Common Backtesting Pitfalls

​

Why pitfalls happen

Backtests feel objective because they produce numbers and charts. But most backtest mistakes come from:

• wrong assumptions (timeframe, execution, intrabar behavior)
• overfitting (optimizing to the past)
• ignoring risk (drawdown, tail losses)
• ignoring data limitations (coverage, gaps, warmup)

This page helps you avoid the traps that create false confidence.

---

​

Pitfall 1: Starting with total return

If you start with the biggest number (PnL / CAGR), you will:

• overfit quickly
• ignore risk
• misunderstand behavior

Fix:

• read results in the correct order

How to read results

Use a stable workflow for every backtest.

---

​

Pitfall 2: Ignoring trade count

A strategy with 7 trades can look incredible by chance. Fix:

• require a meaningful trade sample
• validate behavior across multiple ranges

---

​

Pitfall 3: Expecting intrabar triggers

Many traders expect:

• entries on candle highs/lows
• instant crossovers mid-candle

Trinigence evaluates logic on candle close by default. Fix:

• design strategies for candle-close evaluation
• validate triggers using trade history

---

​

Pitfall 4: Confusing events and states

Common mistake:

• using a crossover (event) as if it were a persistent condition (state)

Example:

• “EMA crosses above EMA” triggers once
• “EMA is above EMA” stays true

Fix:

• use events for triggers
• use states for filters

Crossovers & trend changes

Events vs states explained clearly.

---

​

Pitfall 5: Over-filtering the strategy

Too many filters often lead to:

• almost no trades
• misleading metrics
• “perfect” equity curves that are just inactivity

Fix:

• start with a baseline strategy
• add one filter at a time

---

​

Pitfall 6: Exit logic dominates (but you don’t notice)

A strategy can appear profitable because:

• TP/SL choices dominate outcomes
• exits are too wide/tight
• one extreme trade drives results

Fix:

• simplify exits
• inspect the biggest wins/losses
• validate distribution, not just averages

---

​

Pitfall 7: Over-optimizing parameters

Tuning parameters to maximize past performance often creates:

• fragile strategies
• collapse out of sample
• regime dependence

Fix:

• small parameter changes should not destroy performance
• test across multiple time ranges
• prefer robust parameter zones

Indicator parameters

Understand how parameters change behavior.

---

​

Pitfall 8: Ignoring data coverage and warmup

Common symptoms:

• “strategy doesn’t trade early in the backtest”
• “results look broken in old periods”
• “backtest range got adjusted”

Fix:

• confirm data coverage
• understand indicator warmup behavior

Data coverage

Learn how coverage affects results.

---

​

Pitfall 9: Assuming one market equals all markets

A strategy that works on BTC may fail on:

• ETH
• altcoins
• other volatility regimes

Fix:

• test on multiple markets
• look for behavior consistency, not peak returns

---

​

Pitfall 10: Ignoring drawdown duration

Many traders only look at max drawdown, not how long it lasts. A strategy can have:

• acceptable max drawdown
• but extremely long recovery periods

Fix:

• evaluate drawdown duration alongside max drawdown

---

​

Quick self-check

Before trusting a backtest, ask:

• Do I understand the first 10 trades?
• Do I have enough trades for confidence?
• Is risk acceptable (DD + duration + tail trades)?
• Does it work across multiple ranges?
• Does it survive small parameter/timeframe changes?

---

​

What to read next