What a strategy based on odds is
A strategy, in this sense, is a rule for choosing among options when the result is uncertain. Probability does not tell you which rule to use. It gives each rule something to work with. Different rules reflect different attitudes to risk, and they can recommend different moves in the same situation. Understanding the main ones helps you see why people disagree about the “right” play, and which rule fits which kind of game.
None of these rules produces a guaranteed win, and none can turn a game with a built-in disadvantage into a profitable one. They are ways of organising a decision.
One table, three rules
Here is an invented example. You must pick one of three actions, A, B or C, before you know which of two situations will occur. Situation X occurs with probability 0.7 and situation Y with probability 0.3. The payoffs, in points, are:
| Action | If X (0.7) | If Y (0.3) |
|---|---|---|
| A | 4 | 4 |
| B | 10 | -6 |
| C | 9 | -1 |
Rule 1: maximise expected value
Weight each payoff by its probability and add.
- A: 0.7 x 4 + 0.3 x 4 = 2.8 + 1.2 = 4.0
- B: 0.7 x 10 + 0.3 x (-6) = 7.0 - 1.8 = 5.2
- C: 0.7 x 9 + 0.3 x (-1) = 6.3 - 0.3 = 6.0
This rule picks C. It uses all the information, both probabilities and size of payoffs, and it is the natural choice if you will face many similar decisions, because averages matter more as repetitions accumulate.
Rule 2: play the most likely outcome
Ask which situation is most probable, which is X at 0.7, and choose the action that does best in it.
- Best payoff if X: B with 10.
This rule picks B. It is simple and often works well when one outcome is very probable. It ignores the unlikely situation entirely, so it can be badly wrong when the neglected case is costly. Here B loses 6 points in Y, and Y still happens three times in ten.
Rule 3: minimise the worst case (maximin)
Find the worst payoff for each action and pick the one whose worst is best.
- A: worst is 4
- B: worst is -6
- C: worst is -1
This rule picks A. It guarantees at least 4 points whatever happens, which can matter if a bad result would be disastrous. The price is the lower average: 4.0 against 6.0 for C.
Reading the result
Three rules, three different choices. None is wrong in itself; each answers a different question.
- Expected value asks which option is best on average.
- Most-likely asks what to do if the likely case happens.
- Worst-case asks what protects against disaster.
In a single, high-stakes decision where a loss would be hard to recover from, the worst-case rule has real appeal. In a long series of small decisions, expected value usually wins out. The situation decides, not the arithmetic alone.
Other named ideas
Satisficing. Instead of seeking the best option, set a target and take the first option that meets it. It saves effort and is reasonable when the decision is minor.
Minimising regret. Compare each action with the best you could have done in each situation and choose the action with the smallest worst-case shortfall. It suits people who dislike second-guessing.
Mixing choices. In games against an intelligent opponent, always picking the same move can be exploited. Mixed strategies randomise between moves at set frequencies so that the opponent cannot predict you. The theory behind this is game theory, and it concerns the opponent’s choices as much as the odds.
Sizing stakes. Some strategies concern how much to risk rather than which option to take. The Kelly criterion is a well-known example, but it assumes that you know your edge exactly, which is rarely true, and it can produce large swings even when the assumptions hold.
Testing a strategy
It is easy to believe in a rule after a lucky run. A fairer test is to simulate it. A strategy testing exercise runs the rule many times against random or fixed opponents and records the average. Short tests mislead for the same reasons that short real-life runs do, so keep the number of trials large, and treat any result as an estimate.
Where the rules run out
Rules based on odds depend on probabilities and payoffs you can actually state. In many real games these are approximate, and in some they are hidden. A strategy is only as sound as its inputs.
Most importantly, no probability strategy defeats a negative expectation. In casino games with a house edge, every standard bet loses on average, so choosing among them only changes how fast the expected loss accumulates, not its direction. Be wary of anyone who sells a system that claims otherwise. If you play for money, treat the cost as entertainment, set limits beforehand, do not chase losses, and never borrow. If gambling is affecting you or someone you know, a local support service can help.
Choosing a rule in practice
- State the options and the situations, with honest probabilities.
- Work out the expected value of each option.
- Check the worst case of the leader. Can you absorb it?
- If you cannot, consider a lower-variance option, even at some cost in average.
- Record the decision and the reason, so you can review it later.
Used this way, the named strategies are tools for thinking clearly, not formulas for winning.