skepsis

Is your backtest result real, or overfit?

skepsis takes the returns of a backtested strategy — and, ideally, the returns of every variant you tried along the way — and produces statistical evidence for or against the result being luck, plus a self-contained HTML report you can hand to a PM.

pip install skepsis

import skepsis

result = skepsis.evaluate(returns, trials=trials_df, params=params_df, freq="daily")
print(result.summary())
result.save_html("skepsis_report.html")

Start with the demo: Deflating the Golden Cross — a Sharpe-1.46, 651× backtest on 11 years of BTC, and what's left of it once you account for the 34 variants tried (spoiler: a DSR of 0.35).

skepsis is not a backtester. It sits downstream of whatever you use to generate returns, and it never silently repairs bad input: NaNs are rejected, strained assumptions are warned about, skipped diagnostics say why. The implementations are verified against reference values in the golden test suite — including reproducing the DSR paper's own numerical example to all four printed decimals.