lichess.org
Donate

Estimated of Level Play

Can a level of play be estimated by analysis of a single game? I was thinking it would be nice to see an estimated level of play along with the regular stockfish game analysis. I have noticed that the daily chess puzzle has a level rating. I just had a nice game and win against a player that seemed to be about my playing level. I would welcome a level of play opinion from Lichess members. :] lichess.org/BT3u9WLkQ5r1
Why do you want to know your estimated level if you do not play rated?
I'm really asking if stockfish could give a level of play... for each game as it stands...for every game... rated or not. :] Sorry... This turned out to be two posts in one... A stockfish question, and a rate my game request...My bad?! :]
Of course a single game is unsuitable to rate a level of play.
This is even an unfinished game: your opponent apparently decided to let his time run out in a better position for him.

The best one could do were to find a statistical correlation between player rating and AcPL and then apply that to approximately convert AcPL to rating.
This question crops up in forums from time to time but it ain’t work. One game is probably not enough to tell a cheater.
In this game the white anonymous reached 43 AcPL and the black anonymous 42 AcPL, so it is safe to say that both are about of equal strength.

I guess there must be a strong correlation between AcPL and rating but I feel no urge to crunch the numbers and prove it, so as to establih a conversion formula

rating = constant + factor * AcPL

with constant and factor to estimate from the data.
what if AcPl was an average of last 10 or so games
int numGames = 10;
avgAcpl[10] = sum()/numGames;
rating = constant + factor * avgAcPL;

just pseudo code.. forgive me
@tpr Unfortunately you are wrong. As a 1285 FIDE Rated (but got +19 from an open tournament and will be 1304, yay) player, I have games ranging from 12 AcPL to 60 AcPL. You can not convert AcPL to a rating.
#3 1. No 2. looks decent for a patzer like myself (game complexity was not high, relative easy to not blunder something)

#4 My opinion: you need at least three parameters: acpl, time control, complexity of game (good luck with the last one)

#8 True (or the result is just a shot in the blue)
In my OP I mentioned the daily chess puzzle. It is a single game assigned a rating.The puzzle often shows say some 2200 vs 2190 player and yet the puzzle is rated 1595 showing that the players level of play wasn't up to their ratings??. Does stockfish decide the puzzle ratings? Just curious. :] #5 I wasn't concerned about those cursed cheaters with this post. I really thought all players would like a level of play estimate for each game as it is played. :]

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