Elo Rating Calculator

First calculate the expected score from the rating difference, then subtract that expectation from the actual score and multiply by K. A win uses actual score 1, a draw uses 0.5, and a loss uses 0. The formula is rating change = effective K x (actual score - expected score).

A higher K factor makes the same result move ratings more. Newer or faster-changing pools often use higher K values, while more established ratings usually use lower ones for stability. In FIDE-style mode, the selected K can be reduced when K multiplied by games in the rating period would exceed the rating-period limit.

Because the underdog was expected to score less. When the actual score beats the expectation by a larger margin, the Elo formula produces a larger positive rating change. A favourite can still gain points for winning, but the gain is usually smaller because the result was already likely.

Yes. A draw is scored as 0.5, so it is only neutral when your expected score is also about 0.5. If you are rated much higher than your opponent, your expected score is above 0.5, and drawing counts as an underperformance. If you are the underdog, the same draw can gain points.

A 400-point rule limits the rating difference used in the expected-score step. In the FIDE 2650+ mode used by this calculator, the cap applies to players rated below 2650, while players rated 2650 or above use the full rating gap. The legacy mode caps every player, and classic Elo mode uses no cap.

Official systems may batch several games together, round only at the rating-period level, or apply player-specific K-factor rules and provisional-rating adjustments that a generic one-game estimator does not reproduce fully. The calculator is designed to show the underlying mechanics and likely direction, not to replace the official publication record.

A simple two-player Elo model with the same K factor and symmetric expectations is zero-sum: one player gains exactly what the other loses. Official systems can break that clean symmetry when players have different K factors, when rating-period caps apply, or when special rules affect one player differently from the other.

Yes, it is useful for chess rating change estimates because it includes win, draw, and loss scoring, K-factor controls, FIDE-style rating-gap options, and a rating-period game count. For official FIDE publication checks, compare the result with the federation’s own calculation details because final rounding and player-specific rules can still differ.

It can explain the broad idea of expected score and rating movement, but Chess.com and Lichess do not publish ratings with a fixed-K classical Elo formula. They use Glicko-style systems that include rating uncertainty, so online rating changes can differ substantially from a standard Elo estimate.

Use the K factor that matches the system you are trying to approximate. Many chess explanations use K=40 for newer or junior players, K=20 for established players below master-level thresholds, and K=10 for highly established top players. For a private game ladder, choose a higher K for faster movement and a lower K for more stable ratings.

The scenario table answers the most common planning question before a match: how much could my rating change for each possible result? Seeing all three outcomes at once makes upset value, draw risk, and favourite penalty easier to compare than recalculating the page three separate times.

Elo is a specific rating model with a defined expected-score formula and K-factor update. MMR usually means matchmaking rating, a broader term that can use Elo, Glicko, TrueSkill-style systems, proprietary platform models, or hidden adjustments. Use classic Elo mode for transparent experiments, but do not assume every MMR system updates exactly this way.