lichess.org
Donate

How do grandmasters play with 0 inaccuracies & mistakes?

This is one of the most common and newbie questions but then again, i want to ask that in many of the games that i have seen Grandmasters have 0 inaccuracies and 0 mistakes, leave apart the blunders. i mean in a game you think that your move is correct but later computer analysis shows it is a mistake or an inaccurate move. So, how do GMs or experts knows that it is indeed an accurate move (according to the engines also)?
I just don't believe that statement to be true, if you check with an engine GMs games are filled with non-top computer choices on a frequent basis, however, they might consistently find the top 3 choices which is enough to win against 99% of all players.
The difference in the top choice and the second choice might be 0,15 on each move and maybe not even seen as a mistake but the accumulative numbers add up. I would say that games that last more than 20 moves where one party plays flawlessly are quite rare unless its a theory battle were all is known before the game.
Regards Richard
Lichess inaccuracy is about half-pawn so it means absolutely nothing for top player when they have time think about the moves. 0.5 below the best move is a mistake in most situations. Couple of those and you are in losing position. But grand master do make non-optimal moves. Just less than the others
@gablusky #1
This also depends on playing style.
Games from Bobby Fischer and Jose Capablanca, both crystal clear positional players with brilliant technique, will show you quite a few 0/0/0 games with Lichess analysis.
GMs who played more adventurous and risky like e.g. Mikhail Tal will likely show less 0/0/0 games with Lichess analysis.

As an example the 0/0/0 3 acpl game by Fischer :

Source with annotated game : http://www.chessgames.com/perl/chessgame?gid=1008419
@gablusky

They don't know. They try to give their best and have worked on their deficiencies over years.

It's like learning a language. If you practise each day you get better. The learning curve is losing steepness. That's all.

This topic has been archived and can no longer be replied to.