Interesting Chess Data: Time Controls And Game Results

Interesting Chess Data: Time Controls And Game Results

| 109 | Fun & Trivia

Every chess game starts with a "time control" - how long each player has to move. And every game ends with a "result" - why did it end? A friend asked me about how each time control impacts the result, and I had no idea! So I took a look. 

I started with a count of the most popular 9 time controls we have on over a period of 24 hours (for "Live" chess, not "Daily" games). Those are: 

Time Games % Games
10 min  436,437 28.5%
5 min  329,264 21.5%
3 min  280,513 18.3%
1 min  241,872 15.8%
2|1  56,929 3.7%
3|2  50,360 3.3%
15|10  49,400 3.2%
30 min  44,276 2.9%
5|5  41,746 2.7%
Total:  1,530,797 100.0%

First thing to notice is that people prefer playing with NO increment. Meaning, they would rather play games with a hard finish (you must make all of your moves in 10 minutes), rather than a game with bonus seconds for every move (like 2|1, where you get two minutes to play with 1 additional second per move). It's not even close. Personally I play increment because I don't like to lose on time, but I understand the draw of the simpler time controls. 

Now, before I show the numbers for results, let me describe what each result is:

Checkmate - This is self-explanatory, someone won with an actual checkmate. 

Draw - The game was drawn either by agreement, stalemate, 50 moves with no captures or pawn moves, or repetition (which is after 3 moves if claimed, or 5 moves automatically claimed).

Insufficient Material - This is a draw that is claimed AFTER the game is over on time, but the person who's clock did NOT run out doesn't actually have enough material to have won on checkmate. So, if you only have a king, but your opponent times out, the best you can get is a draw. 

Timeout - This means someone lost by timeout. 

Resign - One player clicked on "Resign" to give up. 

Abandon - One of the players disconnected, either from bad internet connection, or rage-quit. (Please be a good sport and resign instead!)

Here are the %s per time control: 

Time Resign Checkmate Draw Insuff. Mat. Timeout Abandon
10 min 49.7% 28.5% 3.0% 0.5% 12.0% 6.3%
5 min 44.7% 26.0% 2.5% 1.0% 23.0% 2.8%
3 min 37.8% 25.8% 2.2% 1.3% 31.4% 1.6%
1 min 13.4% 23.0% 0.8% 1.2% 61.0% 0.6%
2|1 35.7% 21.7% 2.1% 0.3% 38.7% 1.5%
3|2 48.0% 22.5% 3.1% 0.1% 24.5% 1.8%
15|10 60.3% 25.6% 3.9% 0.0% 4.4% 5.8%
30 min 55.7% 30.0% 4.0% 0.1% 3.2% 7.1%
5|5 55.8% 22.4% 3.4% 0.0% 14.3%


Fascinating! Now you can see the possibly outcomes from your games. 

  • If you play 10 minute games, most of your games will end with someone "winning" due to resignation or checkmate. There will be some draws, and then about 18% of games ending in timeout or abandonment. 
  • If you play 5 minute games, it's similar to 10 minute, but with almost TWICE the timouts!
  • If you play 3 minute games, few people resign, and now you are looking at almost 1/3 of your games ending by timeout. 
  • If you play 1 minute bullet, you're pretty much expecting that way more than half the time someone will lose by timeout. And maybe in a better position. 
  • If you play 2|1, your game might still be about 3 minutes long, and have similar numbers to 3 min. Slightly more timeouts actually (I'm surprised!)
  • If you play 3|2, once again, you aren't that different from 5 min games, and I didn't expect that. 
  • If you play 15|10, obviously you like LONG games with very few timeouts - but you will get a higher rate of "Abandoning" players. (Keep in mind - those players ARE punished by being put into a special playing pool for other people who abandon.)
  • If you play 30 min games, you also have a very high chance your game will end with a resignation or checkmate. Very few timeouts, but the most abandonment. 
  • If you play 5|5, it looks like you actually get some of the best overall numbers in terms of conclusive games and not too high of abandonment or timeout. 

Now, all this is interesting, but I wanted to know ONE more thing - when people timeout, were they winning, losing, or drawing? So I took 50 games from the 3 min time control and eyeballed the position at the end of each Timeout game. And here's what I found. If you lost by Timeout, your position would have been:

Losing          44%

Winning       38%

Draw/Even   18%

Let that sink in. In 3 min non-increment games, you would be WINNING or EVEN more than half of the time. Wow. 

I guess the next step would be to look at those same results for each time control and the associated timeout, but by the smell of curry in the air, it feels like dinner time. 

What are your takeaways?


[Please Note: Don't send me Tech Support, Abuse Reports, or other "Site" issues. Please contact Support for those things. But I do love thank you notes and personal messages! Thank you!]

Yes, I am the guy who started (along with Jay, Igor, Piotr, and many others). I have 3 amazing daughters, one wild son, and a wonderful, patient wife! I learned chess when I was 8 years old. Since then I have been playing, studying, and enjoying it regularly. I prefer semi-open and closed positions that blow up tactically (like the Closed Sicilian, King's Indian, Glek etc). 

Want to know more about how got started? Read about it here!

More from erik
How's 400-Person Virtual Team Works Together

How's 400-Person Virtual Team Works Together

'His Pawn Cheated And Killed My Pawn!'

'His Pawn Cheated And Killed My Pawn!'