No mistake, no blunder

fischermx

Over the time of playing here and used to the post-game computer analysis, I've got somehow the idea that a game where I don't get any mistake, or blunders marked, it's a good game overall. Not that is perfect, not that's brilliant, just, a good game, something to be a little proud of.

 

Is it real a good parameter to consider? Can we consider these stockfish evaluation as, somehow, solid?

ArtNJ

Many people have done that, including some that are low rated (by virtue of the type of position and luck).  It would likely be very hard for someone below expert to do it on purpose, but that doesn't mean it is much of an achievement or any sort of marker of growing skill - it isn't.  Many times odd decisions that a better human wouldn't make don't trigger the threshold for an inaccuracy because they don't happen to matter that much due to factors that only the computer can see -- this is quite clear when low rates post "flawless" games.  

fischermx

Thanks for your answer. I understand your point.

If not, then, how can we properly use the post-game computer analysis to measure game quality?

poodle_noodle
fischermx wrote:

 

Is it real a good parameter to consider? Can we consider these stockfish evaluation as, somehow, solid?

Eh, sort of.

If some auto analysis says your game is perfect, usually it means the opponent didn't put up a fight (or the position was very dull). Still, it's something you can be happy about.

Is it a good parameter? Not really. Engines get things wrong too. You have to work with them and give them plenty of time, so a website's auto analysis, even if it's an engine like stockfish, is mostly useful for pointing out your really big blunders. Also engines don't understand practical moves. Sometimes the eval is mate in 10, but it's nearly impossible to play it during a real game. Sometimes the eval is 0.00 but it's almost impossible to draw (lets say like mutual attack after opposite side castling).

How can we use the post analysis to measure quality?

Probably for just one game you can't. If you were comparing 100 games to another 100 then some kind of average might be telling though.

wizardinyourbed4days

I can be your first mistake wink.png but i promise i won't be a blunder <3

 

fischermx

Thank you DeirdeSkye and Poodle_Noodle for your comments, they are very helpfull.

 

goldenwriter

me too 

 

tlay80
fischermx wrote:

Thanks for your answer. I understand your point.

If not, then, how can we properly use the post-game computer analysis to measure game quality?

You can't. That's not what engines are for.  They're tools to help work through the possibilities of a position. Once you've worked through a game with an engine, then, if you really want to evaluate how well you played, the tool you need to go to is your human brain.

tlay80

Oops.  Just replied to a zombie thread.  Mea culpa.