lichess.org
Donate

Can we get "Sum centipawn loss" in game analysis?

Preface: Yes, I know you can just add inaccuracies and multiply the average, but at a glance it would be nice, even as an optional setting.

Average centipawn loss is a cool metric to sees who's mistakes were usually larger, but it doesn't give an accurate at a glance picture of who played a more solid / accurate game. One big mistake can be lost among 10 .1 inaccuracies.

Sum centipawn loss would also be a nice self check metric to gauge progression. ie. "I used to make 5 pawns worth of mistakes a game, this month I'm giving up 4."

Note that I'm still only suggesting using the same data that the AVG los metric uses, not every minute change in score. Just your inaccuracies/mistakes.

Thoughts? Feelings? Reasons why I'm dumb? Thanks.

p.s. Lichess is awesome, and I love you guys for building it.
in average, the longer the game, the higher the sum centipawn loss will be, since every move loses at least 0 centipawns. so yeah, dumb :p

in the end average centipawn loss is just a cool metric, and gauging progression has to be done through ratings. average centipawn loss is highly dependant on the kind of position and the opponent
Thats why I meant only include your mistakes, not every score deviation on a move by move basis. The same way Average is done (I'm assuming). Yes, longer games give you more opportunity to make mistakes, though I don't consider that an excuse to be making mistakes, so its still a relevant metric.

To your point: Perhaps a threshold for being included? Though I'm sure there already is, or else "Average" would be near 0 after 60 moves and one 3pt mistake.
Average CP loss is for all moves, not just your mistakes.
average centipawn loss counts everything since opening move 1, and the threshold for something to be counted as inaccuracy/mistake/blunder are just certain fixed numbers.

anyway right, i see your point. i think average centipawn loss does all of that, but better. the thing is, if you made just one large blunder worth like 6 pawns, that's probably gameover regardless of how well played the game was before that.

This topic has been archived and can no longer be replied to.