Quality of Game Reviews

I am confused by the initial game assessment that occurs immediately after the end of a game. I mean the minimal assessment that typically says “x” Best, “y” Excellent and “z” Miss.

If you click on a game in your game archive on the home page the app shows the final position, does the confetti animation and repeats the minimal assessment. My confusion began when I noticed the result was different to that shown at the original end of the game. Repeatedly revisiting the game in this way seems to give a different result every time.

He is an example of 6 minimal assessments of the same game.


Assessment Number Best Excellent Miss
1. 7 2 3
2 4 4 4
3 7 1 2
4 4 4 2
5 5 2 2
6 5 3 1


With every one coming out different it seems to me there is very little value in what is shown.

Even when I do a full game review the assessments are in my opinion dubious. It’s not uncommon for a move that saves a hanging queen to be assessed as an inaccuracy. I’m very dubious about subscribing to a premium account to get more full game reviews if the apparent quality is the same as the single review per day that I can currently see.

Would value any insight others can provide about this. Are the game reviews significantly better with a premium subscription?

I mean no disrespect towards this site, but all this "great" "brilliant" etc is just an ego feed, and does nothing to make you a better player. You want to improve?

How To Analyze Your Games.

Run your game through a chess engine.
Focus only on the most critical moments. Moves that change the evaluation from even to losing, from winning to losing, or from winning to even.
When a move changes the position drastically. Ask yourself: “Why did this move change the position so drastically?”
1. Does my move lose material? “Is my move safe?
2. Does my move weaken my pawn structure?
Focus on just a few critical moments of each game. What were the 3 most critical moments in the game? What do I need to do to not make the same mistakes?

Thanks, that’s sound advice

Thats sound great

131 Brilliant Moves.

Regarding the fact that the auto-analysis returns different results even for the same game:

The depth of the analysis will of course affect the results. Even one extra half-move depth can change a !? into a !!.

It's very likely that the auto-analysis feature adjusts its search depth to the network resources that are available at the time. So three runs of the analysis can all return different results, because the number of players online (and the number of client shells, number of concurrent analyses, etc) might change from minute to minute.

While I agree with you. My comment was directed at the OP and his references to the initial analysis right after the game. If the OP had mentioned anything about doing his own analysis then my advice would have been different. Im sure they are here, but the comments I see on analysis is always about the game review and not how to do self-analysis.

blueemu wrote:

Regarding the fact that the auto-analysis returns different results even for the same game:

The depth of the analysis will of course affect the results. Even one extra half-move depth can change a !? into a !!.

It's very likely that the auto-analysis feature adjusts its search depth to the network resources that are available at the time. So three runs of the analysis can all return different results, because the number of players online (and the number of client shells, number of concurrent analyses, etc) might change from minute to minute.

Thanks, that’s could be it, but the size of the difference just seconds apart is still surprising. What ever it is it means the results contain little of value.

As others have written doing your own analysis, with or without an engine, is important. At my level I think it’s helpful to identify the kinds of errors I am making and work towards eliminating them.