Skip to content
Lexicon · Testing

What is overfitting?

Overfitting is when a trading strategy is tuned so closely to past data that it captures noise instead of a real pattern, and then fails on new data.

Updated

(01) Explained

Overfitting, in plain terms.

Every price history contains random noise. Add enough conditions and fine-tune enough settings, and a strategy can be made to fit that noise almost perfectly.

The fit looks impressive in a backtest, but the noise doesn't repeat, so the strategy breaks down on data it hasn't seen.

(02) Example

A rule that fits the past too well

  1. "Buy when RSI is below 27.5, only on Tuesdays, only when volume is between 1.3x and 1.45x its average."
  2. Tuned on one period, a rule like this can show excellent results.
  3. There's no reason those exact values should matter, so the edge usually disappears in the next period.
(03) Why it matters

What to watch for.

  • Warning signs. Many conditions, very precise settings, few trades, and results that collapse when a setting changes slightly.
  • Keep rules simple. Fewer, broader conditions are more likely to capture a real pattern.
  • Test on unseen data. Design rules on one period, then check them on a period you didn't use to build them.
(04) In The Market Lexicon

Where it shows up in the product.

Backtests in The Market Lexicon show the trade count next to return and drawdown, so you can see when a result rests on too few trades.

Backtesting in The Market Lexicon →

Test ideas like this on real market history.

Start free trialPro $50 · Elite $100 per month