Ask a room of traders what a good win rate is and you will get a number. Sixty percent, usually. Maybe fifty.
The number is meaningless. Not "depends on context" meaningless — genuinely, arithmetically unanswerable, because a win rate cannot be judged without knowing the one figure almost nobody calculates.
That figure is the break-even win rate: the percentage of trades you have to win just to finish level. It takes about four seconds to work out, it is fixed before you place a single trade, and it decides the outcome more reliably than the entry signal everyone spends their time arguing about.
This guide is the formula, a table you can read off, and four real backtests across 503 stocks showing what happens when a strategy lands just under the line.
The formula
Break-even win rate = stop ÷ (target + stop)
That is the whole thing. If you risk 3% to make 6%, it is 3 ÷ (6 + 3) = 0.333, so you need to win 33.3% of your trades to end level.
The logic is straightforward. Ten trades at that payoff: three wins earn 18%, seven losses cost 21%. You are behind. Four wins earn 24% against six losses costing 18%, and you are ahead. The tipping point sits between them, and the formula finds it exactly.
Note what is not in that formula: your indicator, your timeframe, your market, your entry rules. The bar is set entirely by your exits. Your strategy's job is to clear it.
The table
Most people use a handful of standard payoffs. Here is what each one demands:
| Target / stop | Payoff | You must win |
|---|---|---|
| +1% / −3% | 1 : 3 | 75.0% |
| +2% / −3% | 1 : 1.5 | 60.0% |
| +3% / −3% | 1 : 1 | 50.0% |
| +4% / −3% | 1.33 : 1 | 42.9% |
| +5% / −3% | 1.67 : 1 | 37.5% |
| +6% / −3% | 2 : 1 | 33.3% |
| +9% / −3% | 3 : 1 | 25.0% |
| +12% / −3% | 4 : 1 | 20.0% |
Read the top and bottom rows together. A strategy winning 70% of the time loses money at 1:3. A strategy winning 25% of the time breaks even at 3:1. The higher win rate is the worse strategy. This is why "what is a good win rate" has no answer.
What this looks like on real trades
We have run four well-known strategies across all 503 S&P 500 stocks, replayed bar by bar on real daily data. Every one is published in full, including the losers. Lined up against their break-even rates:
| Strategy | Exits | Needed | Got | Gap | Result |
|---|---|---|---|---|---|
| Support bounce | +5 / −3 | 37.5% | 54.2% | +16.7 | cleared |
| MACD crossover | +6 / −3 | 33.3% | 32.3% | −1.0 | lost |
| Hammer candle | +4 / −3 | 42.9% | 38.7% | −4.2 | lost |
| Breakout + volume | +6 / −3 | 33.3% | 25.6% | −7.7 | lost |
Four strategies, four different ideas about what makes price move, four different sets of rules. The one that made money is the one that cleared its line. The three that lost are the three that did not. Nothing else in the table separates them.
The gap does not have to be big
The MACD row is the one worth sitting with. It missed by one percentage point.
Across 1,761 trades, the average trade lost 0.15% — a figure so small that no trader alive would notice it, react to it, or remember it. There is no bad trade in that ledger to point at. Yet compounded across all 1,761, the equity curve fell 99.4% from its peak, including a run of 26 consecutive losses.
That is what "one point short" actually means. Not a near miss. Not "almost profitable." The whole account, lost invisibly, a sixth of a percent at a time.
The trap: you cannot just widen the target
The obvious move, once you see the table, is to stretch the target until the maths works. If the hammer needed 42.9% at +4/−3 and only got 38.7%, why not use +6/−3, where the bar drops to 33.3%?
Because the win rate is not independent of the target. A 6% move happens less often than a 4% move, so as you move down the table your win rate falls too. You are not lowering a fixed bar; you are lowering the bar and the jumper at the same time.
There is no formula for how much the win rate falls — it depends on how the instrument actually moves after your signal fires. That is precisely the thing a backtest measures and nothing else can tell you. The formula sets the target. Testing tells you whether you hit it.
How to use this before you trade
- Write down your target and stop before anything else. If you cannot state them as numbers, you do not have a strategy yet.
- Calculate stop ÷ (target + stop). Write that number down too. This is the bar.
- Backtest the strategy and read its win rate.
- Compare the two numbers. Not the profit figure — the two win rates. Profit on a short window can be luck; the gap between these numbers is structural.
- Treat "close" as a failure. A strategy landing within a point or two of break-even is not nearly viable. It is a coin flip that charges commission, and the MACD row above is what that costs.
You can run this on the strategy builder in a few minutes: build the rules from blocks, set the exits, and the backtest reports the win rate against real historical bars using the same evaluation code as the live scanner.
Two things this formula does not cover
Costs. Break-even here means break-even before commissions, spread and slippage. Every one of those pushes the real requirement above the formula's number, and the more frequently you trade the more it matters. Treat the formula's output as a floor.
Variable exits. The formula assumes a fixed target and stop. If you trail your stop, scale out, or exit on a signal rather than a price, wins and losses stop being uniform and the arithmetic needs replacing with the expectancy of the actual distribution. Fixed exits are simpler, and being able to calculate the bar in advance is a genuine argument for keeping them.
Within those limits, this is the closest thing to a free lunch in trading: a number that tells you what your strategy has to do, available before you risk anything, and ignored by almost everyone.