policy) match exact classical planners on
11
+ exact-truth navigation. Second, does the ordered ring topology contribute
12
+ beyond matched nulls. The primary metric is **path optimality**: per instance,
13
+ `optimal_steps / planner_steps` when the planner reaches home, else `0.0`, where
14
+ `optimal_steps` is the exact breadth-first shortest path.
15
+
16
+ No biological, AGI or programming-expertise claim is made. The ring is a
17
+ declared synthetic attractor; no connectome is used as an input. A null result
18
+ is retained and reported plainly, and a bounded CPU study is not evidence of
19
+ absence.
20
+
21
+ ## Preregistered design
22
+
23
+ - **Tasks.** Deterministic 15x15 gridworlds: `open` (no obstacles), `rooms`
24
+ (independent obstacles at probability 0.22, pruned to the connected component
25
+ of home) and `maze` (recursive-backtracker perfect maze). A smooth heading
26
+ random walk (turn sd 0.08 rad/step, 45 steps) is quantised to the closest free
27
+ grid move from home; the agent must then return home within a budget of
28
+ `8 * grid_size` steps.
29
+ - **Heading observations.** The desired heading plus independent Gaussian noise
30
+ (sd 0.5 rad). Every arm consumes the identical noise draws for a given
31
+ instance, so the contrast is paired on the same trajectory and noise.
32
+ - **Treatment `ring`.** A 36-node von Mises bump on an ordered ring
33
+ (each node drives its 3 neighbours each way, `kappa=4`); the ring is mixed by
34
+ the row-normalised local operator and injected with the noisy observation
35
+ (`alpha=0.15`), and the direction is read out by the population vector. The
36
+ decoded heading integrates the home vector.
37
+ - **Matched nulls.**
38
+ - `no_ring`: the recurrent heading module is removed; the estimate is the raw
39
+ noisy observation.
40
+ - `shuffled_ring`: the ordered ring is relabelled by a uniform random node
41
+ permutation (degrees and weight multiset preserved, locality destroyed).
42
+ - `degree_preserving_ring`: the ring support is rewired with the full directed
43
+ double-edge-swap null of `ffb_playground.graph_nulls` (`5*|E|` swaps) and the
44
+ weight multiset is permuted onto the new edges; in/out degrees and self-loops
45
+ are verified unchanged.
46
+ - **Exact classical planners.** BFS, A* (Manhattan) and value iteration to a
47
+ fixed point, each with full map access. They are the optimality ceiling, not a
48
+ compute-matched arm.
49
+ - **Ground truth.** Every instance's exact shortest path is recomputed with all
50
+ three classical methods and replayed edge by edge; the run aborts on any
51
+ disagreement. All 30 instance checks agreed.
52
+ - **Statistics.** Paired per seed (10 seeds, 3 instances per seed, 30 paired
53
+ instances). Primary contrasts: `ring - BFS`, `ring - A*`,
54
+ `ring - value_iteration`, `ring - no_ring`, `ring - shuffled_ring`,
55
+ `ring - degree_preserving_ring`. Paired percentile-bootstrap 95% intervals and
56
+ exact two-sided sign-flip with a Holm correction over the six contrasts; the
57
+ analytic normal-theory MDE and observed power are reported per contrast.
58
+ - **Admissible outcomes.** `ring-advantage`, `classical-advantage`,
59
+ `null-result`, decided by the preregistered rule in the protocol.
60
+
61
+ ## Results (10 seeds)
62
+
63
+ Per-arm mean path-optimality score:
64
+
65
+ | arm | mean score | success rate | mean steps | outbound heading error (rad) |
66
+ | --- | --- | --- | --- | --- |
67
+ | `ring` | 0.2106 | 0.367 | 90.57 | 0.175 |
68
+ | `no_ring` | 0.2052 | 0.333 | 89.80 | 0.385 |
69
+ | `shuffled_ring` | 0.2137 | 0.300 | 91.83 | 0.349 |
70
+ | `degree_preserving_ring` | 0.2057 | 0.333 | 91.87 | 0.403 |
71
+ | `bfs` | 1.0000 | 1.000 | 8.47 | n/a |
72
+ | `astar` | 1.0000 | 1.000 | 8.47 | n/a |
73
+ | `value_iteration` | 1.0000 | 1.000 | 8.47 | n/a |
74
+
75
+ Paired contrasts (`a - b`):
76
+
77
+ | contrast | role | effect | 95% CI | sign-flip p | Holm p | MDE |
78
+ | --- | --- | --- | --- | --- | --- | --- |
79
+ | `ring - bfs` | exact | -0.7894 | [-0.8852, -0.6954] | 0.0020 | 0.0117 | 0.1423 |
80
+ | `ring - astar` | exact | -0.7894 | [-0.8852, -0.6954] | 0.0020 | 0.0117 | 0.1423 |
81
+ | `ring - value_iteration` | exact | -0.7894 | [-0.8852, -0.6954] | 0.0020 | 0.0117 | 0.1423 |
82
+ | `ring - no_ring` | null | +0.0054 | [-0.0012, +0.0148] | 0.3750 | 1.0000 | 0.0128 |
83
+ | `ring - shuffled_ring` | null | -0.0031 | [-0.0239, +0.0145] | 0.7500 | 1.0000 | 0.0278 |
84
+ | `ring - degree_preserving_ring` | null | +0.0048 | [-0.0011, +0.0145] | 0.3750 | 1.0000 | 0.0129 |
85
+
86
+ **Outcome: `classical-advantage`.** The exact planners beat the ring by a large,
87
+ decisive margin (about 0.79 score, CI excluding zero), and the ring does not beat
88
+ every matched null. The overall study is not a claim that path integration is
89
+ useless: it is a statement that on these exact-truth gridworld tasks a reactive
90
+ home-vector policy is far from exact search.
91
+
92
+ ## Mechanism readout
93
+
94
+ The preregistered secondary readout is heading-estimation error, and it behaves
95
+ as designed: the ordered ring roughly halves the **outbound** heading error
96
+ relative to the memoryless arm (0.175 vs 0.385 rad), and also improves on the
97
+ shuffled (0.349) and degree-preserving (0.403) rings. That mechanism does not
98
+ translate into path optimality: the ring's path-optimality advantage over the
99
+ nulls is inside the equivalence region (`+0.005`, CI including zero, MDE 0.013),
100
+ because the greedy reactive policy is limited by obstacle geometry and by its
101
+ own attractor lag during the return phase (total heading error 0.463 for the
102
+ ring versus 0.391 for the memoryless arm). The honest reading is that better
103
+ outbound heading integration is necessary but not sufficient for better
104
+ vector navigation in this benchmark.
105
+
106
+ ## Power
107
+
108
+ The paired differences are small for the ring-null contrasts (sd 0.015-0.031),
109
+ so the study is well powered for small effects: MDE 0.013-0.028 at 10 seeds. The
110
+ preregistered `reference_sd=0.15` power curve is `MDE 0.133` at 10 seeds,
111
+ `0.094` at 20 and `0.066` at 40; the observed ring-null differences are far
112
+ below these, so the null is informative rather than merely under-powered.
113
+
114
+ ## Non-claims
115
+
116
+ - No biological advantage, AGI or programming-expertise claim is made.
117
+ - The ring is synthetic and untrained; it is not derived from any connectome or
118
+ from measured EPG physiology, and this study does not localise the computation
119
+ to any nervous system.
120
+ - The exact planners have full map access and are not compute-matched to the
121
+ reactive policy; the classical contrast is an upper bound.
122
+ - A null is retained; a bounded CPU study is not evidence of absence.
123
+
124
+ ## Limitations
125
+
126
+ - Ten paired seeds and three instances per seed; the task suite is small and
127
+ deterministic.
128
+ - The gridworld is a coarse discretisation and the policy is a simple greedy
129
+ home-vector steer with wall-following, not a full CX model.
130
+ - Heading observations are synthetic Gaussian noise; no behavioural or odometry
131
+ data is used.
132
+ - The metric penalises longer successful paths, so a planner that reaches home
133
+ but wanders scores low.
134
+ - The single outcome label does not generalise beyond these tasks.
135
+
136
+ ## Reproduction
137
+
138
+ ```bash
139
+ uv run --no-project --with pytest --with numpy --with pyyaml --with scipy \
140
+ python -m pytest tests/test_cx_planner.py -q
141
+
142
+ uv run --no-project --with numpy --with pyyaml --with scipy \
143
+ python scripts/run_cx_planner_bench.py \
144
+ --protocol hf_space/assets/cx_planner_protocol.json \
145
+ --seeds 10 --out /home/user/dev/hf/hf-stage/publish/cx-planner-2026-09-15/report.json
146
+ ```
147
+
148
+ The staged artifact also carries `report.json` (with the protocol SHA-256 and
149
+ design digest), `SHA256SUMS` and `manifest.json`.
artifacts/cx-planner-2026-09-15/SHA256SUMS ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ e40b2678d4d24c1bb0dc11defc461fb24cde8e0a9ebcb47834a4d8479015d2d7 report.json
2
+ cbeb6d21a6beea98397967e1a67eda03b1bdc30a618f5705180a3e0ebbc21d22 CX_PLANNER.md
3
+ 716745286a555a336948597394fd49bcea01949e87e6e5793eec43c959a953d4 manifest.json
artifacts/cx-planner-2026-09-15/manifest.json ADDED
@@ -0,0 +1,33 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "artifact": "cx-planner-2026-09-15",
3
+ "created_utc": "2026-09-16T00:00:00Z",
4
+ "branch": "wt/cxplanner",
5
+ "protocol_id": "cx-planner-v1",
6
+ "protocol_path": "hf_space/assets/cx_planner_protocol.json",
7
+ "protocol_sha256": "4e335e493f5808d46f1e6510a0aa8225abd8c489d28a57355ec9ef9919f46fdc",
8
+ "design_digest": "6a6bd63517cf1ae9131a1f2fe003f71b9245db318d75cc7bef15996e1ca55982",
9
+ "seeds": 10,
10
+ "instances_per_seed": 3,
11
+ "primary_metric": "path-optimality-score",
12
+ "outcome": "classical-advantage",
13
+ "ground_truth_all_agree": true,
14
+ "arms": {
15
+ "ring": 0.21055469717220093,
16
+ "no_ring": 0.20517718023005785,
17
+ "shuffled_ring": 0.21369489723759888,
18
+ "degree_preserving_ring": 0.20571290508476228,
19
+ "bfs": 1.0,
20
+ "astar": 1.0,
21
+ "value_iteration": 1.0
22
+ },
23
+ "files": {
24
+ "report.json": "e40b2678d4d24c1bb0dc11defc461fb24cde8e0a9ebcb47834a4d8479015d2d7",
25
+ "CX_PLANNER.md": "cbeb6d21a6beea98397967e1a67eda03b1bdc30a618f5705180a3e0ebbc21d22"
26
+ },
27
+ "non_claims": [
28
+ "No biological advantage, AGI or programming-expertise claim is made.",
29
+ "No connectome is used as an input; the ring is a declared synthetic attractor.",
30
+ "The exact planners have full map access and are not compute-matched.",
31
+ "A null result is retained; a bounded CPU study is not evidence of absence."
32
+ ]
33
+ }
artifacts/cx-planner-2026-09-15/report.json ADDED
@@ -0,0 +1,1030 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "schema": "cx-planner-report-v1",
3
+ "protocol_id": "cx-planner-v1",
4
+ "seeds": 10,
5
+ "instances_per_seed": 3,
6
+ "base_seed": 20260915,
7
+ "primary_metric": "path-optimality-score",
8
+ "arms": {
9
+ "ring": {
10
+ "mean_score": 0.21055469717220093,
11
+ "per_seed": [
12
+ 0.3333333333333333,
13
+ 0.0,
14
+ 0.04552058111380145,
15
+ 0.3333333333333333,
16
+ 0.33962264150943394,
17
+ 0.3333333333333333,
18
+ 0.0035087719298245615,
19
+ 0.0502283105022831,
20
+ 0.3333333333333333,
21
+ 0.3333333333333333
22
+ ],
23
+ "success_rate": 0.36666666666666664,
24
+ "mean_steps": 90.56666666666666,
25
+ "mean_heading_error": 0.4625837733061393,
26
+ "mean_heading_error_outbound": 0.17469844357483696
27
+ },
28
+ "no_ring": {
29
+ "mean_score": 0.20517718023005785,
30
+ "per_seed": [
31
+ 0.3333333333333333,
32
+ 0.0,
33
+ 0.030927835051546393,
34
+ 0.34108527131782945,
35
+ 0.33962264150943394,
36
+ 0.3333333333333333,
37
+ 0.0,
38
+ 0.006802721088435374,
39
+ 0.3333333333333333,
40
+ 0.3333333333333333
41
+ ],
42
+ "success_rate": 0.3333333333333333,
43
+ "mean_steps": 89.8,
44
+ "mean_heading_error": 0.39116022020728164,
45
+ "mean_heading_error_outbound": 0.38536423871366066
46
+ },
47
+ "shuffled_ring": {
48
+ "mean_score": 0.21369489723759888,
49
+ "per_seed": [
50
+ 0.41314553990610325,
51
+ 0.0,
52
+ 0.05084745762711865,
53
+ 0.3333333333333333,
54
+ 0.33962264150943394,
55
+ 0.3333333333333333,
56
+ 0.0,
57
+ 0.0,
58
+ 0.3333333333333333,
59
+ 0.3333333333333333
60
+ ],
61
+ "success_rate": 0.3,
62
+ "mean_steps": 91.83333333333333,
63
+ "mean_heading_error": 0.3593786578949606,
64
+ "mean_heading_error_outbound": 0.34941995413928134
65
+ },
66
+ "degree_preserving_ring": {
67
+ "mean_score": 0.20571290508476228,
68
+ "per_seed": [
69
+ 0.3333333333333333,
70
+ 0.0,
71
+ 0.043478260869565216,
72
+ 0.33879781420765026,
73
+ 0.3333333333333333,
74
+ 0.3333333333333333,
75
+ 0.0,
76
+ 0.005128205128205128,
77
+ 0.3363914373088685,
78
+ 0.3333333333333333
79
+ ],
80
+ "success_rate": 0.3333333333333333,
81
+ "mean_steps": 91.86666666666666,
82
+ "mean_heading_error": 0.3980021120259829,
83
+ "mean_heading_error_outbound": 0.40252230541561324
84
+ },
85
+ "bfs": {
86
+ "mean_score": 1.0,
87
+ "per_seed": [
88
+ 1.0,
89
+ 1.0,
90
+ 1.0,
91
+ 1.0,
92
+ 1.0,
93
+ 1.0,
94
+ 1.0,
95
+ 1.0,
96
+ 1.0,
97
+ 1.0
98
+ ],
99
+ "success_rate": 1.0,
100
+ "mean_steps": 8.466666666666667,
101
+ "mean_heading_error": 0.0,
102
+ "mean_heading_error_outbound": 0.0
103
+ },
104
+ "astar": {
105
+ "mean_score": 1.0,
106
+ "per_seed": [
107
+ 1.0,
108
+ 1.0,
109
+ 1.0,
110
+ 1.0,
111
+ 1.0,
112
+ 1.0,
113
+ 1.0,
114
+ 1.0,
115
+ 1.0,
116
+ 1.0
117
+ ],
118
+ "success_rate": 1.0,
119
+ "mean_steps": 8.466666666666667,
120
+ "mean_heading_error": 0.0,
121
+ "mean_heading_error_outbound": 0.0
122
+ },
123
+ "value_iteration": {
124
+ "mean_score": 1.0,
125
+ "per_seed": [
126
+ 1.0,
127
+ 1.0,
128
+ 1.0,
129
+ 1.0,
130
+ 1.0,
131
+ 1.0,
132
+ 1.0,
133
+ 1.0,
134
+ 1.0,
135
+ 1.0
136
+ ],
137
+ "success_rate": 1.0,
138
+ "mean_steps": 8.466666666666667,
139
+ "mean_heading_error": 0.0,
140
+ "mean_heading_error_outbound": 0.0
141
+ }
142
+ },
143
+ "ground_truth": {
144
+ "all_agree": true,
145
+ "checks": [
146
+ {
147
+ "task_type": "open",
148
+ "seed": 0,
149
+ "optimal_steps": 17,
150
+ "bfs": 17,
151
+ "astar": 17,
152
+ "value_iteration": 17,
153
+ "bfs_expansions": 138,
154
+ "astar_expansions": 83,
155
+ "value_iteration_sweeps": 20,
156
+ "replay_ok": true,
157
+ "agrees": true
158
+ },
159
+ {
160
+ "task_type": "rooms",
161
+ "seed": 0,
162
+ "optimal_steps": 9,
163
+ "bfs": 9,
164
+ "astar": 9,
165
+ "value_iteration": 9,
166
+ "bfs_expansions": 48,
167
+ "astar_expansions": 22,
168
+ "value_iteration_sweeps": 23,
169
+ "replay_ok": true,
170
+ "agrees": true
171
+ },
172
+ {
173
+ "task_type": "maze",
174
+ "seed": 0,
175
+ "optimal_steps": 23,
176
+ "bfs": 23,
177
+ "astar": 23,
178
+ "value_iteration": 23,
179
+ "bfs_expansions": 72,
180
+ "astar_expansions": 39,
181
+ "value_iteration_sweeps": 63,
182
+ "replay_ok": true,
183
+ "agrees": true
184
+ },
185
+ {
186
+ "task_type": "open",
187
+ "seed": 1,
188
+ "optimal_steps": 7,
189
+ "bfs": 7,
190
+ "astar": 7,
191
+ "value_iteration": 7,
192
+ "bfs_expansions": 28,
193
+ "astar_expansions": 7,
194
+ "value_iteration_sweeps": 22,
195
+ "replay_ok": true,
196
+ "agrees": true
197
+ },
198
+ {
199
+ "task_type": "rooms",
200
+ "seed": 1,
201
+ "optimal_steps": 9,
202
+ "bfs": 9,
203
+ "astar": 9,
204
+ "value_iteration": 9,
205
+ "bfs_expansions": 29,
206
+ "astar_expansions": 12,
207
+ "value_iteration_sweeps": 22,
208
+ "replay_ok": true,
209
+ "agrees": true
210
+ },
211
+ {
212
+ "task_type": "maze",
213
+ "seed": 1,
214
+ "optimal_steps": 1,
215
+ "bfs": 1,
216
+ "astar": 1,
217
+ "value_iteration": 1,
218
+ "bfs_expansions": 1,
219
+ "astar_expansions": 1,
220
+ "value_iteration_sweeps": 77,
221
+ "replay_ok": true,
222
+ "agrees": true
223
+ },
224
+ {
225
+ "task_type": "open",
226
+ "seed": 2,
227
+ "optimal_steps": 9,
228
+ "bfs": 9,
229
+ "astar": 9,
230
+ "value_iteration": 9,
231
+ "bfs_expansions": 39,
232
+ "astar_expansions": 27,
233
+ "value_iteration_sweeps": 20,
234
+ "replay_ok": true,
235
+ "agrees": true
236
+ },
237
+ {
238
+ "task_type": "rooms",
239
+ "seed": 2,
240
+ "optimal_steps": 9,
241
+ "bfs": 9,
242
+ "astar": 9,
243
+ "value_iteration": 9,
244
+ "bfs_expansions": 46,
245
+ "astar_expansions": 20,
246
+ "value_iteration_sweeps": 20,
247
+ "replay_ok": true,
248
+ "agrees": true
249
+ },
250
+ {
251
+ "task_type": "maze",
252
+ "seed": 2,
253
+ "optimal_steps": 3,
254
+ "bfs": 3,
255
+ "astar": 3,
256
+ "value_iteration": 3,
257
+ "bfs_expansions": 5,
258
+ "astar_expansions": 3,
259
+ "value_iteration_sweeps": 81,
260
+ "replay_ok": true,
261
+ "agrees": true
262
+ },
263
+ {
264
+ "task_type": "open",
265
+ "seed": 3,
266
+ "optimal_steps": 7,
267
+ "bfs": 7,
268
+ "astar": 7,
269
+ "value_iteration": 7,
270
+ "bfs_expansions": 40,
271
+ "astar_expansions": 19,
272
+ "value_iteration_sweeps": 21,
273
+ "replay_ok": true,
274
+ "agrees": true
275
+ },
276
+ {
277
+ "task_type": "rooms",
278
+ "seed": 3,
279
+ "optimal_steps": 11,
280
+ "bfs": 11,
281
+ "astar": 11,
282
+ "value_iteration": 11,
283
+ "bfs_expansions": 15,
284
+ "astar_expansions": 16,
285
+ "value_iteration_sweeps": 24,
286
+ "replay_ok": true,
287
+ "agrees": true
288
+ },
289
+ {
290
+ "task_type": "maze",
291
+ "seed": 3,
292
+ "optimal_steps": 1,
293
+ "bfs": 1,
294
+ "astar": 1,
295
+ "value_iteration": 1,
296
+ "bfs_expansions": 1,
297
+ "astar_expansions": 1,
298
+ "value_iteration_sweeps": 65,
299
+ "replay_ok": true,
300
+ "agrees": true
301
+ },
302
+ {
303
+ "task_type": "open",
304
+ "seed": 4,
305
+ "optimal_steps": 13,
306
+ "bfs": 13,
307
+ "astar": 13,
308
+ "value_iteration": 13,
309
+ "bfs_expansions": 79,
310
+ "astar_expansions": 13,
311
+ "value_iteration_sweeps": 28,
312
+ "replay_ok": true,
313
+ "agrees": true
314
+ },
315
+ {
316
+ "task_type": "rooms",
317
+ "seed": 4,
318
+ "optimal_steps": 13,
319
+ "bfs": 13,
320
+ "astar": 13,
321
+ "value_iteration": 13,
322
+ "bfs_expansions": 58,
323
+ "astar_expansions": 24,
324
+ "value_iteration_sweeps": 24,
325
+ "replay_ok": true,
326
+ "agrees": true
327
+ },
328
+ {
329
+ "task_type": "maze",
330
+ "seed": 4,
331
+ "optimal_steps": 1,
332
+ "bfs": 1,
333
+ "astar": 1,
334
+ "value_iteration": 1,
335
+ "bfs_expansions": 1,
336
+ "astar_expansions": 1,
337
+ "value_iteration_sweeps": 79,
338
+ "replay_ok": true,
339
+ "agrees": true
340
+ },
341
+ {
342
+ "task_type": "open",
343
+ "seed": 5,
344
+ "optimal_steps": 13,
345
+ "bfs": 13,
346
+ "astar": 13,
347
+ "value_iteration": 13,
348
+ "bfs_expansions": 102,
349
+ "astar_expansions": 25,
350
+ "value_iteration_sweeps": 27,
351
+ "replay_ok": true,
352
+ "agrees": true
353
+ },
354
+ {
355
+ "task_type": "rooms",
356
+ "seed": 5,
357
+ "optimal_steps": 17,
358
+ "bfs": 17,
359
+ "astar": 17,
360
+ "value_iteration": 17,
361
+ "bfs_expansions": 107,
362
+ "astar_expansions": 35,
363
+ "value_iteration_sweeps": 25,
364
+ "replay_ok": true,
365
+ "agrees": true
366
+ },
367
+ {
368
+ "task_type": "maze",
369
+ "seed": 5,
370
+ "optimal_steps": 1,
371
+ "bfs": 1,
372
+ "astar": 1,
373
+ "value_iteration": 1,
374
+ "bfs_expansions": 1,
375
+ "astar_expansions": 1,
376
+ "value_iteration_sweeps": 61,
377
+ "replay_ok": true,
378
+ "agrees": true
379
+ },
380
+ {
381
+ "task_type": "open",
382
+ "seed": 6,
383
+ "optimal_steps": 13,
384
+ "bfs": 13,
385
+ "astar": 13,
386
+ "value_iteration": 13,
387
+ "bfs_expansions": 91,
388
+ "astar_expansions": 25,
389
+ "value_iteration_sweeps": 25,
390
+ "replay_ok": true,
391
+ "agrees": true
392
+ },
393
+ {
394
+ "task_type": "rooms",
395
+ "seed": 6,
396
+ "optimal_steps": 9,
397
+ "bfs": 9,
398
+ "astar": 9,
399
+ "value_iteration": 9,
400
+ "bfs_expansions": 45,
401
+ "astar_expansions": 23,
402
+ "value_iteration_sweeps": 16,
403
+ "replay_ok": true,
404
+ "agrees": true
405
+ },
406
+ {
407
+ "task_type": "maze",
408
+ "seed": 6,
409
+ "optimal_steps": 1,
410
+ "bfs": 1,
411
+ "astar": 1,
412
+ "value_iteration": 1,
413
+ "bfs_expansions": 1,
414
+ "astar_expansions": 1,
415
+ "value_iteration_sweeps": 83,
416
+ "replay_ok": true,
417
+ "agrees": true
418
+ },
419
+ {
420
+ "task_type": "open",
421
+ "seed": 7,
422
+ "optimal_steps": 15,
423
+ "bfs": 15,
424
+ "astar": 15,
425
+ "value_iteration": 15,
426
+ "bfs_expansions": 142,
427
+ "astar_expansions": 59,
428
+ "value_iteration_sweeps": 24,
429
+ "replay_ok": true,
430
+ "agrees": true
431
+ },
432
+ {
433
+ "task_type": "rooms",
434
+ "seed": 7,
435
+ "optimal_steps": 11,
436
+ "bfs": 11,
437
+ "astar": 11,
438
+ "value_iteration": 11,
439
+ "bfs_expansions": 46,
440
+ "astar_expansions": 28,
441
+ "value_iteration_sweeps": 21,
442
+ "replay_ok": true,
443
+ "agrees": true
444
+ },
445
+ {
446
+ "task_type": "maze",
447
+ "seed": 7,
448
+ "optimal_steps": 1,
449
+ "bfs": 1,
450
+ "astar": 1,
451
+ "value_iteration": 1,
452
+ "bfs_expansions": 1,
453
+ "astar_expansions": 1,
454
+ "value_iteration_sweeps": 73,
455
+ "replay_ok": true,
456
+ "agrees": true
457
+ },
458
+ {
459
+ "task_type": "open",
460
+ "seed": 8,
461
+ "optimal_steps": 3,
462
+ "bfs": 3,
463
+ "astar": 3,
464
+ "value_iteration": 3,
465
+ "bfs_expansions": 5,
466
+ "astar_expansions": 5,
467
+ "value_iteration_sweeps": 27,
468
+ "replay_ok": true,
469
+ "agrees": true
470
+ },
471
+ {
472
+ "task_type": "rooms",
473
+ "seed": 8,
474
+ "optimal_steps": 3,
475
+ "bfs": 3,
476
+ "astar": 3,
477
+ "value_iteration": 3,
478
+ "bfs_expansions": 5,
479
+ "astar_expansions": 4,
480
+ "value_iteration_sweeps": 26,
481
+ "replay_ok": true,
482
+ "agrees": true
483
+ },
484
+ {
485
+ "task_type": "maze",
486
+ "seed": 8,
487
+ "optimal_steps": 1,
488
+ "bfs": 1,
489
+ "astar": 1,
490
+ "value_iteration": 1,
491
+ "bfs_expansions": 1,
492
+ "astar_expansions": 1,
493
+ "value_iteration_sweeps": 71,
494
+ "replay_ok": true,
495
+ "agrees": true
496
+ },
497
+ {
498
+ "task_type": "open",
499
+ "seed": 9,
500
+ "optimal_steps": 15,
501
+ "bfs": 15,
502
+ "astar": 15,
503
+ "value_iteration": 15,
504
+ "bfs_expansions": 122,
505
+ "astar_expansions": 51,
506
+ "value_iteration_sweeps": 23,
507
+ "replay_ok": true,
508
+ "agrees": true
509
+ },
510
+ {
511
+ "task_type": "rooms",
512
+ "seed": 9,
513
+ "optimal_steps": 17,
514
+ "bfs": 17,
515
+ "astar": 17,
516
+ "value_iteration": 17,
517
+ "bfs_expansions": 62,
518
+ "astar_expansions": 27,
519
+ "value_iteration_sweeps": 23,
520
+ "replay_ok": true,
521
+ "agrees": true
522
+ },
523
+ {
524
+ "task_type": "maze",
525
+ "seed": 9,
526
+ "optimal_steps": 1,
527
+ "bfs": 1,
528
+ "astar": 1,
529
+ "value_iteration": 1,
530
+ "bfs_expansions": 1,
531
+ "astar_expansions": 1,
532
+ "value_iteration_sweeps": 67,
533
+ "replay_ok": true,
534
+ "agrees": true
535
+ }
536
+ ]
537
+ },
538
+ "contrasts": [
539
+ {
540
+ "schema": "ffb-evaluation-v1",
541
+ "kind": "paired-report",
542
+ "name": "ring_minus_bfs",
543
+ "effect": -0.7894453028277991,
544
+ "mean": -0.7894453028277991,
545
+ "ci_level": 0.95,
546
+ "ci_low": -0.8851712983410465,
547
+ "ci_high": -0.6954479418886199,
548
+ "includes_zero": false,
549
+ "sign_flip_p": 0.001953125,
550
+ "sign_flip_method": "exact",
551
+ "holm_p": 0.01171875,
552
+ "n": 10,
553
+ "resamples": 100000,
554
+ "seed": 707,
555
+ "per_seed_differences": [
556
+ -0.6666666666666667,
557
+ -1.0,
558
+ -0.9544794188861986,
559
+ -0.6666666666666667,
560
+ -0.6603773584905661,
561
+ -0.6666666666666667,
562
+ -0.9964912280701754,
563
+ -0.9497716894977168,
564
+ -0.6666666666666667,
565
+ -0.6666666666666667
566
+ ],
567
+ "a": "ring",
568
+ "b": "bfs",
569
+ "role": "exact-reference",
570
+ "mean_a": 0.21055469717220093,
571
+ "mean_b": 1.0,
572
+ "sd": 0.1606146551809073,
573
+ "mde": 0.1422947925904716,
574
+ "observed_power": 1.0
575
+ },
576
+ {
577
+ "schema": "ffb-evaluation-v1",
578
+ "kind": "paired-report",
579
+ "name": "ring_minus_astar",
580
+ "effect": -0.7894453028277991,
581
+ "mean": -0.7894453028277991,
582
+ "ci_level": 0.95,
583
+ "ci_low": -0.8851712983410465,
584
+ "ci_high": -0.6954479418886199,
585
+ "includes_zero": false,
586
+ "sign_flip_p": 0.001953125,
587
+ "sign_flip_method": "exact",
588
+ "holm_p": 0.01171875,
589
+ "n": 10,
590
+ "resamples": 100000,
591
+ "seed": 707,
592
+ "per_seed_differences": [
593
+ -0.6666666666666667,
594
+ -1.0,
595
+ -0.9544794188861986,
596
+ -0.6666666666666667,
597
+ -0.6603773584905661,
598
+ -0.6666666666666667,
599
+ -0.9964912280701754,
600
+ -0.9497716894977168,
601
+ -0.6666666666666667,
602
+ -0.6666666666666667
603
+ ],
604
+ "a": "ring",
605
+ "b": "astar",
606
+ "role": "exact-reference",
607
+ "mean_a": 0.21055469717220093,
608
+ "mean_b": 1.0,
609
+ "sd": 0.1606146551809073,
610
+ "mde": 0.1422947925904716,
611
+ "observed_power": 1.0
612
+ },
613
+ {
614
+ "schema": "ffb-evaluation-v1",
615
+ "kind": "paired-report",
616
+ "name": "ring_minus_value_iteration",
617
+ "effect": -0.7894453028277991,
618
+ "mean": -0.7894453028277991,
619
+ "ci_level": 0.95,
620
+ "ci_low": -0.8851712983410465,
621
+ "ci_high": -0.6954479418886199,
622
+ "includes_zero": false,
623
+ "sign_flip_p": 0.001953125,
624
+ "sign_flip_method": "exact",
625
+ "holm_p": 0.01171875,
626
+ "n": 10,
627
+ "resamples": 100000,
628
+ "seed": 707,
629
+ "per_seed_differences": [
630
+ -0.6666666666666667,
631
+ -1.0,
632
+ -0.9544794188861986,
633
+ -0.6666666666666667,
634
+ -0.6603773584905661,
635
+ -0.6666666666666667,
636
+ -0.9964912280701754,
637
+ -0.9497716894977168,
638
+ -0.6666666666666667,
639
+ -0.6666666666666667
640
+ ],
641
+ "a": "ring",
642
+ "b": "value_iteration",
643
+ "role": "exact-reference",
644
+ "mean_a": 0.21055469717220093,
645
+ "mean_b": 1.0,
646
+ "sd": 0.1606146551809073,
647
+ "mde": 0.1422947925904716,
648
+ "observed_power": 1.0
649
+ },
650
+ {
651
+ "schema": "ffb-evaluation-v1",
652
+ "kind": "paired-report",
653
+ "name": "ring_minus_no_ring",
654
+ "effect": 0.005377516942143121,
655
+ "mean": 0.005377516942143121,
656
+ "ci_level": 0.95,
657
+ "ci_low": -0.0011995104039167713,
658
+ "ci_high": 0.014837828623362281,
659
+ "includes_zero": true,
660
+ "sign_flip_p": 0.375,
661
+ "sign_flip_method": "exact",
662
+ "holm_p": 1.0,
663
+ "n": 10,
664
+ "resamples": 100000,
665
+ "seed": 707,
666
+ "per_seed_differences": [
667
+ 0.0,
668
+ 0.0,
669
+ 0.014592746062255059,
670
+ -0.007751937984496138,
671
+ 0.0,
672
+ 0.0,
673
+ 0.0035087719298245615,
674
+ 0.04342558941384773,
675
+ 0.0,
676
+ 0.0
677
+ ],
678
+ "a": "ring",
679
+ "b": "no_ring",
680
+ "role": "ring-null",
681
+ "mean_a": 0.21055469717220093,
682
+ "mean_b": 0.20517718023005785,
683
+ "sd": 0.014460510179712734,
684
+ "mde": 0.012811130431759164,
685
+ "observed_power": 0.21737987612167953
686
+ },
687
+ {
688
+ "schema": "ffb-evaluation-v1",
689
+ "kind": "paired-report",
690
+ "name": "ring_minus_shuffled_ring",
691
+ "effect": -0.003140200065397948,
692
+ "mean": -0.003140200065397948,
693
+ "ci_level": 0.95,
694
+ "ci_low": -0.023943661971830982,
695
+ "ci_high": 0.01453580549935321,
696
+ "includes_zero": true,
697
+ "sign_flip_p": 0.75,
698
+ "sign_flip_method": "exact",
699
+ "holm_p": 1.0,
700
+ "n": 10,
701
+ "resamples": 100000,
702
+ "seed": 707,
703
+ "per_seed_differences": [
704
+ -0.07981220657276994,
705
+ 0.0,
706
+ -0.0053268765133171955,
707
+ 0.0,
708
+ 0.0,
709
+ 0.0,
710
+ 0.0035087719298245615,
711
+ 0.0502283105022831,
712
+ 0.0,
713
+ 0.0
714
+ ],
715
+ "a": "ring",
716
+ "b": "shuffled_ring",
717
+ "role": "ring-null",
718
+ "mean_a": 0.21055469717220093,
719
+ "mean_b": 0.21369489723759888,
720
+ "sd": 0.03133147161714678,
721
+ "mde": 0.027757773724287886,
722
+ "observed_power": 0.0615849370057619
723
+ },
724
+ {
725
+ "schema": "ffb-evaluation-v1",
726
+ "kind": "paired-report",
727
+ "name": "ring_minus_degree_preserving_ring",
728
+ "effect": 0.004841792087438726,
729
+ "mean": 0.004841792087438726,
730
+ "ci_level": 0.95,
731
+ "ci_low": -0.0011130020871097945,
732
+ "ci_high": 0.014465848105550804,
733
+ "includes_zero": true,
734
+ "sign_flip_p": 0.375,
735
+ "sign_flip_method": "exact",
736
+ "holm_p": 1.0,
737
+ "n": 10,
738
+ "resamples": 100000,
739
+ "seed": 707,
740
+ "per_seed_differences": [
741
+ 0.0,
742
+ 0.0,
743
+ 0.0020423202442362356,
744
+ -0.005464480874316946,
745
+ 0.0062893081761006275,
746
+ 0.0,
747
+ 0.0035087719298245615,
748
+ 0.04510010537407797,
749
+ -0.003058103975535187,
750
+ 0.0
751
+ ],
752
+ "a": "ring",
753
+ "b": "degree_preserving_ring",
754
+ "role": "ring-null",
755
+ "mean_a": 0.21055469717220093,
756
+ "mean_b": 0.20571290508476228,
757
+ "sd": 0.014509921890798118,
758
+ "mde": 0.012854906195387433,
759
+ "observed_power": 0.1840833971177256
760
+ }
761
+ ],
762
+ "outcome": "classical-advantage",
763
+ "outcome_reason": "at least one exact planner beats the ring with a 95% interval below zero and the ring does not beat every matched null",
764
+ "non_claims": [
765
+ "No biological, AGI or programming-expertise claim is made.",
766
+ "The planner does not use any connectome as an input; the ring is a declared synthetic attractor and this study connects the concept to planning, it does not localise it to a nervous system.",
767
+ "The classical planners are exact references with full map access and are not compute-matched to the reactive policy.",
768
+ "A null result is retained and reported plainly; a bounded CPU study is not evidence of absence."
769
+ ],
770
+ "design_digest": "6a6bd63517cf1ae9131a1f2fe003f71b9245db318d75cc7bef15996e1ca55982",
771
+ "protocol_sha256": "4e335e493f5808d46f1e6510a0aa8225abd8c489d28a57355ec9ef9919f46fdc",
772
+ "contrasts_by_name": {
773
+ "ring_minus_bfs": {
774
+ "schema": "ffb-evaluation-v1",
775
+ "kind": "paired-report",
776
+ "name": "ring_minus_bfs",
777
+ "effect": -0.7894453028277991,
778
+ "mean": -0.7894453028277991,
779
+ "ci_level": 0.95,
780
+ "ci_low": -0.8851712983410465,
781
+ "ci_high": -0.6954479418886199,
782
+ "includes_zero": false,
783
+ "sign_flip_p": 0.001953125,
784
+ "sign_flip_method": "exact",
785
+ "holm_p": 0.01171875,
786
+ "n": 10,
787
+ "resamples": 100000,
788
+ "seed": 707,
789
+ "per_seed_differences": [
790
+ -0.6666666666666667,
791
+ -1.0,
792
+ -0.9544794188861986,
793
+ -0.6666666666666667,
794
+ -0.6603773584905661,
795
+ -0.6666666666666667,
796
+ -0.9964912280701754,
797
+ -0.9497716894977168,
798
+ -0.6666666666666667,
799
+ -0.6666666666666667
800
+ ],
801
+ "a": "ring",
802
+ "b": "bfs",
803
+ "role": "exact-reference",
804
+ "mean_a": 0.21055469717220093,
805
+ "mean_b": 1.0,
806
+ "sd": 0.1606146551809073,
807
+ "mde": 0.1422947925904716,
808
+ "observed_power": 1.0
809
+ },
810
+ "ring_minus_astar": {
811
+ "schema": "ffb-evaluation-v1",
812
+ "kind": "paired-report",
813
+ "name": "ring_minus_astar",
814
+ "effect": -0.7894453028277991,
815
+ "mean": -0.7894453028277991,
816
+ "ci_level": 0.95,
817
+ "ci_low": -0.8851712983410465,
818
+ "ci_high": -0.6954479418886199,
819
+ "includes_zero": false,
820
+ "sign_flip_p": 0.001953125,
821
+ "sign_flip_method": "exact",
822
+ "holm_p": 0.01171875,
823
+ "n": 10,
824
+ "resamples": 100000,
825
+ "seed": 707,
826
+ "per_seed_differences": [
827
+ -0.6666666666666667,
828
+ -1.0,
829
+ -0.9544794188861986,
830
+ -0.6666666666666667,
831
+ -0.6603773584905661,
832
+ -0.6666666666666667,
833
+ -0.9964912280701754,
834
+ -0.9497716894977168,
835
+ -0.6666666666666667,
836
+ -0.6666666666666667
837
+ ],
838
+ "a": "ring",
839
+ "b": "astar",
840
+ "role": "exact-reference",
841
+ "mean_a": 0.21055469717220093,
842
+ "mean_b": 1.0,
843
+ "sd": 0.1606146551809073,
844
+ "mde": 0.1422947925904716,
845
+ "observed_power": 1.0
846
+ },
847
+ "ring_minus_value_iteration": {
848
+ "schema": "ffb-evaluation-v1",
849
+ "kind": "paired-report",
850
+ "name": "ring_minus_value_iteration",
851
+ "effect": -0.7894453028277991,
852
+ "mean": -0.7894453028277991,
853
+ "ci_level": 0.95,
854
+ "ci_low": -0.8851712983410465,
855
+ "ci_high": -0.6954479418886199,
856
+ "includes_zero": false,
857
+ "sign_flip_p": 0.001953125,
858
+ "sign_flip_method": "exact",
859
+ "holm_p": 0.01171875,
860
+ "n": 10,
861
+ "resamples": 100000,
862
+ "seed": 707,
863
+ "per_seed_differences": [
864
+ -0.6666666666666667,
865
+ -1.0,
866
+ -0.9544794188861986,
867
+ -0.6666666666666667,
868
+ -0.6603773584905661,
869
+ -0.6666666666666667,
870
+ -0.9964912280701754,
871
+ -0.9497716894977168,
872
+ -0.6666666666666667,
873
+ -0.6666666666666667
874
+ ],
875
+ "a": "ring",
876
+ "b": "value_iteration",
877
+ "role": "exact-reference",
878
+ "mean_a": 0.21055469717220093,
879
+ "mean_b": 1.0,
880
+ "sd": 0.1606146551809073,
881
+ "mde": 0.1422947925904716,
882
+ "observed_power": 1.0
883
+ },
884
+ "ring_minus_no_ring": {
885
+ "schema": "ffb-evaluation-v1",
886
+ "kind": "paired-report",
887
+ "name": "ring_minus_no_ring",
888
+ "effect": 0.005377516942143121,
889
+ "mean": 0.005377516942143121,
890
+ "ci_level": 0.95,
891
+ "ci_low": -0.0011995104039167713,
892
+ "ci_high": 0.014837828623362281,
893
+ "includes_zero": true,
894
+ "sign_flip_p": 0.375,
895
+ "sign_flip_method": "exact",
896
+ "holm_p": 1.0,
897
+ "n": 10,
898
+ "resamples": 100000,
899
+ "seed": 707,
900
+ "per_seed_differences": [
901
+ 0.0,
902
+ 0.0,
903
+ 0.014592746062255059,
904
+ -0.007751937984496138,
905
+ 0.0,
906
+ 0.0,
907
+ 0.0035087719298245615,
908
+ 0.04342558941384773,
909
+ 0.0,
910
+ 0.0
911
+ ],
912
+ "a": "ring",
913
+ "b": "no_ring",
914
+ "role": "ring-null",
915
+ "mean_a": 0.21055469717220093,
916
+ "mean_b": 0.20517718023005785,
917
+ "sd": 0.014460510179712734,
918
+ "mde": 0.012811130431759164,
919
+ "observed_power": 0.21737987612167953
920
+ },
921
+ "ring_minus_shuffled_ring": {
922
+ "schema": "ffb-evaluation-v1",
923
+ "kind": "paired-report",
924
+ "name": "ring_minus_shuffled_ring",
925
+ "effect": -0.003140200065397948,
926
+ "mean": -0.003140200065397948,
927
+ "ci_level": 0.95,
928
+ "ci_low": -0.023943661971830982,
929
+ "ci_high": 0.01453580549935321,
930
+ "includes_zero": true,
931
+ "sign_flip_p": 0.75,
932
+ "sign_flip_method": "exact",
933
+ "holm_p": 1.0,
934
+ "n": 10,
935
+ "resamples": 100000,
936
+ "seed": 707,
937
+ "per_seed_differences": [
938
+ -0.07981220657276994,
939
+ 0.0,
940
+ -0.0053268765133171955,
941
+ 0.0,
942
+ 0.0,
943
+ 0.0,
944
+ 0.0035087719298245615,
945
+ 0.0502283105022831,
946
+ 0.0,
947
+ 0.0
948
+ ],
949
+ "a": "ring",
950
+ "b": "shuffled_ring",
951
+ "role": "ring-null",
952
+ "mean_a": 0.21055469717220093,
953
+ "mean_b": 0.21369489723759888,
954
+ "sd": 0.03133147161714678,
955
+ "mde": 0.027757773724287886,
956
+ "observed_power": 0.0615849370057619
957
+ },
958
+ "ring_minus_degree_preserving_ring": {
959
+ "schema": "ffb-evaluation-v1",
960
+ "kind": "paired-report",
961
+ "name": "ring_minus_degree_preserving_ring",
962
+ "effect": 0.004841792087438726,
963
+ "mean": 0.004841792087438726,
964
+ "ci_level": 0.95,
965
+ "ci_low": -0.0011130020871097945,
966
+ "ci_high": 0.014465848105550804,
967
+ "includes_zero": true,
968
+ "sign_flip_p": 0.375,
969
+ "sign_flip_method": "exact",
970
+ "holm_p": 1.0,
971
+ "n": 10,
972
+ "resamples": 100000,
973
+ "seed": 707,
974
+ "per_seed_differences": [
975
+ 0.0,
976
+ 0.0,
977
+ 0.0020423202442362356,
978
+ -0.005464480874316946,
979
+ 0.0062893081761006275,
980
+ 0.0,
981
+ 0.0035087719298245615,
982
+ 0.04510010537407797,
983
+ -0.003058103975535187,
984
+ 0.0
985
+ ],
986
+ "a": "ring",
987
+ "b": "degree_preserving_ring",
988
+ "role": "ring-null",
989
+ "mean_a": 0.21055469717220093,
990
+ "mean_b": 0.20571290508476228,
991
+ "sd": 0.014509921890798118,
992
+ "mde": 0.012854906195387433,
993
+ "observed_power": 0.1840833971177256
994
+ }
995
+ },
996
+ "power_plan": {
997
+ "schema": "ffb-evaluation-v1",
998
+ "kind": "power-curve",
999
+ "sd": 0.15,
1000
+ "alpha": 0.05,
1001
+ "power": 0.8,
1002
+ "reference_standardized_effect": 0.5,
1003
+ "reference_effect": 0.075,
1004
+ "n_grid": [
1005
+ 10,
1006
+ 20,
1007
+ 40
1008
+ ],
1009
+ "curve": [
1010
+ {
1011
+ "n": 10,
1012
+ "mde": 0.1328908552244489,
1013
+ "reference_effect": 0.075,
1014
+ "power_at_reference": 0.35260808244470265
1015
+ },
1016
+ {
1017
+ "n": 20,
1018
+ "mde": 0.09396802488688756,
1019
+ "reference_effect": 0.075,
1020
+ "power_at_reference": 0.6087794846454568
1021
+ },
1022
+ {
1023
+ "n": 40,
1024
+ "mde": 0.06644542761222445,
1025
+ "reference_effect": 0.075,
1026
+ "power_at_reference": 0.8853791407623515
1027
+ }
1028
+ ]
1029
+ }
1030
+ }