What Is a Good Win Rate for a Trading Strategy? It Depends

There is no good win rate in isolation, and that is the honest answer rather than a dodge. A 30% win rate can be excellent and a 90% win rate can bankrupt you. What decides profitability is expectancy: your win rate multiplied by your average win, minus your loss rate multiplied by your average loss. A strategy that wins 35% of the time with wins three times the size of its losses makes money. One that wins 90% of the time and gives it all back in the tenth trade does not.

So the useful question is not "what win rate should I have" but "what win rate does my payoff ratio require". Those two numbers move together, and looking at either one alone is how traders end up proud of a losing system.

The only formula you need

Expectancy is what a strategy earns per trade on average, and it is the number that survives contact with reality:

Expectancy = (Win rate x Average win) - (Loss rate x Average loss)

Positive expectancy means the rule makes money over enough trades. Negative means it loses, no matter how good it feels. Run the numbers on a system that wins 40% of the time, makes $300 on winners and loses $150 on losers: (0.40 x $300) - (0.60 x $150) = $120 - $90 = $30 per trade. It wins less than half the time and it is a good system.

Now a system that wins 80% of the time, makes $50 on winners and loses $250 on losers: (0.80 x $50) - (0.20 x $250) = $40 - $50 = a loss of $10 per trade. It wins four times out of five and it is quietly bleeding you.

The break-even win rate for any payoff ratio

Given a payoff ratio (average win divided by average loss), the win rate you must clear just to break even before costs is fixed. This table is the whole article in six rows.

The formula behind the middle column is simply 1 / (1 + payoff ratio). Note that every row is break-even before costs. Add commissions and slippage and each threshold moves against you, which is why high-frequency, low-payoff systems die first when costs are honest.

Why high win rates are so seductive and so dangerous

Winning feels like being right, and being right is what most people are actually optimizing for without admitting it. This is why strategies with 85% and 90% win rates sell so well: the equity curve rises in a smooth, satisfying line, and every screenshot looks like competence.

The shape of the danger is always the same. High win rates are usually bought by taking small profits quickly and holding losers until they come back. That works, repeatedly, until the one that does not come back. Selling options naked is the purest version: you collect a small premium over and over with a win rate above 90%, and one gap eats two years of it. The win rate was never the risk. The payoff asymmetry was, and the win rate was hiding it.

Trend following sits at the opposite pole and is the more honest business. Win rates around 30% to 40% are normal, most trades are small losses, and a handful of enormous winners carry the whole record. It is psychologically miserable and mathematically sound.

What is a realistic win rate by strategy type?

These are ranges observed across common published systems, not promises, and any individual result depends entirely on the rule, the period, and the costs. Read them as "where this style tends to live", not as a target to hit.

How many trades before your win rate means anything?

More than you have. This is the most common way traders fool themselves: a 70% win rate across 20 trades is not a 70% win rate, it is 14 wins, and 14 wins happens by chance often enough that it should not move your beliefs much at all.

A rough working guide: under 30 trades, ignore the number entirely. Between 30 and 100, treat it as a hint. Above 100, and spread across different market conditions, it starts to be evidence. Above 300 you can begin to argue about the second decimal place.

And sample size is not just a count, it is a spread. Two hundred trades all taken during the 2023 to 2024 melt-up is one market wearing a large number. The rule has to face a bear market, a chop, and a crash before its win rate describes anything durable.

The numbers worth watching instead

1. Expectancy per trade. The one number that says whether the rule makes money. Everything else is diagnostic.
2. Maximum drawdown. The deepest peak-to-trough fall, how long it lasted, and how long it took to recover. This decides whether you could actually have held on, and a strategy you abandon has an expectancy of zero.
3. Profit factor. Gross wins divided by gross losses. Above 1.5 is respectable; above 2 across a large sample is strong and deserves suspicion about how it was fitted.
4. Longest losing streak. A 35% win rate mathematically implies streaks of eight or nine losses. If you have not decided in advance that you will sit through that, you will not.
5. Expectancy after costs. Recompute everything with commissions and slippage charged. Around 0.1% per trade is a reasonable assumption for liquid US stocks, and it is where most high-frequency edges go to die.

Past performance does not guarantee future results. For educational and informational purposes only. Not financial advice. Consult a licensed advisor.

Getting your real numbers rather than a rule of thumb

Everything above is arithmetic you can do on your own trades, and you should. But your own trades are a small, biased sample: they are the ones you took, in the market you traded, filtered by the entries you skipped because you were nervous. That sample will not tell you the win rate or the payoff ratio of the rule itself.

A backtest will. Type the thesis in plain English, such as "buy SPY when the 14-day RSI drops below 30 and sell above 55", and Agenttrading restates it as an explicit rule and shows it to you before running anything. Then it tests across 20+ years of split- and dividend-adjusted daily data with a 0.1% cost per trade charged by default, so the win rate, the payoff ratio, the trade count, and the worst drawdown all come out of a sample big enough to argue with. The verdict at the end is honest: HELD UP, MIXED, or UNDERPERFORMED. Plans start at $19 per month, with no execution and no brokerage connection.

Next: what is a good Sharpe ratio covers the risk-adjusted view of the same record, maximum drawdown explained takes the number that actually hurts, and position sizing turns expectancy into a bet size. On the tooling side, backtesting software explains the bench, investment risk analysis reads the risk in words, and do you need a trading journal covers the half of the picture a backtest cannot see.

Analysis and education only. Agenttrading does not execute trades, connect to a brokerage, or provide personalized investment advice.