How to evaluate my chess progress?

um take notes of your progress, like practice the theory, or your game style(what i mean is practice playing with bots without the backspace key ) then look at your accuracy, and take the percentage, and do this:

percentage accuracy x your actual rating, or the bots if it has

thats my manner, hope this helps!

@mettana said in #22:

percentage accuracy x your actual rating,
Thanks for the tip! Where can I find more info on using accuracy and rating together? Are there other ways to measure chess progress with math?

Actually, accuracy is a poor metric of game/play quality as it depends a lot on what the game looks like. Active and sharp games tend to have relatively low accuracy even among very strong players, on the other hand boring draws where nobody tried anything often have very high accuracy values. Thus if you take accuracy as an indicator of performance or progress, it might affect your style in a way that may not be desirable.

@Toscani said in #21:

This S-curve pattern is a natural progression that chess players must anticipate and plan for.
well using the S-curve to track chess progress is generally a good idea, but it doesn't account for individual differences between players. For example, think about a student who hates reading chess books, only uses computer analysis for their own games, and stops watching GM videos if they hear something they don’t like. These personal quirks can really affect how well the S-curve applies. Plus, the S-curve doesn’t tell you which specific areas need improvement.

@encerec said in #25:

For example, think about a student who hates reading chess books, only uses computer analysis for their own games, and stops watching GM videos if they hear something they don’t like.

Don't generalise from 98% of chess players to all! :-)

PS: books ordered yesterday: 9. pages read yesterday: 0.

@mkubecek said in #24:

Actually, accuracy is a poor metric of game/play quality as it depends a lot on what the game looks like. Active and sharp games tend to have relatively low accuracy even among very strong players, on the other hand boring draws where nobody tried anything often have very high accuracy values. Thus if you take accuracy as an indicator of performance or progress, it might affect your style in a way that may not be desirable.
yes i know but if you do it in the right way, it actually improves your chess skills. I know its not a long term evaulation, but it still works for me, like you can take it as your rating(level?) for one game.Thank you for the advice though, maybe i dont know something about chess

@mettana said in #27:

I know its not a long term evaulation, but it still works for me, like you can take it as your rating(level?) for one game.
That's exactly where I see the biggest problem. When I look at my games, I'm not utterly convinced that those with highest accuracy are those where I played best. Quite the opposite, they tend to be uneventful and rather boring. There are games which I feel much more proud about but the accuracy is much lower because in difficult and sharp positions with many tactical opportunities it's almost impossible not to deviate from "engine best".

Mean values over bigger samples might be relevant to compare one player in different time intervals but there is still the risk that this metric would encourage "playing it safe" and avoiding risks which doesn't sound like what you want for long term progress.

@mkubecek said in #24:

Actually, accuracy is a poor metric of game/play quality as it depends a lot on what the game looks like
It sounds interesting, but for testing purposes, I'd play several games in a row against different players, declining rematch requests each time (by the way, that's my style). Assuming the server matches you with opponents randomly, you could then average the accuracy of all those games. The question is, how many games do I need to play for that, since I'm not a mathematician?

In cases like this calculating the confidence interval is supposed to give you an idea. But the problem is that I have no idea what can be assumed about the distribution of accuracy values (I just don't think it's normal). You may need someone more experienced in probability and statistics to tell more, it's been a while since I had my course and it wasn't my primary field.