Survivorship Bias: The Hidden Reason Your Backtest Looks Better Than It Should

This is survivorship bias, and it's one of the quietest ways a backtest lies to you without you ever noticing. Here's what it actually is, how it sneaks into your testing, and what it costs you when you trade the strategy live.

What Survivorship Bias Actually Means

Survivorship bias occurs when a historical dataset only includes assets that are still active or available at the end of the observation period, rather than all the assets that were active at each point along the timeline.

Think of it like this: if you want to know whether a physical training program works, and you only survey the people who made it to the Olympics, you'll conclude the program is foolproof. You never heard from the hundreds of athletes who got injured, burnt out, or dropped out along the way.

In trading, when you download historical data for "the S&P 500" or "the top 100 crypto tokens," most platforms give you historical prices for the assets that are in that group right now. Every company that collapsed, merged at a massive discount, or was removed for failing to meet listing requirements simply disappears from the record.

How This Inflates Your Backtest Numbers

Academic studies have consistently shown that survivorship bias in US equity data overstates historical returns by anywhere from 1% to 4% annually. In certain strategies — especially mean-reversion, value investing, and small-cap momentum — the distortion is even worse:

Where Survivorship Bias Hides Beyond Stock Indexes

Stock indexes aren't the only place survivorship bias poisons testing data. It's pervasive across almost every retail trading asset class: