Performance Rating Calculator
Find your tournament performance rating with the FIDE dp table, the linear method and the exact zero-change rating.
What a performance rating says
A performance rating, often written TPR or Rp, is the rating at which a player's results in an event would have been exactly as expected. If your performance is above your rating, you played better than your rating predicts and your rating will rise; below it, it will fall. Performance ratings are also how FIDE decides whether a result is good enough for a title norm.
The FIDE method
FIDE calculates performance as Rp = Ra + dp. Ra is the average rating of the opponents, rounded to the nearest whole number. dp is a rating difference read from a table using the percentage score p, rounded to two decimals. The table is symmetrical: 50% gives 0, 60% gives +72, 64% gives +102, 75% gives +193, 90% gives +366, 99% gives +677 and 100% gives +800. Scoring 4½ from 7, 64%, against opponents averaging 2010 gives 2010 + 102 = 2112. Because the table is based on a normal distribution, its values differ slightly from the logistic Elo curve, and the table's cap at ±800 means a perfect score cannot give an unlimited performance.
The linear and exact methods
Two simpler and two more exact methods are widely quoted. The linear method, often used for quick estimates, is Ra + 400 × (wins − losses) ÷ games. For the same event, three wins, three draws and a loss, it gives 2010 + 400 × 2 ÷ 7 = 2124. The exact method finds the single rating at which the sum of expected scores against each individual opponent equals the actual score; it is the true "no rating change" figure, here 2116. The three methods agree to within a few points for most results and diverge for extreme scores. A common misconception is that a performance rating is a prediction of your future strength; it measures one event, and a seven-game sample has a standard error of well over 100 points.
Why averages hide things
FIDE's method uses only the average opponent rating, so a field of one 2400 and one 1600 counts the same as two 2000s. That is a practical simplification; the exact method accounts for each opponent separately and gives a more faithful figure when the field is very uneven. Unplayed games and byes are left out of performance calculations, and in norm events the regulations allow the lowest-rated opponents to be raised to a rating floor.
How to use it
- Enter each opponent's rating and your result, round by round.
- Leave unplayed games and byes empty.
- Read the FIDE performance from the dp table, with the average opponent rating and percentage.
- Compare with the linear and exact methods, especially for very high or low scores.
Questions from the tournament hall
How do you calculate a performance rating in chess?
Under FIDE rules, take the average rating of your opponents and add dp, a value from FIDE's table for your percentage score. 4.5 out of 7 is 64%, giving dp = +102; against opponents averaging 2010, the performance is 2112.
What is the linear performance rating formula?
Average opponent rating plus 400 times wins minus losses, divided by games played. It is quick to work out by hand and close to the FIDE method for moderate scores, but less accurate for very high or very low percentages.
What is the performance rating for a perfect score?
In FIDE's method, dp is capped at +800 for 100%, so the performance is the average opponent rating plus 800. Mathematically, the exact zero-change performance for a perfect score is unbounded, which is why the table sets a cap.
Does a high performance rating mean my rating will go up a lot?
It means your rating will rise, but the amount depends on your K-factor and the number of games. A performance 100 points above your rating over 7 games at K = 20 typically adds around 20 to 25 points.
Disclaimer: Chess results apply FIDE's published rating, title and tiebreak rules to the figures you enter. They are not official ratings, norm certificates or tournament rulings. FIDE's published lists, the event's regulations and the chief arbiter decide; national federations and online servers use their own rating systems.