6

Wouldn't be surprise if this is a duplicate (related even more related), but:

Obviously there are two main ways to lose a game: Blundering it away in a single move or be slowly crushed by a better player without knowing what's happening.

We now have a giant database of games and engines that can tell us who is on top at any time in a game, and they could answer my question: which way is more frequent?

Very evidently, the answer might depend on ELO or ELO difference, so you can take that in account too.

Hauke Reddmann
  • 16,242
  • 3
  • 27
  • 70
  • 2
    There could be a bit more nuance here I think, for instance there could be a chaotic position where both players are making "blunders" all over the place and only in the end when the dust settles does it become clear what's going on. Also, I suspect the answer relies quite a bit on time control as well as rating difference. – Scounged Apr 18 '21 at 09:35
  • Problem is, how do you distinguish between the two? Like, in engine chess, it is arguably possible to define a "blunder" as "whenever the eval increases tremendously in one move". But if you have a 32-piece tablebase, then you'd have 0.00 eval the whole way before suddenly increasing to a mate score. That would classify every lost game as lost due to a blunder. I don't see how to define this objectively unfortunately. – Allure Apr 18 '21 at 11:08
  • @Scounged: This is a very interesting spinoff question I already pondered to ask: Maximum of blunders (defined as changing the objective outcome) in one game. – Hauke Reddmann Apr 18 '21 at 15:30
  • @Allure: I see the problem and already tested independently the following: Kh2 Ph3 Pb5 - Kh4 Pc7 Pb7, Black to move. He can lose on the spot with 1...b6??, ruining his majority. Lichess/Stockfish, with tablebases, rated it as +7something. (The point is it didn't just jump to +99.) The one question I linked arbitrarily used +3. Unfortunately I either don't see an alternative to treating this pragmatically. – Hauke Reddmann Apr 18 '21 at 15:36
  • It's hard for a computer to tell when you've been "crushed without knowing what happened". Objectively speaking, every game that is lost had at least one critical mistake. – David Apr 18 '21 at 17:47
  • I think there is a lot of grey area; a lot of games are somewhere in between: the superior side is pressuring and the inferior side cracks under pressure. How long that pressure was on in terms of number of moves as well, objective and practical strength vary substantially. – Akavall Apr 18 '21 at 18:39

2 Answers2

1

At the GM level, players mostly resign after the second "slow crushing" method. In amateur chess usually a blunder of a queen or checkmate or major piece wins. Before 1000 it is the first around 99% of the time (talking chess.com elo) because they cannot calculate well enough to find tactics.

Even at the IM and GM level players will still blunder, but they are usually positional errors and not major piece blunders (of course, they do occur such as Giri vs Carlsen in the Chessable Masters).

Most victories in classical and longer time formats at the GM level are resignations, but in bullet it is usually timeout or a blunder.

blunderchess1236
  • 95
  • 4
-2

Not sure if the question is accurately stated in English as was intended.

You lose either by getting mated, losing on time when clock runs out, or you give up for some reason. Or the director forfeits the game for other reasons like your cell phone ringing, or you are caught using the internet for help.

Within the reason to give up it is usually for losing material, or by being outplayed so you eventually are behind material when they promote a pawn and are able to mate soon.

Are you asking for the ways you can get outplayed or something else?

expert
  • 140
  • 1
  • 3