What Is Backtesting?
Every trader eventually asks the same question: "Would this strategy have worked in the past?" That instinct, to validate an idea before risking real money, is the foundation of backtesting.
Backtesting is the process of applying a trading strategy to historical market data to see how it would have performed. You take your rules, entry conditions, exit conditions, position sizing, stop losses, and run them against months or years of real price data to simulate the outcomes.
It is one of the most important practices in systematic trading, and also one of the most misunderstood.
Why Backtesting Matters
The alternative to backtesting is trading live with real money and hoping your strategy works. This is how most retail traders operate, and it is also why most retail traders lose money.
Backtesting provides several critical insights:
Performance expectations: How much could you expect to make (or lose) over a given period? What is the typical winning streak? The typical losing streak? Without backtesting, you are flying blind.
Risk assessment: What is the maximum drawdown? How long does it take to recover from the worst losing period? These numbers determine whether you can psychologically and financially survive the strategy's bad patches.
Strategy validation: Does the strategy actually have an edge, or does it just look good on a handful of cherry-picked examples? Backtesting across different market conditions reveals whether the edge is real and robust.
Parameter sensitivity: How much do results change if you tweak the parameters slightly? A strategy that works beautifully with a 14-period RSI but fails with a 12 or 16-period RSI is likely overfitted and will disappoint in live trading.
Confidence building: Knowing that a strategy has performed well across years of data, including bear markets, flash crashes, and sideways periods, gives you the confidence to stick with it during inevitable rough patches.
Key Metrics to Evaluate
When reviewing backtest results, focus on these metrics:
Win Rate: The percentage of trades that are profitable. A win rate of 55-65% is typical for trend-following strategies. Note that win rate alone tells you very little, a 40% win rate with large winners and small losers can be highly profitable.
Profit Factor: Total gross profit divided by total gross loss. A profit factor above 1.5 is generally considered good. Above 2.0 is excellent. Below 1.0 means the strategy lost money.
Sharpe Ratio: Measures risk-adjusted return. It tells you how much return you earn per unit of risk. A Sharpe ratio above 1.0 is acceptable; above 2.0 is strong. This is one of the most widely used metrics in professional finance.
Maximum Drawdown: The largest peak-to-trough decline in portfolio value. This is arguably the most important metric because it tells you the worst-case scenario you need to survive. A strategy with great returns but 60% drawdowns will test most traders beyond their breaking point.
Calmar Ratio: Annual return divided by maximum drawdown. This directly quantifies return per unit of worst-case risk. A Calmar above 1.0 means you are earning more annually than your worst drawdown.
Total Number of Trades: A backtest with 10 trades is statistically meaningless. You need hundreds of trades minimum for results to be reliable. More trades = more statistical confidence.
Average Trade Duration: Understanding how long positions are typically held helps you know what to expect and whether the style matches your preferences.
Common Backtesting Pitfalls
Backtesting is powerful, but it is also dangerously easy to get wrong. Here are the most common traps:
Overfitting (Curve Fitting)
This is the cardinal sin of backtesting. Overfitting happens when you optimise your strategy parameters so precisely to historical data that they capture noise rather than genuine market patterns.
An overfitted strategy might show spectacular historical results but fail immediately in live trading because the patterns it "learned" were random coincidences that will not repeat.
Signs of overfitting include:
- Extremely specific parameter values (e.g., a 17.3-period RSI with a 2.847 standard deviation threshold)
- Results that are dramatically better than simpler versions of the strategy
- Performance that deteriorates sharply with even small parameter changes
- An excessive number of rules and conditions
Survivorship Bias
If your historical data only includes assets that still exist today, you are ignoring all the tokens that crashed to zero and were delisted. This creates an artificially rosy picture of strategy performance because the worst outcomes are excluded from the data.
Look-Ahead Bias
This occurs when your backtest accidentally uses information that would not have been available at the time of the trade. For example, using the daily closing price to make a decision that supposedly happened during the trading day.
Ignoring Transaction Costs
A strategy that trades frequently might look profitable before fees but lose money after accounting for trading commissions, slippage, and funding rates. Always include realistic transaction costs in your backtests.
Unrealistic Execution Assumptions
In a backtest, orders fill instantly at exactly the price you specify. In live markets, there is slippage, the difference between the expected price and the actual fill price. For large positions or in low-liquidity markets, slippage can significantly impact results.
Walk-Forward Validation
Walk-forward validation is the gold standard for honest strategy testing. It works like this:
- Divide your historical data into multiple periods (e.g., 12 monthly segments)
- For each period, train (optimise) the strategy on the preceding data
- Test the optimised strategy on the next unseen period
- Repeat this process across all periods, each time training on past data and testing on future data
The key principle: the strategy is always tested on data it has never seen before. This mimics what happens in live trading, you develop a strategy based on what you know, then face an uncertain future.
Walk-forward results are almost always worse than standard backtest results. If they are dramatically worse, the strategy is likely overfitted. If they hold up reasonably well, you have more confidence in the strategy's robustness.
In-Sample vs Out-of-Sample Testing
A simpler version of walk-forward validation involves splitting your data into two parts:
- In-sample data (typically 60-70%): Used to develop and optimise the strategy
- Out-of-sample data (remaining 30-40%): Reserved for testing and never touched during development
The out-of-sample results are far more meaningful than in-sample results because the strategy had no opportunity to adapt to that data. If results are strong in both segments, the strategy likely has a genuine edge.
From Backtest to Paper Trading
Backtesting tells you how a strategy would have performed in the past. Paper trading tells you how it performs in real-time market conditions without risking real money.
Paper trading bridges the gap between historical simulation and live execution. It introduces elements that backtesting cannot capture:
- Real-time data latency
- Actual order execution timing
- Market conditions that did not exist in the historical data
- Your own psychological response to watching trades unfold
A recommended workflow is: Backtest → Optimise → Walk-forward validate → Paper trade for 2-4 weeks → Begin live trading with small capital → Scale gradually.
How TradingGenie Handles Backtesting
TradingGenie provides built-in backtesting with several features designed to produce honest results:
- Walk-forward validation: Strategies are tested using the walk-forward method, not simple historical optimisation
- Realistic transaction costs: Backtests include actual Hyperliquid trading fees and estimated slippage
- Multiple market conditions: Results are broken down by market regime (trending, ranging, volatile) so you can see where the strategy excels and where it struggles
- Comprehensive metrics: All key metrics (Sharpe, Calmar, drawdown, profit factor, win rate) are displayed alongside a complete trade log
- Free paper trading: After backtesting, you can transition to paper trading on the Starter plan at no cost, testing in live market conditions before committing capital
Interpreting Results Honestly
When reviewing any backtest results, ours or anyone else's, keep these principles in mind:
- Past performance does not guarantee future results: This is not just a legal disclaimer. Markets evolve, and what worked yesterday may not work tomorrow.
- Focus on risk metrics as much as return metrics: A strategy that made 100% with a 60% drawdown is far less attractive than one that made 50% with a 15% drawdown.
- Demand statistical significance: Results from fewer than 100 trades should be viewed with extreme scepticism.
- Compare to benchmarks: Is the strategy actually better than simply holding Bitcoin? If not, the complexity may not be justified.
- Expect live results to be worse: Slippage, latency, and changing market conditions almost always mean live results underperform backtests. Plan for this.
Backtesting results are based on historical data and do not guarantee future performance. Trading cryptocurrency involves substantial risk of loss. Only trade with capital you can afford to lose.