Performance rating: what your result was worth

The rating a player would need in order for their result in this event to be the expected one. It is calculated from the tournament alone and ignores the player's own rating entirely.

• It is independent of your current rating. Which is exactly the point: it measures this event rather than your history.
• It depends on two things only. The average rating of your opponents and the share of points you scored against them.
• It can be far above or below your rating. A good tournament shows a performance two hundred points higher; a bad one, the reverse.
• It is not a rating change. The rating moves by a fraction of the gap; the performance is the raw signal, not the correction - how ratings change.

The version you can do in your head at the end of an event, and it is accurate enough for club purposes.

An example: seven rounds, opponents averaging 1700, and you score five wins, one draw, one loss. The difference is four, divided by seven games, times 400 - about 229. Your performance is roughly 1929.

Two consequences worth noticing. Scoring 100% makes the formula unusable - it would give “average plus 400” regardless of how dominant the win was. And a single game tells you almost nothing: the shorter the event, the noisier the number.

Official calculations do not use the rule of 400. They use a conversion table mapping the percentage scored to a rating difference.

1. 1Work out the average rating of your opponents
2. 2Work out your score as a percentage of the games played
3. 3Look that percentage up in the conversion table to get a rating difference
4. 4Add the difference to the average opponent rating

The table is the same one used for expected scores, read backwards: 50% gives a difference of zero, 76% gives about +200, 90% about +366. It is capped at the extremes, which is how the method avoids the infinity problem that breaks the rule of 400 at 100%.

The details, including how unplayed games are treated, are set in the current FIDE rating regulations and are revised from time to time - check those rather than any article if the number matters officially.

• It shows who actually played well. Fourth place with a performance of 2000 is a better tournament than second place with 1750 - and worth mentioning at the prize-giving.
• It is the fairest basis for a “best result” prize. Especially in a field of mixed strength, where the top places are always taken by the same people - prizes for a mixed field.
• It identifies improving players. Somebody whose performance has climbed two hundred points over three events is worth telling - that is the feedback that keeps people coming.
• It sanity-checks your seeding. If performances are wildly out of line with your seeding, the ratings you used were wrong.

At a club without ratings you can still use it: substitute your internal club rating for the official one and the arithmetic works the same way - visible progress and retention.

This is where the number stops being informational and becomes a requirement.

• A norm is defined by performance. Roughly 2600 for a grandmaster norm and 2450 for an international master norm - the exact figures are in the FIDE handbook.
• The event must qualify. Titled opponents, several federations, a classical control and advance registration. A high performance in an ordinary open produces nothing - what a norm requires.
• There are floors on opponent ratings. Very low-rated opponents are counted at a minimum figure, so a performance cannot be inflated by beating unrated players.
• A norm needs a minimum number of games, normally nine.

For an amateur event none of this applies - which is worth saying plainly, because “my performance was 2400” at a club blitz evening means something quite different from a norm.

Use it as it is meant: a way to compare results within one event, and to track your own progress across several. Not as a verdict on a single evening.