A simple comparison
The risk-reward ratio compares what you stand to lose with what you stand to gain on a single choice. If you risk 1 point to try to gain 3, the ratio is 1 to 3, often read as “risk one to make three”. It is a quick way of describing how lopsided a decision is, and it works in any setting where you can put the downside and upside in the same units: points in a board game, resources in a strategy game, chips at a table.
The ratio is only half the story. It says nothing about how likely each side is. A choice with a handsome ratio can still be poor if the gain is rarely achieved, and a modest ratio can be sensible when success is very likely. To use the ratio well, you pair it with a probability.
The break-even chance
The ratio tells you the chance of success you need to break even on average. If you risk R to gain G, the break-even probability of winning is:
R / (R + G)
Check it with an example. Risk 1 to gain 2. Break-even chance = 1 / (1 + 2) = 1/3, about 33 percent. If your chance of success is exactly 1/3, then on average you gain 2 a third of the time and lose 1 two thirds of the time:
(1/3) x 2 - (2/3) x 1 = 2/3 - 2/3 = 0.
The average is zero, as expected. If your chance is above 1/3, the choice is favourable on average, and if it is below, it is not.
More examples:
- Risk 1 to gain 3: break-even chance = 1/4, or 25 percent.
- Risk 1 to gain 1: break-even chance = 1/2, or 50 percent.
- Risk 3 to gain 1: break-even chance = 3/4, or 75 percent.
The last case matters. Small, frequent gains paired with rare large losses can look safe, because the gain usually arrives, yet the break-even chance is high. If your true chance is only 70 percent, the choice loses on average: 0.7 x 1 - 0.3 x 3 = 0.7 - 0.9 = -0.2.
Connecting it to expected value
The break-even chance is just the expected value rule rearranged. The average result of a choice is the chance of success times the gain, minus the chance of failure times the loss. Setting that equal to zero gives the formula above. The practical value of the ratio is speed: you can judge a choice in your head without a full calculation, by asking whether your chance of success clears the bar the ratio sets.
The same thinking sits behind pot odds in poker, where the price of a call is compared with the size of the pot to find the chance of success needed to continue.
Everyday examples
A board game gamble. You can spend 2 resources for a 40 percent chance at a prize worth 5 resources. The risk is the 2 you spend and the gain is 5 if you win, so with R = 2 and G = 5 the break-even chance is 2/7, about 29 percent. At 40 percent the gamble is favourable on average: 0.4 x 5 - 0.6 x 2 = 2.0 - 1.2 = +0.8.
A video game run. You could attempt a hard challenge that risks losing 50 minutes of progress for a reward you value at 100 minutes of play. The break-even chance is 50 / 150, which is 1/3. If you believe you have much less than a one-in-three chance, skip it.
These examples only work if the units are comparable. A reward you value for fun and a loss you value in hours cannot always be put on one scale, so decide what you are measuring.
What the ratio cannot tell you
How sure you are of the chance. A ratio is easy to compute, while the probability is usually an estimate. If you overestimate your chance of success, a ratio that looked good on paper is not. This is a common source of error, closely linked to risk perception, since vivid prizes tend to feel more likely than they are.
What you can afford. A choice that is favourable on average can still be unwise when the loss is something you cannot absorb. Staying in the game matters, which is where bankroll concepts and risk of ruin enter. A ratio treats all losses as equal in kind; in practice, one big loss can end your participation.
Repeated exposure. One favourable gamble differs from taking the same one many times. Over repeated choices, the order of wins and losses and the size of your stake determine how bumpy things get.
Sunk costs. What you have already put in does not belong in the ratio. Only the gain and loss from this point on count. Letting past spending raise the stakes of the next choice is the sunk cost trap.
Casino-style games
In games with a house edge, the payouts are arranged so that the break-even chance is slightly higher than your real chance. A bet can pay at an attractive ratio, and still be unfavourable, because its chance of winning is lower than the payout implies. In such a game, no choice of ratio fixes the average. Treat money wagered as an entertainment cost set in advance, never borrow to play, and talk to a local support service if gambling is causing problems.
A quick routine
- Write the loss and the gain in the same units.
- Compute the break-even chance, R divided by R plus G.
- Estimate your chance honestly, and consider a lower figure too.
- If your chance clears the bar by a wide margin, the choice is favourable on average. If it is close, treat it as a toss-up.
- Check that you can afford the loss.