Why Backtest Results Mislead Most Traders
A strategy shows 500% returns in backtest. You go live. The first month is a loss. What happened?
Almost always: overfitting, curve fitting, or misreading the statistics. This guide explains what to actually look for.
The Metrics That Matter
Net Profit
The total return over the backtest period. This number alone means very little without context.
A 500% return over 10 years on a 1D chart is very different from a 500% return over 6 months on a 1M chart. The second is almost certainly overfitted.
Profit Factor
Net profit divided by net loss. A profit factor above 1.5 is good. Above 2.0 is excellent. Below 1.3 suggests the strategy barely covers losing trades.
Max Drawdown
The largest peak-to-trough decline during the backtest. This tells you the worst period you would have experienced.
A strategy with 300% returns and 50% max drawdown requires you to hold through a 50% account decline to achieve those returns. Most traders cannot do this emotionally.
Win Rate
Percentage of trades that close at a profit. A high win rate sounds good but means nothing without knowing the average win versus average loss.
A 40% win rate with 3:1 reward-to-risk is more profitable than an 80% win rate with 0.5:1 reward-to-risk.
Red Flags in Backtest Results
Too few trades: A strategy with 15 trades over 3 years is not statistically meaningful. You need at least 100 trades to draw conclusions.
Results only on one symbol: A strategy that only works on BTC/USDT is less robust than one that works across multiple pairs.
No slippage or fees modeled: Real trading has costs. Always include commission and slippage in your backtest settings.
Realistic Expectations from Backtesting
A well-designed strategy typically achieves 40-70% of its backtest returns in live trading due to slippage, fees, execution delays, and the natural degradation of any edge over time.
Use backtests to eliminate bad strategies, not to predict exact future returns.