Who were the players?
Qf8
@alkapwn3d
If a player has 8 ACL they are a cheater
If a player has 8 ACL they are a cheater
#1 Most of the time (unless both players make many blunders), to estimate average centipawn loss just divide 1000 by the number of moves, and round up a little bit.
@RomantheGreat
I have had games with sub 10 ACL. The game in my profile (lichess.org/1VIymAWj/black) I played Leela and I knew the position all the way to the end of the game. Just before this I had the same line but forgot the line and messed up near the end.
I had 4 ACL and didn't cheat. I just knew the position.
I have had games with sub 10 ACL. The game in my profile (lichess.org/1VIymAWj/black) I played Leela and I knew the position all the way to the end of the game. Just before this I had the same line but forgot the line and messed up near the end.
I had 4 ACL and didn't cheat. I just knew the position.
I find the metric to be fairly unhelpful.
When I get ahead against a lower rated opponent, for example, and then continue the game by denying counter chances, this tends to drive up CPL because it is "less accurate" to secure the win instead of finding the shortest route. <-- If you could measure CPL for the part of the game where the advantage was generated - and then turn it off when the advantage is being solidified and turned into a win, it might be useful for something.
The metric does not measure practical ability or practical chances in any way - and in order to achieve a lower metric it is often necessary to forgo practical wins for riskier play.
If anyone has a helpful way to use this metric, I'd love to hear about it.
When I get ahead against a lower rated opponent, for example, and then continue the game by denying counter chances, this tends to drive up CPL because it is "less accurate" to secure the win instead of finding the shortest route. <-- If you could measure CPL for the part of the game where the advantage was generated - and then turn it off when the advantage is being solidified and turned into a win, it might be useful for something.
The metric does not measure practical ability or practical chances in any way - and in order to achieve a lower metric it is often necessary to forgo practical wins for riskier play.
If anyone has a helpful way to use this metric, I'd love to hear about it.
@drunkintuition a 1900 doesn't play a bot and not make any inaccuracies, mistakes or blunders without consulting an engine, especially in blitz
I am sorry @drunkintuition I didn't realize your rating was provisional
@RomantheGreat llimit for inaccuracy in server analysis is so high that even avererage person can occasionally play 0/0/0 or 1/0/0 game. Current version for server analysis takes into account does the difference have meaning like if you are rook up and blunder a pawn it really does not matter and will not count as inaccuracy.
Old version limit for inaccuracy was 50 centipawns. Which means things like in a game reaching move about 10-15 I was clearly worse and basically losing and having 0/0/0 still at that point. Obviously in critcal position like that finding a good move is impossible. Also lenght of a game has huge impact - obviously - i think I have few "perfect" games from falkbeer counter gambit. everynow someone hits with F-pawn ans one way roud from there like in this game -- not min just something I looked for
and since the analysis on new version it is easy see that moves were not most accurate possible but as white shoudl have resigned long before he did. Had he resigned when game was ove CPS loss would have been less than 10. And d5 would have been the majos CPS loss
Old version limit for inaccuracy was 50 centipawns. Which means things like in a game reaching move about 10-15 I was clearly worse and basically losing and having 0/0/0 still at that point. Obviously in critcal position like that finding a good move is impossible. Also lenght of a game has huge impact - obviously - i think I have few "perfect" games from falkbeer counter gambit. everynow someone hits with F-pawn ans one way roud from there like in this game -- not min just something I looked for
and since the analysis on new version it is easy see that moves were not most accurate possible but as white shoudl have resigned long before he did. Had he resigned when game was ove CPS loss would have been less than 10. And d5 would have been the majos CPS loss
This topic has been archived and can no longer be replied to.