How to Detect Trading Leaks in Your Own Performance Data

A trading leak is a combination of conditions where you consistently lose money. They are invisible without data. This guide shows you how to find yours.

A trading leak is a specific combination of conditions — instrument, session, direction, emotional state — where you consistently lose money without realising it. It is called a leak because it drains your account slowly, hidden inside your overall P&L.

Leaks are nearly invisible without data. You might be profitable overall while one specific pattern is costing you 30% of what you could be making. You would never know without segmenting your results.

What a Trading Leak Looks Like

Here are real examples of trading leaks:

• A trader with 58% overall win rate, but 31% win rate on all trades taken after a losing trade (revenge trading pattern)
• A forex trader profitable on EUR/USD and GBP/USD, but losing consistently on GBP/JPY without realising it, because GBP/JPY losses are masked by EUR/USD wins
• A trader with positive expectancy in London session but negative expectancy in New York afternoon — trading the same setup in both sessions
• A trader whose win rate drops from 61% to 39% on days when their emotional state rating was 1 or 2 vs. 4 or 5

In each case, the trader had a real edge. They just were not applying it selectively enough.

The Three Types of Trading Leaks

1. Segment Leaks

A segment leak is when a specific category of trade (a session, an instrument, a direction) has negative expectancy. Expectancy is calculated as: (win rate × average winner) − (loss rate × average loser). Any segment with negative expectancy is losing you money on net.

To find segment leaks, you need at least 15–20 trades per segment for the data to be meaningful. With fewer trades, random variation can look like a pattern.

2. Behavioral Leaks

Behavioral leaks are patterns tied to your state or actions rather than market conditions. The most common ones:

• Revenge trading — taking trades immediately after a loss, often with larger size or lower quality setups
• Overtrading — days where you take significantly more trades than your average, usually on losing days
• Plan deviation — trades taken without meeting your entry criteria, which lose at higher rates
• Emotional correlation — win rate drops measurably when your emotional state is low

3. Combination Leaks

The most damaging leaks are combinations: a specific instrument in a specific session in a specific direction. For example, short GBP/JPY during the Asian session might be consistently losing while short EUR/USD during London open is consistently profitable. Aggregate data will not show this — only segmented data will.

How to Find Your Leaks: A Step-by-Step Process

1. Collect at least 30 trades in your journal with full context fields (session, instrument, direction, emotional state, followed plan)
2. Calculate expectancy for each session you trade. Compare them. Are any negative?
3. Calculate expectancy for each instrument you trade. Are any dragging your average down significantly?
4. Split your trades into "followed plan" vs. "deviated from plan." What is the expectancy difference?
5. Split by emotional state: trades taken at 1–2 vs. trades at 4–5. Is there a meaningful difference?
6. For any segment that shows negative expectancy, check if you have enough sample size (15+ trades). If yes, that segment is a confirmed leak.

What to Do Once You Find a Leak

The first instinct is to try to fix the leak — to figure out why GBP/JPY afternoon trades lose and work on them. That is usually the wrong move.

The right move is to stop trading the leak entirely and concentrate your capital on the segments with positive expectancy. You are not trying to become profitable at everything. You are trying to identify where your edge exists and trade only there.

A trader who goes from trading 6 sessions across 8 pairs to trading 2 sessions across 3 pairs — because the data showed those are their only profitable segments — is not being restrictive. They are being precise.

The Minimum Data Requirement

Leak detection requires data. With fewer than 30 trades, you will see patterns that are not really there — variance disguised as signal. With 50 trades, you can start drawing conclusions. With 100 trades, the patterns become reliable.

This is why logging every trade matters — not just the interesting ones. Every trade you skip logging is data you cannot use.