Computer Analysis

Sort:
Avatar of Michahellis

I did a computer analysis of one of my games and it game the following information.

Computer Analysis (~2500 strength)

Inaccuracies: 2 = 5.4% of moves

Mistakes: 17 = 45.9% of moves

Blunders: 2 = 5.4% of moves

 

I assume this means that 5.4% of my moves were not necessarily wrong, but were not moves from common lines, or not the worst move? Then almost half my moves were outright wrong and the remainling 5.4% were critically wrong. Then it says ~ 2500 strength.I don't really understand that part.

Avatar of zfonzy

The ~2500 strength is the strength of the computer analysing your game, the higher the strength the more in-depth it looks at positions before making a recomendation.

I think that's the answer to what you're asking?

Avatar of Michahellis

That was it, thanks! So, obviously I made a lot of mistakes, but how bad did I play,or how can I tell how bad I played. What would be a usual % of mistakes etc.I did actually end up winning in the end, despite a couple of big mistakes.

Avatar of baddogno

The 2500 strength is the strength of the engine used for analysis.  Used to be premium members would get a little stronger engine, but I think everyone gets the same now.  We also used to have an engine with an opening book but that seems to have dissapeared, so if the very beginning of the analysis seems a bit weird, that's why.  Generally the analysis is good for pointing out tactical errors but if you want a serious engine to take a look you have to do it offsite. Oh look, I'm wrong again! Laughing

  • Basic/Free: 1 Computer Analysis per week - with a search depth of 8 ply (estimated at 2000 strength)
  • Gold:  2 Computer Analysis per week - with a search depth of 10 ply (estimated at 2200 strength)
  • Platinum:  25 Computer Analysis per week - with a search depth of 14 ply (estimated at 2500 strength)
  • Diamond: 100 Computer Analysis per week - with a search depth of 14 ply (estimated at 2500 strength)

Ke

Avatar of Irontiger

The problem with chess.com analysis, at least at free levels, is that mistakes are called even in trivially winning positions. If the computer sees a mate in 26 and what you play gets you "only" one more queen or a mate in 27, he will call it a mistake.

So the repartition of mistakes is more important than the number of mistakes. Look at them and the suggested lines and you should be able to tell the difference between a tactical miss, a positional overlook, and complete computer crap.

 

I suppose other players could give you their stats to compare, but that's pretty much useless...