vvsandClaude Opus 5
2498c66358
A with the fixed metric: still PASSES A1 and A2 in both configurations - flat 0.26 -> 2.82 (proposed) and 0.35 -> 0.77 (reading 3), canonical 6.32 -> 8.49 and 9.00 -> 10.44 - but a correction to the first report: for reading 3 on the flat families step A (0.77) beats our step and NOT bisection (0.99); the first run said 1.16. B (outcome memory) FAILS its falsifier: flat 0.31 -> 0.97 and 0.58 -> 0.82 (B1 holds) but canonical 4.74 -> 2.49 and 5.52 -> 2.12 (B2 fails), bisecting on 66% and 53% of canonical states. Mechanism: the known trap - a good step can collapse the error without moving the bracket, so an outcome measured in bracket shrink condemns it; the same trap that sank the bracket-progress gate (4449) and the bracket guard (7260). Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
31 lines
1.6 KiB
Plaintext
31 lines
1.6 KiB
Plaintext
######## A (fixed metric)
|
|
|
|
== proposed (projection)
|
|
group states S>0 A lands out paired n ours step A bisect (bracket halvings per eval)
|
|
APS 1408 48.9% 23.1% 530 6.32 8.49 1.07
|
|
CHA fun1 51 64.7% 33.3% 22 6.48 8.14 1.00
|
|
CHA flat fun3-6 1640 87.1% 6.4% 1337 0.26 2.82 1.00
|
|
CHA other 177 62.1% 7.3% 102 3.74 5.20 1.00
|
|
|
|
== reading (3), as for m>=2
|
|
group states S>0 A lands out paired n ours step A bisect (bracket halvings per eval)
|
|
APS 1103 51.3% 21.6% 444 9.00 10.44 1.09
|
|
CHA fun1 37 51.4% 15.8% 16 6.71 10.89 1.00
|
|
CHA flat fun3-6 620 68.2% 28.1% 304 0.35 0.77 0.99
|
|
CHA other 129 52.7% 14.7% 58 6.31 8.30 1.00
|
|
######## B (fixed metric)
|
|
|
|
== proposed (projection)
|
|
group states B bisects ours B (bracket halvings per eval, all states with a decision)
|
|
APS 1258 65.6% 4.74 2.49
|
|
CHA fun1 46 60.9% 4.74 2.12
|
|
CHA flat fun3-6 1623 83.6% 0.31 0.97
|
|
CHA other 165 73.3% 3.91 1.60
|
|
|
|
== reading (3), as for m>=2
|
|
group states B bisects ours B (bracket halvings per eval, all states with a decision)
|
|
APS 958 52.6% 5.52 2.12
|
|
CHA fun1 32 43.8% 4.52 1.31
|
|
CHA flat fun3-6 603 48.3% 0.58 0.82
|
|
CHA other 120 46.7% 5.22 2.03
|