That number sits oddly against what chess looks like from the outside. In the 2018 World Championship, Carlsen and Caruana drew all twelve classical games and settled the title on rapid tiebreaks. If your picture of chess is the one from the broadcast, you're picturing a game that barely exists at the level you play it. Let's dig in!

Why Do So Few Amateur Games Finish Level?

Because a draw is the hardest result to produce. It asks both players to go the whole game without a losing mistake, and that is a much taller order than winning is.

Think about what has to happen. One side gets a small edge, then gives it back. Nobody hangs a piece. Nobody misses a tactic in a position with fifteen pieces still on. Nobody loses on time in a scramble. Every one of those has to hold for forty moves at once, and at 1400 the odds of that are dreadful.

Winning needs one of those things to go wrong exactly once, which is why it's so much more common. A decisive game only needs one mistake, so it happens 96.1% of the time, while a draw needs an entire game's worth of accuracy from two people at the same moment. Grandmasters draw constantly for the same reason in reverse: at that level, the mistake never arrives.

Which Openings Draw The Most?

One opening stands clear of the field, and it's the one you'd guess. The Berlin Defense drew 337 of its 3,051 games, which is 11.0%, or one in nine. Nothing else we measured comes close.

The Ruy Lopez as a whole draws 5.8% across 13,922 games, roughly one in 17, and the Berlin drags that figure up by itself. At the other end sit the gambits. The King's Gambit Accepted drew 2.9% of 3,975 games, about one in 34, and the Falkbeer Countergambit managed 2.7% of 1,097, one in 37.

The big mainlines cluster in the middle and barely move. 1.d4 drew 3.9% of 207,413 games and the Sicilian Defense drew 3.8% of 75,974, which is one in 26 either way. Openings with completely opposite reputations, the solid queen's pawn and the sharpest defence in chess, produce the same draw rate to within a tenth of a percent.

Is The Berlin Really A Drawing Weapon?

Yes, and this is the rare case where the folklore survives contact with the data. The Berlin earned its name in 2000, when Kramnik used it to blunt Kasparov's 1.e4 completely, and it has been shorthand for a drawing weapon ever since.

Four times the draw rate of the King's Gambit is a real effect, not noise. It comes from the structure: the main line trades queens early, and a queenless middlegame removes most of the ways an amateur game explodes.

Now the honest part. We can't separate the opening from the people who choose it. The Berlin is a theory-heavy line that attracts stronger and better prepared players, and stronger players draw more whatever they open with. Our sample carries no rating filter, so some of that 11.0% belongs to the players rather than the pawns. The direction of the effect is solid, but treat its exact size as an upper bound rather than a measurement of the opening alone.

Should You Choose An Opening To Avoid Draws?

No. Don't bother, and if you have ever passed over a line because you heard it was drawish, you gave up something real to avoid something that was never going to happen.

Look at what the choice actually buys. Switching from the most drawish opening we measured to the least takes you from one draw in nine to one in 37.

In 100 games that's the difference between 11 draws and 3. You've moved eight results, and you're still winning or losing 89 games out of 100 either way.

Now compare that to what the same decision costs. An opening you picked for its draw rate is an opening you don't understand, and every game you play in it starts from a position you can't navigate. The eight games you claw back from the draw column are worth far less than the games you throw away playing a structure you don't know. Pick the opening you'll enjoy studying instead.

How Did We Count This?

Every rated standard game in the Lichess database for July 2026, with no rating filter, so the figure is what happens across the whole player base rather than one slice of it.

For each opening we look at its defining position, the point where the line becomes itself, and tally how the games that reached it finished. That's the same data behind the win rates on all 137 of our chess openings pages.

The six openings behind the headline figure are 1.d4, 1.c4, and the Sicilian, French, Caro-Kann and Alekhine. They're there because they can't overlap: each one is fixed by a different first or second move, so no game is counted twice. That matters more than it sounds, because a Najdorf game is also a Sicilian game, and adding those two totals together would count 2,137 games twice and quietly inflate everything built on top.

So What Should You Take Away?

The draw is a master-level problem that you have almost certainly never had. At 3.9%, it's a rounding error in your results, and the opening you pick moves it by a handful of games in a hundred.

Check your own last fifty games. Two draws is par. If you have more than three, you are already the exception, and it says more about how you handle level positions than about anything you played on move one!

Spend the attention somewhere it pays. Your games are being decided by the one mistake that arrives in every single one of them, so play the openings you actually want to play and go and find that mistake instead. Our openings section has the real numbers behind every line, and the free opening trainer will drill any of them until the position stops surprising you.

Sources

• Lichess Open Database, standard rated games, July 2026. Released CC0.
• World Chess Championship 2018, Carlsen against Caruana, twelve drawn classical games.