The Psychology of Looking at Backtest Results

The backtest is complete. Numbers fill the screen: win rate, profit factor, max drawdown, Sharpe ratio, total return. The equity curve paints a picture. You stare at this data and form a conclusion about your strategy. But here's what most traders don't realize: the conclusion you form is influenced as much by your psychology as by the data itself. The same backtest results lead different traders to opposite conclusions — because human brains are interpretation machines, not data-processing machines, and the interpretations are riddled with predictable cognitive biases.

Confirmation Bias: Seeing What You Want

You built this strategy. You spent hours designing it, choosing indicators, setting parameters. You have emotional investment in it working. When the backtest results appear, your brain is primed to find evidence of success and dismiss evidence of failure.

The equity curve has a rough patch in 2022? "That was the crypto winter — everyone lost money then." The win rate is 48%? "But the profit factor is 1.6, so wins are bigger than losses." The max drawdown is 35%? "That only happened once and it recovered."

Each rationalization might be valid individually. But notice the pattern: every negative metric gets explained away while positive metrics are accepted at face value. This is confirmation bias in action. If you had built a different strategy and gotten the same results, you'd evaluate them more objectively — because you'd have no emotional stake in the outcome.

Anchoring: The First Number Wins

The first metric you see disproportionately influences your overall assessment. If the backtest dashboard shows "Total Return: +187%" at the top, you're anchored. Every subsequent metric is evaluated through the lens of "+187% return." A 30% max drawdown? "Worth it for 187% return." A 0.8 Sharpe? "Below average, but 187% return makes up for it."

Reverse the order. If you saw "Max Drawdown: 35%" first, your assessment would be: "Painful. Is the return worth this risk?" The same data leads to different conclusions based on viewing order. Professional analysts read risk metrics first specifically to prevent return anchoring.

Survivorship Bias in Self-Evaluation

You've probably tested 10-20 strategies before finding one that "works." You discard the failures and focus on the winner. But statistically, if you test 20 strategies, some will look profitable by random chance alone. The one you select as "your strategy" might simply be the random winner of the bunch — and you've unconsciously data-mined by selecting the best-performing variation from a large set of trials.

The Visual Deception of Equity Curves

Equity curves are powerful visual tools — and powerful deception tools. A curve that trends upward from left to right looks "good" regardless of the actual metrics. Your visual system processes the shape before your analytical brain examines the numbers. A 45-degree uptrend on the equity curve creates a positive impression that's hard to override even when the numbers tell a different story.

Worse, the Y-axis scale dramatically changes perception. The same equity curve looks like a rocket ship on a compressed Y-axis and flat on an expanded one. A 50% return with 30% drawdown can look spectacular or mediocre depending entirely on how the chart is scaled.

Loss Aversion in Strategy Evaluation

Humans feel losses approximately 2× as intensely as equivalent gains. This distorts how you evaluate backtest periods. You might stare at the 2022 drawdown for 10 minutes and glance at the 2023 recovery for 30 seconds — even though both are equally important for evaluating the strategy. The result: you either reject a good strategy because the drawdown "feels too scary" or you add unnecessary filters to avoid historical drawdowns, inadvertently overfitting the strategy.

The Narrative Fallacy

After seeing backtest results, your brain constructs a story explaining WHY the strategy works. "It captures the momentum shift after consolidation periods." This narrative feels explanatory but is often post-hoc rationalization. You're fitting a story to results rather than predicting results from a theory. The danger: if the narrative is wrong, you won't know when to stop trading the strategy because your mental model doesn't match reality.

How to Analyze Results Objectively

1. Read Risk First

Before looking at returns, examine: maximum drawdown, worst month, longest losing streak, and worst drawdown duration. Form your initial impression from risk, not reward. Ask: "Would I hold through this drawdown with real money?" If no, the strategy's returns are irrelevant.

2. Use Benchmarks

Compare your strategy against buy-and-hold for the same instrument and period. A strategy that returns 100% while BTC returned 150% just underperformed passive holding. Compare against a simple benchmark strategy (200 SMA filter). If your complex strategy barely beats a simple one, the added complexity may not be worth the overfitting risk.

3. Focus on Profit Factor and Expectancy

These metrics are harder to misinterpret than win rate or total return. Profit factor below 1.3 is marginal. Expectancy tells you the dollar value of each trade. These numbers resist the psychological biases that distort win rate and return figures.

4. Pre-Register Your Evaluation Criteria

Before running the backtest, write down what results you'd need to see to accept or reject the strategy. "I'll trade this if profit factor > 1.5, max drawdown < 25%, and it works on at least 3 instruments." This prevents post-hoc rationalization — you can't move the goalposts after seeing results if the goalposts were written down first.

Let the Data Decide

StratBase.ai presents comprehensive metrics for every backtest — risk metrics alongside returns, with clear benchmarks. But the final interpreter is always you. Understanding the biases in your own interpretation is as important as understanding the strategy's mechanics. The best backtest in the world is useless if you can't read the results objectively.