Chess Rating Calculator (Elo)
Win chance and rating points at stake from two Elo ratings.
What it does
Arpad Elo modeled each game as evidence about strength: the rating gap converts to a win probability, and results move ratings toward that expectation. A 200 point favorite expects about 76%, so its wins pay little while its losses cost plenty.
The tool above takes both ratings with the K factor. It returns each expected score and the points player A banks on a win.
K factors tune volatility by level. New and junior players develop fast at K 40, club regulars sit at 32, and masters stabilize at 16 to 24, which is why the same upset pays differently across events.
Draws split the difference at half a point each. The same formula prices them: score half against expectation moves half the win exchange.
How to use
- Enter both current ratings.
- Pick the event K factor.
- Read expected scores as win chances.
- Check the win gain before risking sharp lines.
Formula
- gap / 400
- Rating difference scaled. 400 points means ten to one odds.
- K
- Development coefficient setting exchange size.
- score
- 1 win, 0.5 draw, 0 loss.
FIDE and most federations use logistic variants of this core with rating floors and bonus rules around it. Club math follows the plain form exactly.
Worked example
Example: 1500 vs 1700 at K 32
Player A 1500, player B 1700, club K factor.
- Gap: (1700 - 1500) / 400 = 0.5.
- Expected A: 1 / (1 + 10^0.5) = 1 / 4.1623 = 24.0%.
- Expected B: 100 - 24.0 = 76.0%.
- A win gains: 32 x (1 - 0.2403) = +24.3 points.
Reference table
| Gap | Favorite | Underdog | Upset pays at K32 |
|---|---|---|---|
| 0 | 50.0% | 50.0% | +16.0 |
| 100 | 64.0% | 36.0% | +20.5 |
| 200 | 76.0% | 24.0% | +24.3 |
| 400 | 90.9% | 9.1% | +29.1 |
Limits
- Plain Elo only. Bonus and floor rules excluded.
- Rapid and blitz pools rate separately.
- Provisional ratings swing wider than shown.
- Color and form sit outside the math.
Common mistakes
- Reading expected score as guaranteed points.
- Using club K for FIDE rated events.
- Fearing higher rated players past the real odds.
- Ducking games to protect ratings.
- Forgetting draws move points too.
Try it yourself
Quick checks. Answers sit tucked under each question.
1800 vs 1800 at K 24. Win gains?
Expected 50%. Gain 24 x 0.5 = +12.0 points.
1200 beats 1600 at K 32?
Gap 400: expected 9.1%. Gain 32 x 0.909 = +29.1 points.
Questions
What is a good chess rating?
Club players run 1200 to 1800, experts past 2000, grandmasters past 2500.
How does K factor work?
It scales exchanges. High K develops juniors fast; low K steadies masters.
What happens on a draw?
Half score each. The favorite drops a little, the underdog gains a little.
Why do upsets pay more?
Points price surprise: beating 76% expectation earns the unexpected 76% share times K.
Do colors matter?
White scores about 55% in practice. Ratings ignore it; preparation should not.
Online vs over the board?
Separate pools with different deflation. Compare within one system.