Backtest overfit check

A great backtest screenshot proves very little. If you tried 26 versions and kept the best one, some of that "edge" is just luck. This check measures how much.

In the first video I tested the 200-day moving average on SPY across 26 lengths. The winner was probably not a fluke, but whether it beats simply buying and holding is inside the margin of error.

How to submit

Email sampledexe.check@gmail.com with the subject "Backtest check" and include:

1. A CSV with two columns, date, daily_return (or date, equity), daily, at least 2 years.
2. Roughly how many parameter sets or versions you tried ("don't know" is fine).
3. The backtest period, and whether trading costs are included.

Optional: one return column per version you tried, so I can also compute the probability of overfitting (PBO).

Your data and report stay private unless you say otherwise. Only the CSV is opened. Not investment advice.