Your backtest results might be lying to you — here’s how to spot it and fix it before going live.

Introduction

Backtesting is critical.

But bad backtesting?

It’s worse than no testing at all — because it gives you false confidence.

Most traders don’t lose because they’re lazy. They lose because they over-trusted a strategy that looked great in hindsight — but was built on invisible flaws.

This post exposes the most dangerous backtest biases and shows you how to eliminate them.

The 3 Most Common Backtest Biases

1. Lookahead Bias

Using future information that wasn’t actually available at the time of trade.

Examples:

Entering based on the close of a candle — before it’s actually closed
Calculating moving average crossovers based on confirmed data
Using signals from a candle that hasn’t fully formed

Why it's dangerous: Your entries and exits appear “accurate,” but they’re unrealistically perfect — because you’re cheating time.

How to fix it:

Only act on closed candles (use bar replay or time-based logic)
Avoid functions that reference “future bars” in code
Simulate entries realistically (e.g. next bar open, bid/ask spreads)

2. Overfitting

Creating a strategy that performs well only on past data — but fails in real-time.

Symptoms:

Too many filters (volume spike + RSI + MA + pattern + moon phase)
Perfect equity curve in one market — but breaks in others
Strategy only works on one pair, one timeframe, one year

Why it's dangerous: You’re not discovering an edge — you’re memorizing noise.

How to fix it:

Test across multiple instruments & time periods
Keep your rules simple and robust
Apply cross-validation: test on unseen out-of-sample data
Try walk-forward testing: adapt model in segments, simulate realistic re-optimization

3. Survivorship Bias

Only testing systems or assets that still exist — ignoring those that failed or changed drastically.

Examples:

Backtesting SPY or BTC without accounting for delisted stocks, forks, splits
Ignoring slippage, spread, delistings, or black swan blowups

Why it’s dangerous: You’re assuming the conditions that created your edge will always exist.

How to fix it:

Use complete historical datasets, not just what exists now
Include “dead” assets in portfolio-level testing
Simulate volatility regimes, liquidity drops, and spreads increasing

Other Biases to Watch For

Bias Type	Description	Fix
Selection bias	Only testing your “favorite” trades	Include every trade in your data sample
Cherry picking	Manually excluding ugly outcomes	Log every result, good or bad
Optimism bias	Assuming you’ll always get filled at ideal prices	Simulate slippage and order book depth realistically
Anchoring bias	Refusing to retest or abandon old systems	Let data guide decisions, not nostalgia

Best Practices for Honest Backtesting

1. Use realistic assumptions

Include fees, spread, and slippage
Account for execution delay (e.g. not entering at candle close)
Simulate partial fills for large size

2. Separate in-sample and out-of-sample periods

Train your strategy on one period
Validate it on a completely different one → If performance holds across both: more robust

3. Keep your strategy as simple as possible

“A system is only as good as its worst assumption.”

Fewer moving parts = less overfitting risk The simpler it is, the easier it is to test, improve, and trust

Final Thought

Most failed traders didn’t skip testing. They trusted flawed testing.

Your system’s performance is only as reliable as the integrity of your backtest.

Don’t fall in love with the curve. Fall in love with clean process, honest data, and stress-tested logic.

Your backtest results might be lying to you — here’s how to spot it and fix it before going live.

Introduction

Backtesting is critical.

But bad backtesting?

It’s worse than no testing at all — because it gives you false confidence.

Most traders don’t lose because they’re lazy. They lose because they over-trusted a strategy that looked great in hindsight — but was built on invisible flaws.

This post exposes the most dangerous backtest biases and shows you how to eliminate them.

The 3 Most Common Backtest Biases

1. Lookahead Bias

Using future information that wasn’t actually available at the time of trade.

Examples:

Entering based on the close of a candle — before it’s actually closed
Calculating moving average crossovers based on confirmed data
Using signals from a candle that hasn’t fully formed

Why it's dangerous: Your entries and exits appear “accurate,” but they’re unrealistically perfect — because you’re cheating time.

How to fix it:

Only act on closed candles (use bar replay or time-based logic)
Avoid functions that reference “future bars” in code
Simulate entries realistically (e.g. next bar open, bid/ask spreads)

2. Overfitting

Creating a strategy that performs well only on past data — but fails in real-time.

Symptoms:

Too many filters (volume spike + RSI + MA + pattern + moon phase)
Perfect equity curve in one market — but breaks in others
Strategy only works on one pair, one timeframe, one year

Why it's dangerous: You’re not discovering an edge — you’re memorizing noise.

How to fix it:

Test across multiple instruments & time periods
Keep your rules simple and robust
Apply cross-validation: test on unseen out-of-sample data
Try walk-forward testing: adapt model in segments, simulate realistic re-optimization

3. Survivorship Bias

Only testing systems or assets that still exist — ignoring those that failed or changed drastically.

Examples:

Backtesting SPY or BTC without accounting for delisted stocks, forks, splits
Ignoring slippage, spread, delistings, or black swan blowups

Why it’s dangerous: You’re assuming the conditions that created your edge will always exist.

How to fix it:

Use complete historical datasets, not just what exists now
Include “dead” assets in portfolio-level testing
Simulate volatility regimes, liquidity drops, and spreads increasing

Other Biases to Watch For

Bias Type	Description	Fix
Selection bias	Only testing your “favorite” trades	Include every trade in your data sample
Cherry picking	Manually excluding ugly outcomes	Log every result, good or bad
Optimism bias	Assuming you’ll always get filled at ideal prices	Simulate slippage and order book depth realistically
Anchoring bias	Refusing to retest or abandon old systems	Let data guide decisions, not nostalgia

Best Practices for Honest Backtesting

1. Use realistic assumptions

Include fees, spread, and slippage
Account for execution delay (e.g. not entering at candle close)
Simulate partial fills for large size

2. Separate in-sample and out-of-sample periods

Train your strategy on one period
Validate it on a completely different one → If performance holds across both: more robust

3. Keep your strategy as simple as possible

“A system is only as good as its worst assumption.”

Fewer moving parts = less overfitting risk The simpler it is, the easier it is to test, improve, and trust

Final Thought

Most failed traders didn’t skip testing. They trusted flawed testing.

Your system’s performance is only as reliable as the integrity of your backtest.

Don’t fall in love with the curve. Fall in love with clean process, honest data, and stress-tested logic.

Biases in Backtesting

Introduction

The 3 Most Common Backtest Biases

1. Lookahead Bias

2. Overfitting

3. Survivorship Bias

Other Biases to Watch For

Best Practices for Honest Backtesting

1. Use realistic assumptions

2. Separate in-sample and out-of-sample periods

3. Keep your strategy as simple as possible

Final Thought

Introduction

The 3 Most Common Backtest Biases

1. Lookahead Bias

2. Overfitting

3. Survivorship Bias

Other Biases to Watch For

Best Practices for Honest Backtesting

1. Use realistic assumptions

2. Separate in-sample and out-of-sample periods

3. Keep your strategy as simple as possible

Final Thought