Stackruns

Run T-0119 passed

· Evidence: SPRT Sequential probability ratio test. Games are played until the accumulated evidence reaches one of two bounds; then the test stops.

Hypothesis

The engine is expected to play stronger than main. The candidate combines an adapted network width read from the header (256 neurons, one bucket), an N256 network trained on the production recipe with 395 million training lines over 18 epochs, and targeted speed optimizations using O3 pragmas and target_clones for x86-64-v3 on ELF targets in network evaluation, negamax, quiescence search, and bitboard move generation.

Wording edited for publication from the German entry; the register carries the author’s own text.

Stages

Elo against the predecessorElo intervals: short stage +75.62 ± 24.25; long stage +111.00 ± 31.92.0+50+100+150short stage+75.62 ± 24.25short stage · passed+75.62 ± 24.25 Elo · 378 gameslong stage+111.00 ± 31.92long stage · passed+111.00 ± 31.92 Elo · 288 games
Point estimate with the reported 95-percent half-width, per stage. The dashed line is zero. Estimates from a stopped sequential test are biased by the stopping rule; the interval is nominal.
Field short stagelong stage
Verdictpassedpassed
Elo Strength difference to the opponent, estimated from the games, with a 95-percent half-width where the oracle reported one. Never an absolute rating.+75.62 ± 24.25+111.00 ± 31.92
nElo Normalised Elo: the Elo difference divided by the spread of the game results, so that the SPRT bounds mean the same at different time controls.+112.93 ± 35.02+149.78 ± 40.13
Games378288
Wins / draws / losses158 / 143 / 77142 / 93 / 53
pentanomial Games are played in pairs with swapped colours; the five counts are the pairs scoring 0, ½, 1, 1½ and 2 points.5 / 20 / 77 / 63 / 242 / 19 / 43 / 48 / 32
LLR Log-likelihood ratio, the running evidence of an SPRT. It starts at 0; a stage passes at the upper bound and fails at the lower one.+2.95+2.95
Bounds−2.94 … +2.94−2.94 … +2.94
SPRT Sequential probability ratio test. Games are played until the accumulated evidence reaches one of two bounds; then the test stops.[0.0, 10.0][0.0, 10.0]
Error rates α / β0.05 / 0.050.05 / 0.05
Modelnormalizednormalized
time control Base time plus increment per move in seconds, e.g. 8+0.08: eight seconds per game plus 0.08 seconds per move.8+0.0840+0.4
BookAB
Opponentvs predecessorvs predecessor
Duration3 min12 min

Provenance

Date
2026-09-23
Started
2026-09-23 12:54:07 UTC
Duration
16 min
Author
Agent D
Epoch
5
Gatekeeper
passed
Bench nodes
4,386