Where Did Your Profit Go? How UltraTrader AI Ranks Your Trading Leaks
A trading leak is a pattern that repeats and takes money out of your account one trade at a time. A winner drifts back below your entry. A stop gets moved once, then again. Each one looks small on the day it happens. Over 100 trades they add up, and the order you fix them in decides how much of your review time pays off.
This post prices them. You’ll put each exit leak in dollars against one benchmark, rank the leaks by cost and count overlapping trades once. The worked example uses a hypothetical journal of 120 trades, and the last section shows where the AI review in UltraTrader AI fits.
The short version
- Price every leak against the same benchmark. Otherwise the ranking compares different things.
- This post prices three exit leaks: winners that turned into losses, winners cut short and losses bigger than the plan.
- A trade can sit in two leaks at once. Count it once when you add up the total.
- The benchmark you choose can change the order. Rerun the ranking at a lower and a higher benchmark before you pick a fix.
- A leak with a small trade count gives you a direction to check. The interval around its rate shows how far it can still move.
What you need from your trade history
For each closed trade you need the entry price, exit price, direction, size, planned stop and the best open profit the trade reached before it closed. The planned stop gives you 1R, the amount you planned to lose. The best open profit is the maximum favorable excursion (MFE), and the MAE and MFE guide explains how to calculate it.
A standard trade history shows the result but not the path the price took. To get MFE you need the highest and lowest prices while each trade was open. Mark trades where that data is missing and leave them out of the leak counts. Don’t estimate it.
Put everything in R. A trade that made $150 on $100 of planned risk is +1.5R. R lets you compare trades of different sizes, and you can convert back to dollars at the end.
Three exit leaks and how to price them
| Leak | Rule used in this post | Cost of each trade |
|---|---|---|
| Winners that turned into losses | Open profit reached +1R, trade closed below zero | Benchmark profit minus the result |
| Winners cut short | Open profit reached +1R, trade closed in profit but kept less than the benchmark | Benchmark profit minus the result |
| Oversized losses | Trade closed worse than −1.2R | The part of the loss beyond −1R |
The benchmark profit in this example is half of the best open profit. Nobody exits at the exact top, so pricing a leak against the full MFE makes it look bigger than any exit rule could recover. Half is a starting point chosen for the example, and the sensitivity check below shows what happens when you move it.
The 1.2R tolerance on losses leaves room for slippage and spread. A −1.1R loss on a −1R plan is more likely execution cost. A −2.5R loss points to a stop that moved or was never placed.
The first two leaks are both about open profit, from opposite ends. One held a winning trade with no protection until it went red. The other closed it before the move was done. Oversized losses are about the stop, and the guide to holding losers longer than winners covers the time side of that pattern.
Worked example: 120 hypothetical trades
The journal in this example is made up. A seeded script generated 120 closed trades with $100 of planned risk on each. The journal won 57 trades (47.5%) and finished at −$2,998 (−30.0R). 52 of the trades reached +1R of open profit at some point.
Ranked by cost, with the benchmark at half of the best open profit:
| Rank | Leak | Trades | Cost (R) | Cost ($) | Share of listed cost |
|---|---|---|---|---|---|
| 1 | Winners that turned into losses | 13 | 23.2R | $2,322 | 53.7% |
| 2 | Oversized losses | 12 | 11.8R | $1,178 | 27.2% |
| 3 | Winners cut short | 23 | 8.3R | $826 | 19.1% |
The listed costs add up to $4,326. Four trades appear in two rows: they reached +1R, reversed and then closed beyond −1.2R. Counted once, at the larger of their two costs, the total is $4,042.
Cut-short winners have the most trades, 23, and the smallest cost. Ranked by count, they’d have come first.
The total isn’t money this trader would have kept. A rule that protects reversing winners also closes some trades that would have gone on to run further. Use the cost to size the problem, then test a rule change on the same trades before you trade it.
Check the benchmark before you trust the order
| Benchmark (share of best open profit) | Winners into losses | Oversized losses | Winners cut short |
|---|---|---|---|
| 30% | $1,887 (13 trades) | $1,178 (12 trades) | $139 (12 trades) |
| 50% | $2,322 (13 trades) | $1,178 (12 trades) | $826 (23 trades) |
| 70% | $2,757 (13 trades) | $1,178 (12 trades) | $1,818 (32 trades) |
At 70%, cut-short winners move from third place to second. Oversized losses stay at $1,178 at every level, because their cost is measured against the stop and has nothing to do with open profit. The top leak stays first at all three levels, so in this journal it’s a stable pick. If your top two leaks swap places when you move the benchmark, the ranking can’t tell you which one to work on first.
Pick the benchmark from your strategy. A trend-following method that waits for rare long moves gives back more open profit by design than one that takes quick targets. The MAE and MFE guide covers this under capture rate.
How to run it on your own journal
- Export the closed trades for one strategy over one period. Mixing strategies blends their exit rules.
- Add 1R, the result in R and the MFE in R for each trade. Flag trades with missing price data.
- Apply the three rules from the table and price each trade they catch.
- Rank the leaks by total cost. Add up the total again with each trade counted once.
- Rerun the ranking at benchmarks of 30% and 70% and note whether the top leak holds its place.
- Write one rule change aimed at the top leak. Test it on the same closed trades, then check it on the next block of trades. The guide to exit rules shows how to compare one rule with the original record.
Write a pass rule down before you look at the results, for example: “I work on the top leak only if it stays first at all three benchmarks and covers at least 10 trades.” Set the numbers to fit your own sample.
Limits of this method
Small samples swing. In the example, 13 of the 52 trades that reached +1R turned into losses, a rate of 25.0%. The 95% Wilson interval for that rate runs from 15.2% to 38.2%, using the method in the NIST Engineering Statistics Handbook. With a range that wide, keep collecting trades and rerun the ranking as they come in.
There’s no statistical threshold behind round checkpoints such as 30, 100 or 200 trades. They’re useful moments to review. The interval tells you how much a number can still move.
Trade records don’t show intent. A winner that turned into a loss may have followed the plan exactly, for example in a strategy that has no breakeven rule on purpose. Check the planned exit before you call a trade a leak.
Fees, swaps and spread sit inside each result. If they’re large in your account, price them as a separate leak so they don’t hide inside the others. And a ranking only covers the patterns you defined. A strategy that loses money on entries can show small exit leaks and still lose, so read the leak table next to your expectancy.
Where UltraTrader AI fits
The AI review in UltraTrader AI ranks exit leaks by cost under “Where the money went”. It reviews your completed trades on an MT5 auto-sync portfolio and lists leaks such as winners slipping into losses, closing winners too early and oversized losses. The review also notes that a trade can appear in more than one pattern.
The same review grades six areas: Risk/Reward, Entry Timing, Profitability, Consistency, Drawdown and Exit Strategy. It compares three exit rules on your own trades (Optimized Stoploss, Optimized Exit and Multi-Level Exit) and marks the best fit for you. If a rule did no better than your own exits, it tells you. These comparisons are replays of past trades.
The review covers your last 6 months of trading or your last 500 trades, whichever comes first, and it needs 30 trades before the full analysis runs. The thresholds in this post belong to the manual method above. Check the source trades behind any leak the review shows, the same way you’d check your own spreadsheet. The AI review is part of Premium, for MT5 portfolios.
FAQ
How do I find the best open profit of a trade?
Take the highest price for a long trade, or the lowest price for a short trade, between entry and exit from chart data. The distance from the entry price to that point, divided by your planned risk, is the trade’s MFE in R. If you can’t get the price path for a trade, leave it out of the leak counts.
How many trades do I need before a leak ranking means anything?
There’s no fixed number. Work out a confidence interval for each leak’s rate, such as the Wilson interval, and look at how wide it is. In the example, 52 trades gave a range of 15.2% to 38.2% for one leak. Rerun the ranking as new trades come in and check whether the order holds.
Should I fix the most expensive leak first?
Start with it if it stays first when you move the benchmark and has enough trades behind it. If the top two leaks swap places, test a rule for each on the same closed trades before you choose.
Is the cost of a leak money I would have made?
No. The cost is measured against a benchmark you picked. A rule that stops one leak also changes other trades, so the effect only shows when you test the rule on past trades and then on new ones.
Run it on your own trades
On Premium, connect an MT5 account to UltraTrader AI and the AI review shows where your profit slipped away, ranked by cost.
For MT5 auto-sync portfolios. Historical analysis, no guarantees.
The trades in this post are hypothetical. Results depend on the inputs and are not a prediction of future performance. This is educational content, not financial advice.