Free data hides the stocks that died. Counting only survivors makes every strategy look better than it was.
Survivorship bias is the error of studying only the things that lasted. In markets, most penny and small-cap names eventually delist — bankruptcy, buyout, failure to meet listing rules. Free data feeds usually drop those, keeping only the survivors.
If your history contains only the names that made it, every bounce looks like it recovered and every breakout looks like it ran — because the ones that went to zero aren’t in the sample. Win rates and average returns come out optimistically wrong, sometimes wildly so.
Measure on a universe that includes delisted names, clip each name’s data to the window it actually traded, and state the bias you can’t remove. That is the difference between a base rate you can trust and a sales pitch. It is also why every stat here is survivorship-free.