Risk/Reward Ratio: How to Calculate It and What It Means for Your Win Rate

The risk/reward ratio of a trade compares the amount you stand to lose if the price hits your stop with the amount you stand to gain if it reaches your target. It is set before entering, from three prices: the entry, the stop and the target. It does not tell you whether the trade will work; it tells you how often it must work for the strategy to pay.

The diagram shows the pieces: the distance from entry to stop is the risk per share, the distance from entry to target is the reward, and here the target sits twice as far as the stop. The position size then turns the risk per share into the amount risked.

How to calculate the risk/reward ratio

For a short trade the signs flip: the stop is above the entry and the target below. A short sold at $50 with a stop at $52 and a target at $44 is also 1:3. Its return when it reaches the target is (entry − exit) ÷ entry = (50 − 44) ÷ 50 = 12%.

Watch the order of the two numbers. This page writes risk first, so 1:3 means risking 1 to make 3; many brokers and sites write reward first and call the same trade 3:1. Quoting the reward in R, as in 3R, avoids the confusion.

The win rate each ratio needs

If a trade wins R times its risk or loses 1 times its risk, it breaks even when the win rate W satisfies W × R = (1 − W), which gives W = 1 ÷ (1 + R). The expected result per trade, the expectancy, is worked out in trading math; here is the break-even line alone:

Costs move the line up: commissions and the bid-ask spread are paid on winners and losers alike, and weigh more on small targets.

What stops and targets in ATR multiples gave

A wider target wins less often. To see by how much on real prices, we entered a hypothetical trade at the close of every day in two daily series, with a stop at 1 ATR(14) (the average true range of the last 14 days) and a target at 0.5, 1, 2 or 3 ATR. The ATR is Wilder's: its first value is the average of the first 14 true ranges, the very first being that bar's high minus its low, and each later value is 13 times the previous one plus the day's true range, divided by 14. The first later bar that touches the stop or the target decides the trade; a bar that touches both counts as a loss, since daily bars do not say which came first. Results are in R, before costs. The S&P 500 ETF series (dividend-adjusted, January 4, 2021 to October 28, 2026) gives 1,449 entries, from January 22, 2021, and the Bitcoin series (March 8, 2025 to October 28, 2026) 587, from March 21, 2025. Trades still open at the end of the data are left out: from 1 to 10 per row on the ETF and from 4 to 17 on Bitcoin.

On the ETF with a 1 ATR target, 767 long trades won, 666 lost, 14 bars touched both and 2 were still open. What the table shows, and what it does not:

• The win rate falls as the target widens, close to the break-even line each time: at 0.5 ATR, 66.5% won on the ETF against 66.7% needed.
• The ratio did not create an edge. Long trades made a little from 1 ATR up and about broke even at 0.5 ATR; short trades lost at every ratio. The ETF rose from 341.59 to 772.31 over the series, so a long bias helped. Bitcoin ended lower (from 86,222.45 to 81,343.50) yet its longs still beat its shorts; one 20-month series supports no rule.
• The trades overlap. A trade starts every day while earlier ones run, so the 1,449 ETF entries are not 1,449 independent results. Year by year, the 3 ATR long result on the ETF ran from −0.27R in 2022 to +0.71R in 2023. Resampling the entries in blocks of 60 consecutive trades (moving blocks, 10,000 draws, seeds 1 to 10 of a mulberry32 generator) gives a 95% range that includes zero for 14 of the 16 cells at every seed. The two others are the ETF 0.5 ATR short, clearly negative (about −0.16R to −0.06R), in the direction of the ETF's rise, and the ETF 3 ATR long, whose lower end runs from −0.008R to +0.009R across the ten seeds and blocks of 40, 60 or 90 trades: it includes zero with some settings and not with others, so neither it nor the ranking between targets is an edge.

High-probability trading: why the win rate alone misleads

A 90% win rate sounds safe. With a reward of 0.1R per win and a loss of 1R per loss, its break-even win rate is 1 ÷ 1.1, about 90.9%: at 90% it loses money. A high win rate comes from a small target or a wide stop, and the rare loss erases many wins. A strategy that wins a third of the time can be the better one if it wins three times what it loses.

Each bar is (win rate × reward) − (loss rate × 1R), in R per trade, before costs. What counts is the win rate and the ratio together, and how long a losing streak can last before the account is hurt, which is the subject of risk of ruin.

A typical high-probability setup, as our earlier article on high-probability trading described it, combines four parts: a trend filter (higher highs and higher lows, or price above a moving average), an entry on a pullback to support or on a confirmed breakout, a stop just beyond that level or below the signal candle, and a near target such as the last swing high. Each part aims to raise the win rate; the near target also shrinks R, which is why the break-even line above decides whether the setup pays. We have not tested this setup on our data, and no win rate is implied.

Trade like a casino: edge and repetition

A casino does not need to win most bets. It needs a small edge on every bet and a large number of bets.

The chart shows why repetition matters. One bet has a standard deviation of about 5.84 times the stake, so over 100 spins the player's average result per $1 can land almost anywhere between a large loss and a large gain. Over 10,000 spins the range narrows, and over a million it sits entirely below zero, around the −2.7% edge.

A trader who wants to "trade like a casino" needs the same three things: a positive expectancy per trade (win rate and ratio together), a bet size small enough to survive the losing streaks, and enough trades for the edge to show. The difference is that the casino knows its edge exactly, from the rules of the game; a trader only estimates it from past trades, and it can change.

Is trading gambling?

Both involve risking money on an uncertain result. The differences are in the odds and in the control:

• The odds. A casino game has a known negative expectancy for the player. A market has no fixed odds; a trade's expectancy depends on the method, the costs and the period.
• The exit. A bet is lost in full. A trade can be closed at a stop for a planned part of the stake.
• The behavior. Trading without a defined risk, raising the size after losses or trading for the thrill is gambling with a different interface. The signs are in the dangers of overtrading.

More definitions are in our trading concepts.