Chess.com Analysis

Sort:
cat_of_chess

I really have a propblem of with chess.com analysis. It makes many many many mistakes. This had let me to not upgreade my account yet. The analysis many times dose not makes sense. And appears to me to be very poor. It reconments moves that would obiously loose games. Also it should offer alternaite moves for every move out of book openings made insted of showing 2-3 when a mistake is made. There are inumerous mistakes in the liens that it recomments. I truly believe that it dose not play the level of a 2200 analysis. I have talked with my friend who is 2099 USCF rated, he seems a far more sound in his analysis of games. If you are going to offer analysis that poor I recoment that you call it 1400 rating analysis. 

ivandh

Post a game and let us see what it says

alaa78

if you have a friend with 2000+ rating and ready to analyse your games, then I'm sure you will benefit alot more than from a chess engine or computer analysis in chess.com because there will always be an interaction and explanation that can never be replaced by an engine...

Shivsky

Computer-based/engine analysis is dead on for tactics. I see  no way your 2099 friend can over-rule a comment on a tactical / forcing line.

For everything else (positions warranting strategic moves or thematic ideas),  an engine's evaluation score can sometimes be misleading, however reputable it is.

Besides, they just print out an eval. score and a principal variation to justify it....that doesn't really explain a position based on the merits of pieces, structure and what moves make more logical sense than the other.  A skilled human can really be much more helpful.

In short, I'd rely on your friend for those kinds of non-analytical positions and the engine for tactics alone.

cat_of_chess
Shivsky wrote:

Computer-based/engine analysis is dead on for tactics. I see  no way your 2099 friend can over-rule a comment on a tactical / forcing line.

For everything else (positions warranting strategic moves or thematic ideas),  an engine's evaluation score can sometimes be misleading, however reputable it is.

Besides, they just print out an eval. score and a principal variation to justify it....that doesn't really explain a position based on the merits of pieces, structure and what moves make more logical sense than the other.  A skilled human can really be much more helpful.

In short, I'd rely on your friend for those kinds of non-analytical positions and the engine for tactics alone.


I like chess.com alot. But the analysis makes the most basic of mistakes. How can I even trust lines it gives to me.

Shivsky

You've made repeated accusations of chess.com's analysis.  

I find it very hard to believe that any engine, even the "meek" and lowly crafty engines (pretty low on the totem pole of chess engines today) could ever be accused of making the most basic of mistakes.

It is time you actually posted the "faulty analysis" that you speak of and let people help you sort this issue out.  Unless you just wanted to vent out a dubious premise.

oinquarki

I am aware that there are occasional glitches in the analysis, but they only happen once every thousand moves or so, and then they're so obvious that you should be able to tell instantly. Other times it is because of the computer's calculation horizon. For example the engine looks ahead eight moves and stops, missing a tactic on move nine. There's nothing chess.com can really do about that because there's always going to have to be limit to the analysis because it takes a lot of precessing power to do.