playing against engines

I have begun a different study approach for about two weeks now, before I was just trying to solve difficult tactics problems and barely ever playing, but then I noticed I have completely no idea on "how to play a chess game" from beggining to end, and now, besides doing tactics, I begun to play a 30 min game a day against Rybka (for those who do not know, it's a very powerful engine).

I had read somewhere that if you play against a lot stronger players all the time, it could be bad for you in comparison to playing against players who are just above your rating.

I do not understand the idea behind this, and cannot think of any good reason my play could be affected badly, do you have any ideas, experience or advice about this?

just a few hints:
*I do not mind losing all the time (and I know I will lose all the time), it just doesn't make an effect on my motivation.
*I do not mind having a defensive style developed, actually, it's just the way I want to learn to play: defensive, careful and safe, not taking too many risks.

Originally posted by diskamyl
I have begun a different study approach for about two weeks now, before I was just trying to solve difficult tactics problems and barely ever playing, but then I noticed I have completely no idea on "how to play a chess game" from beggining to end, and now, besides doing tactics, I begun to play a 30 min game a day against Rybka (for those who do not know, ...[text shortened]... ason my play could be affected badly, do you have any ideas, experience or advice about this?

I don't know about the generalization you gave, but something similar: playing against strong computer engines (as they are presently constituted) all the time might not be the best idea. The thing is, engines aren't generally all that flexible, so what happens is that you end up being "programmed" by the engine in order to survive (generally playing to draw), and you end up expecting a certain combination or series of moves, and basing your play around this combination or series, but when playing OTB or correspondence as at RHP your opponents will generally play in a completely different way -- much more variation in terms of moves and move orders. And if you are simply learning to respond by rote (which is what I mean by being "programmed" by the chess engine) you will not be prepared for this.

So, it may be much better, especially as a relatively low-rated player (a category I include myself in) to get a broader selection of opponents; by this means you will likely learn more about "how to play chess". Also, struggling just to draw all the time is kind of depressing, especially within the confines of a narrow set of moves/openings. And if you get burned out because you are constantly losing (and nothing but) and/or because you are concentrating on a very narrow and stylized set of moves, which lack the interest of variation and which may not have broader general application, then what is the point?

Originally posted by diskamyl
just a few hints:
*I do not mind losing all the time (and I know I will lose all the time), it just doesn't make an effect on my motivation.
*I do not mind having a defensive style developed, actually, it's just the way I want to learn to play: defensive, careful and safe, not taking too many risks.

In response to the information added in your edit:

Supplementing your regular chess play by playing against an engine like Rybka for 30 minutes a day isn't going to hurt. But again, if you *only* (or primarily) play against strong chess engines you are going to be left in the dark when it comes to playing other opponents, because their moves (for good or bad) will not be identical to the engine's.

Again, it is very easy to become conditioned by an expected series of moves (which is what happens if you play the same opponent -- an engine of the sort currently commercially available) over and over again, without playing other opponents. In order to know how to play good chess, you must learn how to "punish" bad moves by your opponent: the reason for this is that bad moves are only bad if you can prove they are bad over the board. If you let your opponent get away with them, you may find that even "bad" moves can be remarkably effective. What happens when you decide to play opponents other than Rybka and find that, due to differing strategy and tactics on their part, you suddenly find that the game is being taken in directions which make your carefully prepared responses irrelevant?

One possible problem: To improve it's important to be dedicated to finding the absolute best move every single time. If you're losing almost every time, especially against a computer where no one else is watching what you're doing, it can be hard to keep up that motivation to keep looking for the best move.

If that doesn't affect you, that's fine, but I think it would affect most people.

Another twist on the question would be, say you could play an unlimited number of games against Kasparov (or choose your favorite world-class player). Also stipulate that he won't be helping you analyze or do post-mortems, etc. Would it be better to keep playing Kasparov, or just play people slightly above you at a local chess club?

With respect to the second item, you can develop a "solid" playing style while playing ANY opponents. True, a strong opponent will put you on your guard, but again, *only* playing a strong chess engine may foster an OVERLY defensive playing style in which your goal is simply to survive and draw, not to win. Paradoxically, this may lead to poorer rather than better performance. The reason is that there is a difference between a style of play which keeps your position tight and avoids unnecessary risks while *simultaneously* pursuing your own plans, and (on the other hand) a style of play which becomes purely defensive. In chess you MUST regard offense as a *type* of defense because, if you allow yourself to be conditioned to reacting to your opponent's play in a purely defensive way, you will allow him to concentrate on his own plans without his having to deal with yours; this gives him the freedom to do what he wants while you react to it, and eventually, if you make one little mistake (as you surely will) you will be overrun. You have to keep a BALANCE between defensive considerations and your own plans.

Originally posted by diskamyl
I have begun a different study approach for about two weeks now, before I was just trying to solve difficult tactics problems and barely ever playing...

Well, I notice that you have no games currently in progress, and you yourself describe your previous study regimen as consisting primarily of abstract tactical studies and "barely ever playing". I don't think I can go amiss by suggesting that, if you want to improve your chess, you should play regularly. Don't play 20 games at once in a scatterbrained fashion, but play every day and devote considerable time to analysis of each position. As someone else here said, look for "the best" move. That requires knowing what your opponent's plans are and formulating your own plans (which must take the former into account); and that requires development of positional consciousness as well as tactical skills; and only by expending a lot of effort trying, will you succeed in developing these abilities. I'm still working on it and expect to be for years to come.

If you have no idea of how to go about formulating plans in chess then you might consider books dealing with the subject. Jeremy Silman's The Amateur's Mind is something I can recommend for someone of your (our) level.

When your choice is between OTB chess (in tournaments or clubs) and playing a computer, the latter can give you some leeway in terms of analysis time, and fit itself into your personal schedule better. But if you are playing correspondence chess anyway, the difference disappears.

Personally, in your position, I would: (a) play regularly; (b) play mostly opponents near your own rating level; (c) play some opponents considerably stronger than you, since this seems to be something you want and (in moderation) won't hurt (except your rating) and may even help; (d) spend time on analysis in every game in each move; (e) study your own games and see where you made mistakes or might have improved your performance.

(e) is problematic (for anyone below Master level) since your evaluations of your own games won't be any too strong: but anyone can make some degree of headway at it.

You might also consider: (f) Studying well-annotated Master games (where the annotations are aimed at readers at your general level of competence) inasmuch as these books can help explain the mysterious ideas behind the otherwise incomprehensible moves of advanced players. Of course, there is only so much that an ordinary student of the game can do at once.

Originally posted by Mark Adkins
I don't know about the generalization you gave, but something similar: playing against strong computer engines (as they are presently constituted) all the time might not be the best idea. The thing is, engines aren't generally all that flexible, so what happens is that you end up being "programmed" by the engine in order to survive (generally playing to of variation and which may not have broader general application, then what is the point?

Mark Adkins, thank you for your incredibly detailed and helpful posts, but I have some objections against some of your arguments.

I don't think that chess engines have an inflexible style which makes it possible to expect a certain combination or series of moves. in computer chess competition, it is very interesting that the number of duplicate games in maybe hundreds of thousand games are neglibly small. I was never able to play a certain, prepared line against any engine, and actually think it's not possible -due to the very advanced opening books they have-.

so, I do not agree that the style of high-level engines are narrow, actually I'm convinced on the idea that they can "hink" more flexibly (going deeper in positions which humans do not even consider as 'candidate moves' and making their way out of those complications).

so, I think, regarding style, it won't make a bad effect on me.

secondly, I do not play against the engine in any feeling of doing well at competition or score. I just do not have any chance of drawing against any of the high-level engines in my level, and therefore I have no intention to "play for a draw." As a matter of fact, I'm trying to play more tactically and get into more open positions to improve my tactical ability, which for sure makes me lose quicker, but that's not important for me as I have mentioned.

for your third argument about punishing mistakes, I'm trying to learn how to punish mistakes by observing (and experiencing as well) how the engine (actually, very skillfully!) punishes my mistakes. I think it makes even a deeper effect on me (in comparision to being the one to punish mistakes), since I'm on the losing side, while trying drastically to survive against those punishments.

for your argument about how to defend in chess, well, I hadn't thought of that, and I think you are certainly right. there's no sitting back, waiting behind your "city wall" and defend against the invastion in chess, you just have to do something, otherwise you simply get crushed slowly (or very quickly!). thanks for making me realize that, I think that's very important.

I'm going to try to vary my opposition from now. again, thanks a lot for your replies.

(for your last post, I should mention, when I said I had no idea on "how to play chess," that was maybe a little exagerration, I have read books that are similar to what you have suggested, and I actually can make logical plans in games, but I have realized that I can't make plans that "work" tactically. I just hang a piece, pawns, rooks, whatever, in the end of a variation I had calculated. So I'm back to more basic and intense tactical study now, instead of giving all my time to finding 5-8 move combinations. -which, I wasn't doing bad, by the way. which is interesting. probably they were actually 'over my level', even if I could manage to find them. sometimes studying simpler things and just a little experience in real play makes you better more quickly.)