# session /tmp/ab_headon
# commit=a82c864c6075d8325589681bae23b371e3c4f1b5 binary_sha256=9d20f530d633982b30735d5ece38da09b707737ae1c57238ee41abe928276601 rounds=7 runs=15 conc=8 ts=2026-09-24T23:45:01+02:00

ARM SUMMARY
arm            runs  dmg/run dmgtk/run    wins   win% shots/run hitstk/run
--------------------------------------------------------------------------
control          15      279       211  48/105   45.7       785       92.9
headon           15       14       228   0/105    0.0       580       72.1

PER-RUN (never just the mean)
  control        dmg:  r1=263 r2=307 r3=294 r4=221 r5=290 r6=235 r7=327 r8=262 r9=330 r10=312 r11=274 r12=292 r13=288 r14=281 r15=217
                 wins: r1=2/7 r2=3/7 r3=4/7 r4=2/7 r5=4/7 r6=3/7 r7=3/7 r8=2/7 r9=6/7 r10=5/7 r11=2/7 r12=4/7 r13=2/7 r14=4/7 r15=2/7
  headon         dmg:  r1=4 r2=8 r3=4 r4=17 r5=14 r6=10 r7=28 r8=18 r9=0 r10=36 r11=17 r12=8 r13=14 r14=14 r15=15
                 wins: r1=0/7 r2=0/7 r3=0/7 r4=0/7 r5=0/7 r6=0/7 r7=0/7 r8=0/7 r9=0/7 r10=0/7 r11=0/7 r12=0/7 r13=0/7 r14=0/7 r15=0/7

PAIRWISE PERMUTATION TEST (per-run values) + MANN-WHITNEY CROSS-CHECK
permutation: exact when C(n,na) <= 20,000,000; otherwise Monte-Carlo 1,000,000 draws, seed=0x5eed5eed, p = (cnt+1)/(B+1), se = sqrt(p(1-p)/(B+1))
metric      A           B             diff(A-B)    perm p method         MC se      MW p     MW U
-------------------------------------------------------------------------------------------------
dmg/run     control     headon         +265.670    0.0000 MC/B=1,000,000  0.0000    0.0000      0.0
round wins  control     headon           +3.200    0.0000 MC/B=1,000,000  0.0000    0.0000      0.0

MINIMUM DETECTABLE EFFECT (two-sample, alpha=0.05 two-sided, 80% power; MDE = 2.8016*sd*sqrt(2/n))
metric       n/arm  sd(control)   MDE(abs)    MDE vs control mean
----------------------------------------------------------------
dmg/run         15       34.956     35.760         12.8% of 279.5
round wins      15        1.265      1.294           40.4% of 3.2

ROUND-LEVEL TEST (pooled rounds, Fisher exact) vs `control` — ANTI-CONSERVATIVE: rounds cluster within runs
arm              ref wins   arm wins        p
----------------------------------------------
headon             48/105      0/105   0.0000

LIVENESS (arm env applied in the bot's own boot report)
  control        OK   (15/15 runs: no arm env; report present)
  headon         OK   (15/15 runs: TR_RACK_PATTERN=off TR_RACK_HEADON=both applied)

[bb] APPLIED-SHIFT CHECK (from bot stdout; needs TR_BITBRAIN_LOG=1). A provably-zero placebo emits ZERO [bb] lines.
  arm             runs w/log   lines       min       max  zeros
  control           0/15         0         -         -      -
  headon            0/15         0         -         -      -

ROUND-WIN ATTRIBUTION (events primary; score tie-break for mutual-kill / timeout rounds)
  control        wins==firstPlaces 15/15 runs OK; single-death rounds agree with score 102/102 (3 tie-broken)
  headon         wins==firstPlaces 15/15 runs OK; single-death rounds agree with score 105/105 (0 tie-broken)
