@JoshuaBakker has made an interesting comment:

Why would you want [weaker competition]? For learning it's better to play against decent players....

But is it?

I believe that such advice used to be conventional during the precomputer era. If Capablanca had time to play chess against me, then I should have seized the opportunity, even if I must certainly have lost.

But is this not chiefly because Capablanca rarely had time to play chess against me?

Today, Stockfish always has time.

One could reasonably infer from the comment, taken in context, that an Elo 1300 player wasted his time by playing any Elo 1000 to 1600 opponents, if he could instead have been playing against nothing but the masters. Yet, is this really true?

Would one really learn chess best by fighting only hopeless odds? Does one truly gain nothing by the occasional victory, having taking advantage of an opponent's blunder? Must steely, grim perseverance in the face of certain destruction be the sole mode of chess play which has any lessons to teach?

(Here are a related question and answers from last year.)

  • 3
    The downvotes are curious. Care to comment? (Or maybe I should just display steel, grim perseverance in the face of certain downvoting!)
    – thb
    Commented Jun 5, 2016 at 12:48
  • Sorry to be dense. If you see what the problem with my question is, would you edit it? Thanks.
    – thb
    Commented Jun 5, 2016 at 12:58
  • 1
    If you only play against players who are much stronger than you, how are you going to learn to play winning positions? (Unless the strong player is giving lessons, e.g. giving odds or otherwise starting you from positions where you are winning.)
    – bof
    Commented Jun 6, 2016 at 7:31
  • 1
    If you only play against computers, you will get lots of practice playing losing positions, but you won't learn how to play such positions agains human beings. In a bad position against a human player, it's often a good idea to try to randomize the outcome by playing for complications or even outright swindles; of course that will get you nowhere against a computer.
    – bof
    Commented Jun 7, 2016 at 2:36
  • 1
    @bof: Actually there is a way; play against an AI that is slightly stronger than you, but undo whenever you feel like you made a mistake, with the primary goal that you win and the secondary goal that you make as few undos as possible. Then you will learn how to win as well as how to identify mistakes before you make them.
    – user21820
    Commented Dec 31, 2021 at 13:45

1 Answer 1


As this question highlights, it can become problematic if you don't even feel like you have a hope of winning a single game. Where's the fun in that? In my opinion, chess should be played first and foremost because it is fun. It isn't very fun never to feel the satisfaction of playing a nice winning combination against an opponent.

I think it's important for chess improvement to have fun while dealing with chess, and thus some success is needed. Otherwise it all becomes a chore, which will impede development in the long run. Therefore, I don't think it's advisable to play against stockfish for practice, unless it's training of some particular technique. Furthermore, stockfish training (i.e playing games against stockfish) will never really be applicable in regular games, as humans cannot, and will not even try to, play like the top computers.

However, facing strong opponents is a crucial component to improving. The ideal opponent is in my view somebody with a strength about 200-400 rating points above oneself. Then they will most likely beat you, and highlight the mistakes you make which makes them easier for you to discover and correct for future games. But you still have a realistic chance of scoring. You will have to struggle though, and really push yourself not to get blown off the board.

If you don't play stronger opponents, you will never improve very much. Why? Because there simply is no need for improvement then!

  • 1
    Well, it looks as though some readers did not care for my question, for some reason which is opaque to me. Maybe the title was too colorful. Maybe the question was too opinion-based (though chess.SE often seems a bit more opinion-based than do other SE sites). Sorry to drag your useful answer down with the question, at any rate. +1
    – thb
    Commented Jun 5, 2016 at 12:55
  • 1
    I think 200-400 is too much. 100-200 is more like it. When you get crushed, you don't actually learn much. It has to make a difference whether you are playing well or not. What do you learn when you box against Mike Tyson? Only that you are deficient in every way and no single thing would have made a difference. Commented Jun 6, 2016 at 12:02
  • 1
    The 200-400 interval is meant to take into account things such as daily form. A chess player's rating is only some sort of average measurement of strength. Usually they can play both much better and much worse than their rating. So an opponent 100 points above you can still play worse in one single game than you normally do, but this is far less likely to happen if you're facing someone 200 points above you. The 200 number as alower bound is to ensure that you actually have to play well to win against them, and not just take them on one of their bad days.
    – Scounged
    Commented Jun 6, 2016 at 12:25

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.