High win rate but still losing money? Find the leak

If you win most of your trades and the account still shrinks, the leak is almost never your entry. It's the size of the losses next to the size of the wins, and the fees you forgot to subtract.

Here's the math in one line. A trader wins 70% of the time, makes 40 dollars on a winner and loses 180 dollars on a loser. Each trade is worth (0.70 × 40) − (0.30 × 180), which is minus 26 dollars. Seven green tickets out of ten, and the account bleeds 26 dollars per trade.

The break-even win rate is the number you're actually trading against

With a 40 dollar winner and a 180 dollar loser, you need to win 81.8% of the time just to stand still. The formula is short: break-even win rate = average loss ÷ (average win + average loss). Here, 180 ÷ 220 = 0.818.

That's the line your setup has to clear before fees. Most traders never calculate it, so they compare their 70% against a feeling instead of against 81.8%.

Read the last two rows together, because that gap is where most accounts die. You plan a 2:1 trade, you take the money at 1.3:1 because green feels safe, and your required win rate jumps from 33.3% to 43.5% without anyone telling you.

Our break-even calculator runs these numbers for your own average win and loss, including the win rate you need after a drawdown.

Leak 1: the losses that go past the planned stop

A planned loss is the amount you accepted when you clicked. Anything beyond that is a second decision, usually made while the position is open and your judgment is expensive.

Log both numbers: the planned risk at entry and the realized loss at exit. If your plan risks 200 dollars and your average loss is 340, your stop is a suggestion, not a stop. In R-multiples that trade is minus 1.7R, and two of them erase three clean 1R winners.

The tell in your log: sort the last 50 trades by realized result, worst first, and read the bottom 5 one by one. Moved stops cluster there. So do the trades you added to.

Leak 2: winners cut before the plan says so

The other half of the same problem. You set a target at 2R, price hits 1.1R, and you close it because a red candle showed up. Nothing about that decision feels reckless, and it moves your break-even win rate up by 10 percentage points.

Track planned reward against realized reward per setup. If your plan targets 2R and your average winner is 0.8R, either your targets are fantasy or your exits are fear. Both are fixable, and they need different fixes.

Note what counts here: you compare each trade against its own plan, not against the high of the day. Trades that ran to your target while you were already out don't count as mistakes unless your rule said to hold.

Leak 3: costs, which never show up in the win rate

Fees, spread, funding and slippage come out of every ticket, and they eat a much bigger share of a 20 dollar winner than of a 200 dollar winner. A 30 dollar gross winner with 8 dollars of round-trip cost is a 22 dollar trade. Do that 100 times and 800 dollars of your year is gone before the market had an opinion.

If you want to see what friction does without risking money, the GlanWick simulator prices crypto trades with 0.4% fees and 0.01% to 0.05% slippage per order, so the PnL you see is net. A spreadsheet full of perfect fills teaches the opposite lesson.

Find your own leak in 20 minutes

1. Export your last 50 closed trades.
2. Add two columns: planned risk at entry, realized result after costs.
3. Compute average win, average loss, win rate, and then expectancy: (win rate × average win) − (loss rate × average loss).
4. Compute the break-even win rate: average loss ÷ (average win + average loss).
5. Tag the bottom 5 losses with what actually happened: moved stop, added to loser, oversized, traded after the daily stop, entered outside a listed setup.

If expectancy is negative and your win rate is above the break-even line, your data is inconsistent and you're missing costs somewhere. If expectancy is negative and your win rate is below the line, you already know which of the three leaks to open first.

GlanWick does this part for you: R-multiple distribution and expectancy in R on every plan, win rate per setup and the drawdown chart from Starter, profit factor and maximum drawdown on Pro. The journal itself is free, so the measurement doesn't cost you anything.

Change one variable, then measure 20 trades

The fastest way to learn nothing is to tighten stops, widen targets, cut size and switch markets in the same week. Pick the leak with the largest total damage in R, change that one rule, and keep everything else still for the next 20 trades.

A short list of changes that are actually checkable:

• Hard stop at the planned level, no exceptions, for 20 trades.
• One partial exit rule, written before entry, with a defined runner.
• A cap on how many trades you take after two losses.

Twenty trades won't prove a strategy. It will show you whether you can follow your own rule under pressure, which is the part that broke in the first place.

The uncomfortable version

A high win rate is a comfortable number because it arrives fast and feels like skill. Expectancy arrives slowly and asks a harder question: over the real mix of your wins and losses, after every fee and bad fill, does the method leave anything behind?

Keep the receipts for the trades you'd rather explain away. That's where the answer lives.