Most retail traders skip backtesting. They see a strategy that looks good, fund an account, and start trading live. This is almost always a mistake.

Backtesting doesn't guarantee future results — but it's the difference between making an informed bet and a blind one. This guide covers how to do it properly: what data to use, which metrics to focus on, and the pitfalls that make backtests misleading.

⚠️ Important: Even a perfect backtest cannot guarantee future profitability. Markets change. Use backtesting to filter out clearly bad strategies — not to find guaranteed winners.

What Is Backtesting?

Backtesting runs a trading strategy against historical price data to see how it would have performed. It answers: "If I had been running this exact logic for the past 6 months, what would my account look like?"

A well-run backtest tells you:

The Metrics That Actually Matter

Most people look at total profit first. That's the wrong starting point. Focus on these metrics in this order:

Max Drawdown
Largest peak-to-trough loss
✓ Good: under 20%
Profit Factor
Gross profit / Gross loss
✓ Good: 1.5 or higher
Win Rate
% of winning trades
✓ Context-dependent
Avg Win / Avg Loss
R:R per trade
✓ Good: 2:1 or higher
Sharpe Ratio
Return per unit of risk
✓ Good: above 1.0
Consecutive Losses
Worst losing streak
✓ Can you survive it?
💡 Profit factor explained: If your strategy made $10,000 in gross profits and lost $6,000 in gross losses, profit factor = 10,000 / 6,000 = 1.67. Anything below 1.0 means the strategy is net losing.

Step-by-Step: How to Backtest Properly

1. Choose Your Timeframe and Asset

Test on the actual asset and timeframe you intend to trade live. A BTC strategy tested on ETH data is not a validated BTC strategy. A 4-hour strategy tested on 1-minute bars tells you nothing about 4-hour performance.

2. Use Sufficient Historical Data

Use at least 12 months of data, ideally 2–3 years. Crypto markets cycle through bull runs, bear markets, and ranging consolidations. A strategy that only worked in a 2021 bull run is not a general strategy — it's a bull market strategy.

Pay attention to whether your backtest period included:

3. Include Realistic Fees and Slippage

This is where most backtests become dangerously optimistic. A strategy that returns 40% before fees might return 15% after. Use realistic assumptions:

4. Run the Backtest and Record Results

Document the full results — not just the profit, but every metric listed in the section above. If you only have total profit, you don't have enough information to evaluate the strategy.

5. Stress Test the Parameters

Change your strategy parameters slightly (±10–20%) and re-run the backtest. A robust strategy should still be profitable with small parameter variations. If changing the RSI period from 14 to 12 collapses the results, the strategy is overfitted — it found patterns that worked in the past but won't generalize.

6. Paper Trade on Live Prices

After a positive backtest, run the strategy in demo/testnet mode against live prices for at least 2 weeks. This reveals issues that backtesting misses: execution latency, API delays, and real-time data differences. Only after this step should you consider live trading.

5 Common Backtesting Mistakes

Overfitting (Curve Fitting)
Tweaking parameters until the backtest looks perfect. Every dataset has noise — optimize too aggressively and you fit the noise, not the signal. The strategy will fail on new data. Solution: test the parameters you chose before looking at results, and validate on out-of-sample data.
Ignoring Fees and Slippage
Many backtests assume zero-cost execution. In reality, fees + slippage on a high-frequency strategy can consume 50–100% of gross profits. Always include realistic costs.
Look-Ahead Bias
Using data that wouldn't have been available at the time of the trade. Example: using the candle's closing price to trigger an entry that happens "at the open." This makes backtests appear much better than they actually are.
Survivorship Bias
Testing only on assets that still exist and performed well. If you build a basket strategy on "top 20 crypto" using today's list, you're including coins that survived — not the average coin from 3 years ago. Test on assets that existed at your historical start date.
Not Testing the Drawdown Periods
A backtest showing 50% annual return with a 45% max drawdown is not a good strategy for most people — the psychological and financial reality of a 45% drawdown is brutal. Always simulate what it would feel like to live through the worst periods, not just the average.

Backtesting vs Paper Trading: When to Use Each

Use backtesting when:

Use paper trading (demo mode) when:

Both are essential. Neither replaces the other. The path to live trading should always be: backtest → paper trade → small live position → scale up.

Backtest Any Strategy Free for 14 Days

Enliko includes a full backtesting module with real historical data from Bybit, HyperLiquid, and Capital.com. Run, compare, and validate strategies before risking any real funds.

Start Free Trial
No credit card · Full backtesting included · 14 days demo

Frequently Asked Questions

How much historical data do I need to backtest?

At minimum 12 months, ideally 2–3 years. You need enough data to include different market conditions: bull, bear, and ranging. Using only 3 months of data from a bull market will make almost any long-biased strategy look great.

What profit factor should I target before going live?

A profit factor above 1.5 is generally considered acceptable. Above 2.0 is good. Above 3.0 should make you suspicious — it may indicate curve fitting. Prioritize consistency over raw numbers.

Should I use 1-minute or daily candles for backtesting?

Use the same timeframe you plan to trade. If your strategy checks for signals every hour, test on 1-hour candles. Testing on higher timeframes than your strategy runs on will miss intracandle volatility and give overly optimistic results.