==============================================================================
LOG-DAY MODEL + COHEN'S f + POWER -- unmerge_d0_10   [metric: success RATE]
==============================================================================
model: behavior ~ stim + log(day) + stim:log(day) + (1|rat)   (behavior = success RATE)
log(day) uses 1-indexed training day (our day 0 = paper "Day 1")
observed groups: stim n=3, control n=3   nrep=120, alpha=0.05

--- fitted on real data ---
stim x log(day) interaction: F(1,57)=9.174  p(resid)=0.003684  p(Satt)=0.003695 (df=57)
  honest per-animal random slope (log-day): F(1,35.1)=8.84  p=0.005302
Cohen's f (interaction, partial eta^2=0.030) = 0.177  (small-medium; f: .10 small, .25 medium, .40 large)

--- power simulation (log-day ground truth: stim:logday=+0.08869, ratSD=0.04059, resSD=0.08029) ---

 true stim:log(day) = +0.08869 (100% of observed)
  N/group  | per-animal power | LME power
  ------------------------------------------
  3        |  0.54            |  0.85  <- observed
  5        |  0.95            |  0.98
  8        |  1.00            |  1.00
  12       |  1.00            |  1.00
  16       |  1.00            |  1.00
  24       |  1.00            |  1.00

 true stim:log(day) = +0.04434 (50% of observed)
  N/group  | per-animal power | LME power
  ------------------------------------------
  3        |  0.17            |  0.33  <- observed
  5        |  0.40            |  0.52
  8        |  0.69            |  0.68
  12       |  0.89            |  0.90
  16       |  0.95            |  0.96
  24       |  0.99            |  0.99

Read per-animal as the honest power; LME matches the paper's power code (optimistic).