==============================================================================
LOG-DAY MODEL + COHEN'S f + POWER -- naive_a2_d6_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=6   nrep=120, alpha=0.05

--- fitted on real data ---
stim x log(day) interaction: F(1,41)=2.375  p(resid)=0.131  p(Satt)=0.1321 (df=36)
  honest per-animal random slope (log-day): F(1,9.0)=1.58  p=0.2398
Cohen's f (interaction, partial eta^2=0.023) = 0.153  (small-medium; f: .10 small, .25 medium, .40 large)

--- power simulation (log-day ground truth: stim:logday=-0.2002, ratSD=0.06161, resSD=0.06563) ---

 true stim:log(day) = -0.2002 (100% of observed)
  N/group  | per-animal power | LME power
  ------------------------------------------
  3        |  0.12            |  0.26  <- observed
  5        |  0.27            |  0.39
  8        |  0.50            |  0.57
  12       |  0.78            |  0.82
  16       |  0.88            |  0.91
  24       |  0.97            |  0.97

 true stim:log(day) = -0.1001 (50% of observed)
  N/group  | per-animal power | LME power
  ------------------------------------------
  3        |  0.03            |  0.10  <- observed
  5        |  0.07            |  0.12
  8        |  0.17            |  0.22
  12       |  0.17            |  0.19
  16       |  0.33            |  0.39
  24       |  0.36            |  0.39

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