What an honest verdict is: why a high Sharpe isn't enough

“It looks like it goes up” is not a verdict. A pretty curve can be pure luck if you don't measure uncertainty. That's why we don't stop at raw return or Sharpe.

We measure the net mean per rebalance and its p-value via a date-clustered bootstrap (to avoid fooling ourselves with correlated observations), the PSR (Sharpe adjusted for noise), and the percentage of positive years.

Only if it beats chance and cost robustly do we say PASS. The bar is high on purpose: we'd rather discard something good out of prudence than sell something false out of enthusiasm.