Why a High Risk-Reward Ratio Can Still Lose Money: Win Rate, Expectancy and Position Sizing

Key takeaway: The risk-reward ratio is merely the proportion of a single winning trade to a single losing trade; it cannot answer whether a method will be profitable in isolation. What truly determines long-term results is the trade expectancy: the combined effect of win rate, average profit, average loss, transaction costs, and execution bias. Even if the planned target is twice the stop-loss, if the actual win rate is below the break-even point, or winning trades are frequently exited prematurely, the account will still experience continuous drawdowns.

First, separate the three easily confused numbers.

The "planned risk-reward ratio" comes from the stop-loss and target set before entering the market. For example, being willing to take $1 of risk to target $2 of profit is denoted as 2R. The "actual risk-reward ratio" is the average profit divided by the average loss in the sample. The "expected value of the trade" measures how much R each trade can bring on average in the long run. The three are not equal. Gaps, slippage, staggered profit-taking, trailing stops, and manual intervention can all turn the planned 2R into 1.3R or even lower in the sample.

If p represents the win rate, W represents the average profit, and L represents the average loss, then:

Trade expectancy E = p × W − (1 − p) × L − average transaction cost

Standardizing the average loss as 1R and the average profit as bR, without considering costs, the break-even winning rate is:

Break-even win rate = 1 ÷ (1 + b)

Therefore, a win rate of at least 50% is required when the actual risk-reward ratio is 1:1; 1.5:1 corresponds to 40%; 2:1 corresponds to approximately 33.3%; and 3:1 corresponds to 25%. This is only the mathematical break-even point, not the safety line that the strategy should achieve, because costs and execution errors will continue to push the threshold higher.

A complete calculation example: Why does the 2R objective ultimately only yield an expected value of 0.18R?

Assuming an account balance of $100,000, the maximum risk per trade is 0.75%, or $750. A stock is planned to be bought at $100, with a stop-loss order placed at $97.50. After adjusting for transaction fees and estimated slippage to $0.10 per share, the true risk per share is $2.60. The acceptable amount should not be calculated as 750 ÷ 2.50, but rather based on the true risk.

Position size = 750 ÷ 2.60 = 288 shares (rounded down)

If the target price is $105, and taking into account the same cost of $0.10, the net profit per share is about $4.90. The actual risk-reward ratio is 4.90 ÷ 2.60 ≈ 1.88R, instead of the 2R seen on the chart.

After reviewing a sufficient number of trades with consistent patterns, it was found that 38% of the trades reached the target, 52% triggered stop-loss orders, and another 10% exited near the cost price but with an average loss of 0.15R. Therefore:

E = 0.38×1.88 − 0.52×1 − 0.10×0.15 ≈ 0.18R

This method still has a positive expected value, but its advantage is far less than the intuitive feeling of the "2R target". If the actual win rate drops from 38% to 32%, with other conditions remaining unchanged, the expected value will shrink to about 0.07R; a little more slippage or two failures to stop loss as planned could turn it into a negative value.

Five reasons why your account is still losing money even though the risk-reward ratio looks high.

1. Mistaking the target price for the average realized win

A target is a design input; realized trade outcomes are statistical evidence. Trend strategies may occasionally achieve a 4R (4% return), but many profitable trades only achieve a 0.5R; showing only the best trade will overestimate the long-term advantage.

2. Underestimating losses beyond the planned stop

Overnight gaps, insufficient liquidity, slippage of stop-loss orders, and widening spreads can all cause average losses to exceed 1R. This type of deviation is particularly prevalent in forex, gold, and index futures before and after major data releases.

3. Win rate is not a constant.

The same set of entry rules can have completely different hit rates during trending, consolidation, and expansion phases of volatility. Applying the win rate of a bull market sample directly to a sideways market is equivalent to hiding environmental changes within the average.

4. Profits and losses do not occur independently or evenly.

Even positive expected value strategies can experience consecutive losses. If each trade carries 5% account risk, six consecutive losses will reduce capital by approximately 26.5%; a return of about 36% is needed to recover. Overly large positions can prevent the realization of correct statistical advantages.

5. The sample is affected by selection bias.

Only analyzing charts that are "understandable," deleting trades exited prematurely, or using data known only after the fact in backtesting can create falsely high risk-reward ratios. A valid sample must include all signals that conform to the written rules.

How to apply the risk-reward ratio to real-world decision-making?

The first step is to standardize the recording method using the denominator R. For each transaction, an initial risk of 1R is defined, and the final result is recorded as +1.4R, -1R, or -0.2R. This allows for comparison of different prices, instruments, and positions on the same scale.

The second step is to separately analyze bullish, bearish, and market conditions. At least separate trends from ranges, and normal volatility from high volatility. If the advantage exists only in one environment, then trading filters are more important than improving the nominal risk-reward ratio.

The third step involves reporting four metrics simultaneously: win rate, average profit, average loss, and expected value per trade. These are all essential; in addition, the maximum consecutive loss, maximum drawdown, and profit factor must be included to determine whether the profit path is sustainable.

The fourth step is to conduct a cost stress test. Increase the spread, commission, and slippage by 25% and 50% respectively, and recalculate the expected values. Strategies that can turn negative results from slight cost changes usually lack sufficient buffer.

Fifth, use fixed risk to deduce position size. First, determine the stop-loss based on the chart structure, rather than deciding how much to buy. The position size formula is: Tolerable account risk ÷ (Difference between entry price and stop-loss price + unit cost).

An executable pre-trade checklist

• Is the stop-loss determined by structural failure, rather than by trying to achieve a nice 2R?
• Is the target price supported by liquidity, previous highs and lows, or volatility?
• Is the actual average profit or loss being used, or the planned value per trade?
• Have spreads, commissions, slippage, and overnight costs been factored in?
• Does the current market environment align with the statistical sample?
• After six consecutive losses, is the account drawdown still within an acceptable range?

The correct use of the risk/reward ratio is not to replace judgment, but to put entry, exit, and position sizing into the same probabilistic framework. Only by first verifying positive expected value, and then controlling individual risk and drawdown path, can a "good opportunity" on the chart be transformed into sustainable trading results.