A repeatable way to choose
When two or three options all look plausible, intuition can stall or lean on whichever option felt best last time. A short method helps. It will not make decisions for you, but it forces the pieces into view: what you could do, how likely each thing is to work, and what it is worth.
The method below has six steps. It is designed to be practised away from the match first, such as in a review, and then compressed into a few seconds of thought when it counts.
Step 1: List the real options
Write down what you could do, including waiting or doing nothing. Two or three options are usually enough; if you list ten, you are probably including ones you would never choose. Keep the list honest. A frequent mistake is to compare a plan you like against a weak version of the alternatives.
Step 2: Describe each outcome
For each option, name the possible outcomes. Often there are just two, it works or it does not, but some options have a middle result, such as partial success. Describe what each outcome does for you in terms of the thing you care about, such as winning the round, keeping a lead or preserving a resource.
Step 3: Estimate the chances
Give each outcome a rough probability. Use fractions or round percentages. Be honest that these are estimates; they come from experience and from what you can see, and they are often wrong by some margin. For options involving things you cannot see, the uncertainty is wider, which is a good reason to hold the number loosely. Tasks involving hidden information, like guessing an opponent’s response, are better handled with the ideas behind mixed strategies, where the aim is not to predict one response but to avoid being predictable.
Step 4: Assign payoffs
Give each outcome a rough value. These can be arbitrary units as long as you are consistent. A gain might be +4, a small gain +2, a failure −1. You are scoring what matters to you, so a play that costs a key resource on failure should have a more negative score than one that merely wastes time.
Step 5: Combine and compare
For each option, multiply each outcome’s payoff by its chance and add them up. That gives an average value per option. The one with the highest average is the leading candidate.
A worked example with three options, using made-up numbers:
| Option | Success chance | If it works | If it fails | Average value |
|---|---|---|---|---|
| A: steady push | 80% | +2 | −1 | 0.8 × 2 + 0.2 × (−1) = 1.4 |
| B: bold flank | 50% | +4 | −1 | 0.5 × 4 + 0.5 × (−1) = 1.5 |
| C: hold position | 100% | +1.2 | none | 1.2 |
On these figures B is slightly ahead of A, and both beat C. But the gap between A and B is small, which brings us to the next step.
Step 6: Stress-test the estimate
The result is only as good as the inputs, so ask how wrong you could be before the answer changes. Suppose the bold flank actually works 40 percent of the time rather than 50. Its value becomes 0.4 × 4 + 0.6 × (−1) = 1.6 − 0.6 = 1.0. That is now below A’s 1.4 and C’s 1.2. A ten-point error in one estimate flipped the ranking.
That tells you something useful: the choice between A and B depends on a number you are not sure about. If you believe the flank is a genuine coin flip or better, B is reasonable. If you suspect the enemy is waiting for it, A is safer. When the answer is sensitive, there is little to be gained by agonising, and some robustness is worth more than a small edge on paper.
What the numbers leave out
The average ignores some real things.
- Variance and the situation. Averages treat a safe +1.2 and a risky gamble with the same average as equal. When you are ahead, you may prefer the steadier one; when behind, the riskier one. Sensible risk-based strategies depend on the position, not just the average.
- Information and tempo. An option can reveal your plan to the opponent or change the pace of the match.
- Goals. If a particular outcome is all that matters, such as needing a win rather than a good position, then goal-oriented decisions may call for a different weighting than average value.
- Correlated risks. If two things each work 50 percent of the time but are linked, they do not combine the way independent ones would.
Making it quick
In a real match, you will not write a table. Compress the steps into a quick mental version: what are my two real options, which is more likely to work, which pays more, and how bad is failing? If one option is both likelier and better on the downside, take it. If one is likelier and the other pays more, that is when the rough sum matters.
Practise the full method in review until the compressed version becomes automatic. Be prepared for outcomes that disagree with your estimates; any single result is weak evidence either way, and the method’s value lies in the quality of the choices over many matches rather than any single one.