Win Rate vs Avg Return: How to Pick Better Trades

You screen stocks by average return, not just win rate, by treating a screen's Win Rate and Avg. Return as one pair, never one number. A high win rate with a small average return can still lose money if the losers are big enough, while a lower win rate with a strong average return often carries more real edge. ChartMath shows both fields on every screen so you can check that math before you risk anything.

Key Takeaways

• Win rate hides losing math: a 70% win rate can still be net negative if average losses run 3x bigger than average wins.
• Average return measures payoff, not frequency: it tells you how much a screen's typical trade actually made or lost across its backtested sample.
• Expectancy is the number that matters: (win rate x average win) minus (loss rate x average loss) is the only figure that shows a real mathematical edge.
• Sample size changes everything: a screen with 15 backtested trades tells you almost nothing, even with strong stats.
• ChartMath pairs both metrics on every screen: Win Rate and Avg. Return sit side by side on each screen's Strategy Analytics, so you never judge a setup on one number alone.

At a Glance: Win Rate vs Avg Return

Why Win Rate Alone Can Mislead You

A screen that wins 8 out of 10 trades sounds like an easy decision. But if those two losses average -6% each while the eight wins average only 1% each, the math flips. Eight wins at 1% is 8%. Two losses at 6% is -12%. That screen loses money overall, despite a 80% win rate.

This is the exact trap covered in why high win rate doesn't equal profitable trading. Win rate answers "how often," not "how much." A trader chasing setups purely on a high percentage badge is optimizing the wrong number.

If you're newer to the concept of backtested win rate itself, start with what is backtest win rate in trading setups and does it actually matter, then come back here to see how average return fills in the rest of the picture.

What Average Return Actually Tells You

Average return is the mean gain or loss per trade across a screen's full backtested history, win and loss trades combined. It's the payoff side of the equation win rate leaves out.

According to Tradeways, the payoff ratio, average win divided by average loss, tells you how many dollars a typical winner returns for every dollar a typical loser costs. A payoff ratio above 1.0 means your average win outweighs your average loss, but it says nothing about how often you actually win. That's why you need both numbers, not one.

A screen with a 45% win rate and a strong average return can beat a screen with a 75% win rate and a thin one. The percentage alone never tells you which.

How Do You Combine Win Rate and Average Return?

You combine them using expectancy: multiply win rate by average win, subtract loss rate multiplied by average loss. The result is the average dollar or percent result you can expect per trade over many trades, and it's the single number that confirms whether a strategy has real edge.

StaxInvesting lays out the same formula: expectancy equals win rate times average win, minus loss rate times average loss. Say Screen A has a 40% win rate with a 4% average win and a 1.5% average loss. Its expectancy is (0.40 x 4) minus (0.60 x 1.5), or 1.6 minus 0.9, which is +0.7% per trade. Screen B has a 65% win rate with a 1% average win and a 2% average loss: (0.65 x 1) minus (0.35 x 2), or 0.65 minus 0.7, which is -0.05% per trade. Screen B has the higher win rate and the worse expectancy.

Research from Traders' Second Brain makes the point sharply: a 40% win rate at a 3:1 risk-reward ratio can be more profitable than a 70% win rate with a poor payoff ratio. Win rate without a payoff number attached is close to meaningless.

1. Check the Sample Size Before You Trust Either Number

A screen's win rate and average return only mean something once they're computed across enough trades to smooth out noise. A screen showing 12 backtested trades can flip from 80% to 40% with a single new result.

Look for the sample size next to the stats, not just the headline percentages. If a screen shows a strong win rate but a tiny trade count, treat it as unproven. This is the same trap covered in paper trade by sample size not calendar days: judging a strategy by a handful of results, whether backtested or lived, produces the wrong lesson almost every time.

2. Compare Win Rate and Avg Return Side by Side on the Same Screen

The fastest way to judge a screen is to open its Strategy Analytics and look at both fields at once, not one, then the other in a different tab. That's the entire point of pairing them on the same card.

ChartMath shows Win Rate and Avg. Return together on every one of its 200+ screens, backtested across a fixed universe of 500+ US equities. Every screen is a deterministic rule, so its record can be recomputed instead of taken on trust, unlike a tip forwarded in a group chat.

Take a screen like RSI Overbought Fade. Its Strategy Analytics tab shows the exit rule, the sample size, and both stats in one view. You never have to guess which number the app is quietly optimizing for. If you want a deeper primer on reading that win rate field specifically before you look at anything else, how to use backtested win rate to pick trades covers it step by step.

3. Weight Avg Return Higher for Momentum and Breakout Setups

Breakout and momentum setups tend to run a lower win rate with a bigger average return, because a chunk of entries fail quickly while the ones that work run far. Mean reversion setups tend to run the opposite: higher win rate, smaller average return, because the trade is designed to close fast at a modest target.

Neither shape is wrong. The mistake is applying one screening standard to both. If you're evaluating a Golden Cross or an Opening Range Breakout screen, a 45-50% win rate paired with a solid average return can be a legitimate setup. Judged only on win rate, it would look weak next to a mean reversion screen that never was built to win that often.

Golden Cross and Bollinger Band Setups: A Quick Case Study

Traders searching for "golden cross stocks today" are usually looking for a scan, not a lecture on moving averages. The useful follow-up question is what that screen's backtested win rate and average return actually look like before adding it to a watchlist.

The Golden Cross screen (daily), SMA(50) crossing above SMA(200), tends to trigger less often and hold for longer, so its average return per trade carries more weight than its win rate. Traders asking about "stocks at bottom of bollinger bands" are looking at a mean reversion setup, the below lower Bollinger screen (daily) stated as a fixed rule, where the win rate usually runs higher and the average return per trade runs smaller by design. Comparing the two screen types by win rate alone would make the Bollinger setup look automatically superior, which isn't a fair read once average return and typical hold time enter the picture. The deeper mechanics of that particular indicator are covered in Bollinger Bands: Two Trades, One Indicator.

4. Watch for Red Flags in Backtested Stats

Some patterns should make you pause before trusting a screen's numbers at all.

• No sample size shown: a win rate without a trade count behind it can't be checked.
• Only a win rate, no average return: if a tool won't show payoff, assume it's hiding a weak one.
• Round, suspiciously high numbers: a 95% win rate on a real technical setup is rare enough to warrant scrutiny, not celebration.
• "Accuracy" instead of win rate and average return: that phrasing usually signals a single blended metric, not the two separate numbers you need to judge edge.
• A tool that surfaces a call from a forwarded Discord or Telegram message with no recomputable backtest attached at all.

Putting It Into Practice With a Watchlist

Start small. Pick three to five tickers you already follow and check both stats on each screen you're considering, the consistent uptrend screen (daily) is a reasonable place to start, before turning alerts on. That's the same approach covered in how to pick stocks for swing trading: a rules-based method.

Once you've confirmed a screen's win rate and average return look reasonable for its sample size, rehearse it with a simulated order before committing real capital. ChartMath's paper trading pre-fills stop, target, and share count off a capital-split sizing method, so you can watch a screen play out with the same numbers you'd trade live, tracked in a Portfolio tab.

If you're a US equity swing trader balancing this against a day job, how to trade stocks with a day job without missing entries walks through fitting that review into a realistic daily schedule.

FAQ

Is a 70% win rate always good?

No. A 70% win rate is only good if the average loss on the remaining 30% doesn't erase the gains from the winners. Check the average return per trade before assuming a high win rate equals a profitable setup.

What counts as a good average return per trade?

There's no universal number; it depends on holding period, sample size, and the screen's typical trade frequency. A modest average return with a large sample size and a reasonable win rate usually beats a flashy average return backed by a handful of trades.

Does ChartMath show an "accuracy" score?

No. ChartMath's card metrics are Win Rate and Avg. Return, shown together on every screen's Strategy Analytics so you can weigh frequency against payoff before you act, rather than relying on one blended figure.

Any backtested figure referenced here is historical and hypothetical performance, not a forecast or promise of future results.

You can check this pairing yourself right now. Browse the full catalog with the web based screener, or download the app to see Win Rate and Avg. Return on every one of ChartMath's 200+ backtested screens before your next trade.