Forge your chess rating.

• K = 40 when you are new to the rating list - until you have completed events with at least 30 rated games.
• K = 40 for all players until the end of the year of their 18th birthday, as long as their rating stays under 2300.
• K = 20 as long as your rating remains under 2400.
• K = 10 once your published rating has reached 2400 - and it stays at 10 permanently, even if you later drop below 2400.

In short: most adult club players who have played 30+ rated games sit at K = 20. New players and juniors move twice as fast with K = 40, which is why ratings converge quickly at first.

One more clause from the same article: if the number of games in a rating period multiplied by K would exceed 700, K is reduced to the largest whole number that keeps K × games at or below 700.

The FIDE method works in two steps. First take the average rating of your opponents, R_a. Then convert your percentage score p into a rating difference d_p using FIDE Table 8.1.1, and add it:

TPR = R_a + d_p

Scoring 50% gives d_p = 0, so your performance equals the field's average. A perfect 100% score gives d_p = +800 - which is really a lower bound: any 100% score against the same opponents produces the same number. An alternative convention anchors the same delta to your own rating instead: performance = your rating + d_p. A third, cruder method ("the algorithm of 400") adds 400 × (wins − losses) / games to the opponents' average.

Try the performance rating calculator →

Each game, your expected score is E = 1 / (1 + 10^{(R_opp − R_you)/400}). After the game your rating changes by ΔR = K × (S − E), where S is your actual score (1, ½ or 0) and K is your development coefficient. Beat a higher-rated player and S − E is a large positive number; lose to a lower-rated one and it bites hard in the other direction.

FIDE itself reads E from its official table (Table 8.1.2) rather than the raw curve - at a 400-point gap the table gives exactly 0.92, where the curve gives about 0.909. Both are shown side by side in our elo calculator.

For rating purposes, FIDE caps the rating difference between you and a single opponent at 400 points (Rating Regulations, art. 8.3.1). If you are rated 1500 and lose to a 2200 grandmaster, the game counts as a 400-point upset loss, not a 700-point one - you lose only what a 1900-rated opponent would lose. From 1 October 2025 the cap applies to players rated below 2650; players rated 2650 or above have their true difference counted, so elite mismatches move ratings again.