Blog
How to Read a Probability Table
Percentages, odds, and 'one in n' say the same thing badly in three different ways.
Percentages, odds-against and one-in-n describe the same number and mislead in three different ways. Most arguments about probability are really arguments about which form somebody used.
This is a short guide to reading whichever one you have been handed.
The three forms
A percentage says how often something happens out of a hundred. Twenty-five percent means twenty-five times in a hundred.
One-in-n says how often out of n. One in four is the same as twenty-five percent, and the temptation to read it as one specific outcome out of four attempts is the error it invites.
Odds-against compares failures to successes. Three to one against is also twenty-five percent, and it is the form most likely to be misread by anyone who does not use it daily, because it counts the wrong side.
How each one misleads
Percentages flatten very small numbers. A 0.01% chance and a 0.001% chance look similar and differ by a factor of ten, which is why extremely rare events are better expressed as one-in-n.
One-in-n invites the false belief that n attempts guarantee one success. Four attempts at a one-in-four event give roughly a 68% chance of at least one success, not certainty, and the gap widens as n grows — a hundred attempts at one-in-a-hundred is about 63%.
Odds-against gets misread as a percentage constantly. Three to one is not 33% and not 75%; it is 25%, and the number of people who will confidently tell you otherwise is high.
Converting between them
Percentage to one-in-n: divide 100 by the percentage. 25% becomes one in four; 4% becomes one in twenty-five.
One-in-n to percentage: divide 100 by n. One in eight becomes 12.5%.
Odds-against to percentage: for odds of a to b against, the probability is b divided by (a plus b). Three to one against is 1/(3+1), which is 25%.
The conversion worth memorising is the last one, because it is the one people get wrong and the one gambling contexts use.
The independent-repeats trap
The most consequential misreading is treating repeated attempts as additive. Two attempts at a 25% chance is not a 50% chance.
The correct calculation is one minus the chance of failing every time. Two attempts at 25%: 1 − 0.75², which is about 44%. Four attempts: 1 − 0.75⁴, about 68%.
The rule of thumb worth carrying is that n attempts at a one-in-n chance gives roughly 63% — never certainty, and the figure barely moves for any n above about ten.
What a table should show
A well-built probability table gives one form consistently, states the denominator, and shows the cumulative figure where repeated attempts are relevant.
The tools on this site show percentages with the pool size alongside, because the pool size is what makes a percentage interpretable — 25% from a pool of four and 25% from a weighted pool of forty mean different things about how the number was arrived at.
If a table gives you a probability without a denominator, the useful question is what it is a probability of, and it is surprisingly often not what you assumed.
The two-column habit worth adopting
If you write probability tables for other people, the cheapest improvement is to give the same number twice in two forms. A column of percentages beside a column of one-in-n figures costs nothing and removes most of the misreading in one step.
The reason it works is that the two forms fail in opposite directions. Percentages flatten small numbers so rare events become indistinguishable; one-in-n exaggerates them and invites the belief that n attempts guarantee a hit. A reader with both in front of them tends to land on the right interpretation without being told which to use.
For anything where repeated attempts are relevant, add a third column: the cumulative chance across the number of attempts people will actually make. That is the number they were trying to work out, and it is the one they get wrong most reliably when left to compute it themselves.
Where the forms come from
The three notations exist because they came from different places, which is why none of them is obviously the right default.
Odds-against is the oldest of the three in common use and it comes from betting, where the form is doing a specific job: it states what you get back relative to what you staked. Three-to-one against means a winning stake of one returns three plus your stake, which makes the notation genuinely convenient in the setting it evolved for and genuinely confusing everywhere else.
Percentages come from arithmetic rather than from any application, and their weakness — flattening small numbers — is the price of a fixed denominator. One-in-n comes from counting and inherits its problem from that: it invites you to think in whole attempts, which is exactly the intuition that produces the belief that n tries guarantee a hit.
Knowing which setting a table came from usually tells you which form it will use, and therefore which misreading to guard against before you start.
Frequently asked questions
What is three-to-one against as a percentage?
25%. For odds of a to b against, the probability is b divided by (a plus b).
Do four attempts at one-in-four guarantee a success?
No. It is about 68% — one minus 0.75 to the fourth power.
What is the rule of thumb?
n attempts at a one-in-n chance gives roughly 63%, and that figure barely moves for any n above about ten.
Which form should I use for rare events?
One-in-n. Percentages flatten small numbers — 0.01% and 0.001% look similar and differ tenfold.
How do I convert a percentage to one-in-n?
Divide 100 by the percentage. 4% becomes one in twenty-five.
Why does a table need a denominator?
Because 25% from a pool of four and 25% from a weighted pool of forty mean different things about how the number arose.