Comments on https://lichess.org/@/jk_182/blog/how-does-the-clock-impact-the-rate-of-mistakes/JSazQplM
Nice blog. By the way, have you thought about any correlation between knight and bishop in short time controls. There is a certain school of thoughts that contend that sole knight against sole bishop in short time control is of a substantial advantage.
It would have been cool if this could have been confirmed in a statistical way.
"if certain players better cope with time trouble than others"
* It depends on the position. A simple, technically won position is easier to play than a sharp, complicated position.
@Nordlandia said in #2:
> Nice blog. By the way, have you thought about any correlation between knight and bishop in short time controls. There is a certain school of thoughts that contend that sole knight against sole bishop in short time control is of a substantial advantage.
> It would have been cool if this could have been confirmed in a statistical way.
I actually thought about something like this recently. My idea was to see how many mistakes in blitz games get punished by lines where a knight move is key. The idea was that this is an indication that players might have missed the knight move. I didn't find a good way to find when a line relied on a knight move, especially when the weaker side has multiple options. But I'll keep thinking about it.
@tpr said in #3:
> "if certain players better cope with time trouble than others"
> * It depends on the position. A simple, technically won position is easier to play than a sharp, complicated position.
That's of course true, but one can compare it to a baseline of mistakes a player makes when not in time trouble. This wouldn't be perfect as there are many complicating factors, but I think it's a good way to at least get an idea about how well different players handle time trouble.
I wonder what a 3D graph with the third axis "game phase" (based upon remaining material, as Stockfish and Lichess define it although I'm struggling to recall the definition) would look like.
#5
"mistakes a player makes when not in time trouble"
* What is time trouble? It is a psychological notion of anxiety. Bullet players regularly play at a pace of 1 second per move. Some classical time control players get anxious when they have to complete 10 moves in 5 minutes. The moves do not happen in a void: they result from all previous thinking about the position.
These results are reassuring, as expected.
As you would expect to find that after some point the probability of a mistakes becomes less and less relevant to player ranking.
However, a mistake is not a mistake if its not capitalized by the opponent. If it is not, the mistake becomes anything from rectifiable to a winning opportunity by the person that actually made the mistake. If your time has run out and you are on a one second increment and you move your queen right up against the opponents king, without support, you will either lose your queen and likely the game, or possibly corner the king and win the game. Its not worth the risk of course, unless you calculate it to be checkmate on the next move, but it just goes to show you, a mistake is not realized until the other player spots it.
Anyway, this is the reason that events like Titled Tuesday work, as you would really not hope to be able to win against the likes of Magnus if the format was rapid or classic.
As it would be awfully boring if not moot to pit an engine to play itself while restricting the time it has for calculation, as it would make the same type of mistakes, although it is likely that an engine would be a lot better at spotting a mistake, than avoiding one. But then again, maybe not.
It's pretty funny seeing a basically linear graph showing long think, wrong think.
Like you said it doesn't mean that much, because players are going to think a long time in a position that they don't feel comfortable in, but on the other hand - such sayings tend to become a thing for a reason!