Elo Probability Calculator: How to Use It

This calculator works out your expected score (win probability) in a game or sport that uses the Elo rating system. You enter your rating, your opponent’s rating, and (optionally) a draw probability, and you get back the probability that you win, the probability of a draw, and the probability that you lose. The result is based on the standard Elo expected score formula used by FIDE and most chess platforms.

Formula

The Elo expected score for a player with rating R against an opponent with rating O is:

E = 1 / (1 + 10^((O – R) / 400))

Where:

• R = your current Elo rating (unitless, typical range 100–3000)
• O = opponent’s current Elo rating (same scale)
• E = expected score, a number between 0 and 1. If you win you get 1 point, a draw gives 0.5, a loss gives 0. So E is your expected average score per game.

To convert expected score to win probability when draws are possible, you need a draw probability D (between 0 and 1). Then:

• Win probability = (1 – D) × E
• Draw probability = D
• Loss probability = 1 – win – draw

If you set D = 0, the win probability equals E exactly (no draws).

Worked Example

Using the default values: your rating = 1500, opponent rating = 1500, draw probability = 0.3.

1. Calculate the rating difference: 1500 – 1500 = 0.
2. Compute 10^(0/400) = 10^0 = 1.
3. Expected score E = 1 / (1 + 1) = 0.5.
4. Win probability = (1 – 0.3) × 0.5 = 0.7 × 0.5 = 0.35, or 35%.
5. Draw probability = 0.3 = 30%.
6. Loss probability = 100% – 35% – 30% = 35%.

So with equal ratings and a 30% draw chance, you win 35%, draw 30%, lose 35%.

How the Elo System Works

Arpad Elo developed this system for chess in the 1960s. The core idea is that the difference in ratings predicts the expected score. A 400-point gap means the higher-rated player is about 10 times more likely to win (if no draws). For example:

• Difference 0: expected score 0.5 (50% win chance if no draws)
• Difference 100: expected score 0.64 (64%)
• Difference 200: expected score 0.76 (76%)
• Difference 400: expected score 0.91 (91%)

The actual rating update after a game uses a K-factor (e.g., 32 for new players, 24 for most, 16 for masters). New rating = old rating + K × (actual score – expected score). This calculator only gives the expected score, not the update, but you can use the expected score to compute the update manually.

How to Choose the Draw Probability

In chess, the draw probability depends on the rating difference and the players’ styles. A common approximation is that draws are more frequent between equal-strength players. For a quick estimate:

• If you don’t know, set D = 0.2 (20%) for chess between club players.
• For blitz or rapid games, draws are less common – try D = 0.1.
• For engine vs engine games, draws can be over 50% – set D = 0.5 or higher.
• For games without draws (like most video games), set D = 0.

If you want a pure win/loss probability, set D = 0. The calculator will then return win and loss probabilities that sum to 100%.

Common Mistakes and Misconceptions

One common mistake is thinking the expected score is the win probability. It is not – it’s the average score per game. In a game with draws, a 0.5 expected score could mean 50% win and 50% loss, or 25% win, 50% draw, 25% loss. Always account for draws if they are possible.

Another mistake is using the wrong base. The standard Elo formula uses base 10, not base e. Some systems (like Glicko) use a different scale. This calculator uses the classic base-10 formula.

Also, the Elo system assumes ratings are stable and that the probability of winning is constant within a game. It doesn’t account for factors like home advantage, fatigue, or specific matchups. For those, you would need a more complex model.

Rounding: the calculator rounds to two decimal places for percentages. That’s enough for practical use. Don’t expect exact probabilities – they are estimates.

Limitations of the Formula

The Elo formula is a logistic curve. It works well when ratings are accurate and players have played enough games. New players with provisional ratings may have large rating fluctuations, making the probability less reliable.

The draw probability is not part of the original Elo system – it’s an addition. The original system only gives expected score. If you use a fixed draw probability, you assume draws are equally likely at all rating differences, which is not true. In practice, draws are more common when ratings are close. For a more accurate model, you would need a separate draw formula (like the one used by FIDE for title norms).

Finally, the Elo system is not designed for games where the outcome is not binary (e.g., rankings in a race). It assumes a zero-sum game with two players.

Practical Uses

This calculator is useful for:

• Chess players preparing for a match – know your odds before you play.
• Tournament organizers estimating the distribution of results.
• Fantasy sports or betting (but be aware of the limitations).
• Understanding how rating systems work.

For example, if you are rated 1600 and your opponent is 1700, the expected score is 1/(1+10^(100/400)) = 0.36. With a 20% draw probability, your win chance is (1-0.2)*0.36 = 28.8%, draw 20%, lose 51.2%. You can use this to decide whether to play aggressively or defensively.

Frequently Asked Questions

What is a good Elo rating?

It depends on the player pool. In chess, a rating of 1200 is a casual club player, 1600 is a strong club player, 2000 is an expert, 2400 is a national master, and 2700+ is world-class. But these are arbitrary – the system is relative to the pool.

How do I convert expected score to win probability?

Multiply the expected score by (1 – draw probability). For example, if expected score is 0.6 and draw probability is 0.2, win probability = 0.6 * 0.8 = 0.48, draw = 0.2, loss = 0.32.

Does this work for other games like League of Legends?

Yes, if the game uses Elo or a variant. Many online games use a modified Elo system (like Glicko or TrueSkill). This calculator gives a rough estimate, but the exact probabilities may differ.

What is the K-factor?

The K-factor determines how much your rating changes after a game. It is not used in the probability calculation, but it affects how quickly your rating converges to your true skill. Typical values: 40 for new players, 20 for established, 10 for masters.

Why is a 400-point difference a 10x advantage?

Because 10^(400/400) = 10. In the formula, the denominator becomes 1+10 = 11, so the expected score is 1/11 ≈ 0.09 for the lower-rated player, and 10/11 ≈ 0.91 for the higher-rated player. That’s a 10:1 odds ratio.

Can I use this for team games?

Only if you treat the team as a single entity with one rating. If players have individual ratings, you need a different model (like Elo for teams, which averages ratings). This calculator works for two-player or two-team contests.

Related

• Axis Deviation Calculator: ECG Heart Axis Explained
• Avg Deviation Calculator: Formula & Examples
• Average Travel Time Calculator: How It Works

in r | Your rating | - | 1500 | 100
in o | Opponent rating | - | 1500 | 100
in d | Draw probability (0-1) | - | 0.3 | 0
out win | Win probability | % | (1 - d) / (1 + 10^((o - r) / 400)) * 100 | 2
out draw | Draw probability | % | d * 100 | 2
out lose | Loss probability | % | 100 - win - draw | 2