Accuracy

This is a bit of an unusual topic and outlook on the game because most  of the time we just accept our Accuracy score calculation, but I would like to know more about that calculation system, partly because I have some wildly fluctuating Accuracy numbers. In some games that I thought I had Aced, I didn't get much respect in terms of Accuracy.

Let me show you what I mean. Also in this blog is an article on "The Petrov" from a few months ago.

I had never played The Petrov before and I Aced this game in 10 moves. I was over the Moon, until I pushed that little button on the lower left of the board, and selected Review (I review almost all my games) and got shown my stats. My Accuracy was shot down in flames to 57 for a game that I won in 10 moves. What the Hell? Does a game from a relative newby like myself get any better? A crap Accuracy rating of 57 for a game that good? Really?

For comparison let's look at a 17 move game that I won last night, by Gentlemanly Resignation. I played Black you may like to flip the board.

A win in 17 moves, and (Holy Crap!) at 96 rating.

So, what gives? It can't just have been the influence of my new coach, although he has really made a difference to my game in just 4 months.

Well, I came across this;

"Chess accuracy, as calculated by platforms like Lichess and Chess.com, measures how closely your moves align with the best moves recommended by a powerful chess engine like Stockfish. It's a percentage that reflects the degree to which you reduced your winning chances by making a move compared to the optimal move for that position. A higher percentage indicates better moves, with 90%+ often considered "expert strength" in longer, non-forced games."

This surprised me and the highlight is on purpose. Our Accuracy measurement has utterly nothing to do with our game performance against an opponent. We're literally playing against, and being compared to what/how a computer would perform, and what pieces the computer would play, and then where the computer would put that piece.

This raised an unusual thought in my head. We're not necessarily playing against an opponent. We've been programmed to play against "what a computer" would do and deviation from that standard is measured and used to determine a few things including possibly cheating, and (ask Hans Niemann) social status. So, poor Accuracy, or excessively good rating is not allowed.

At the end of the tournament it will still be the winner who goes home with the prize, but that winner may simply be the player who has best learned what our computers expect the players to play, and the best of the players who do that will wind up the winner (with the obvious exception of Magnus Carlsen, who does what he wants no matter what).

So how is your Accuracy? Wildly fluctuating, like mine, or "steady as she goes"?