Chapters in Technical Analysis & Trading Tools
Chapter 7 of 10 · 6 min read
Backtesting: testing ideas on history
How to evaluate a rule before risking money, and the traps that make results look better than they are.
A backtest applies a rule to past data and records the outcomes: win rate, average gain and loss, drawdown. It answers whether the idea had an edge, not whether it will in future.
Traps
- Overfitting: tuning until the past looks perfect.
- Look-ahead bias: using information not available at the time.
- Survivorship bias: testing only stocks that still exist.
- Ignoring costs: brokerage, taxes and slippage reduce real returns.
- Too few trades: small samples are noise.
What to look at
Look beyond total return: maximum drawdown, win rate versus payoff ratio, and consistency across years. Past results do not predict future returns.
Key takeaways
- A backtest supports or rejects an idea; it never proves it.
- Watch for overfitting, bias and costs.
- Prefer simple rules that hold across regimes.
Put it into practice
Open the live page on Market Intelligence and see this concept on real data.
Keep learning
- AI Signals: a next-day model and its track record · Sentiment, News & AI Research
Education only, not investment advice. Updated 2026-10-09.