studying endgames

hamworld

Only Chess

17 Jun '13 14:31

hamworld
Joined
31 Oct '05
Moves
47
17 Jun '13 14:31
Let's say hypothetically you're studying a particular endgame(3 pieces including king each side and 7 pawns each side). Let's say you set it up with Rybka and play a real match against Rybka from that position, and end up drawing it in all 10 games.

Is it safe to say that you've mastered that endgame or is it possible that a human might try something you've never seen before(since humans play differently than computers) and just outplay you?
tvochess
Joined
08 Apr '09
Moves
19509
17 Jun '13 16:02
I think it depends on the complexity of the endgame. If the endgame is a theoretical draw, your computer will not find a usefull strategy and just play moves that keep its position level. As long as you don't blunder, you'll manage to hang on till a draw. However, in a complex endgame, a human opponent may steer the game into positions which are harder to solve for humans, hoping he can outplay you there untill you blunder.

In your case, each side has 11 moving pieces, so that's still quite a lot of options, and probably complex. So possibly a human may do better.
SwissGambit
Caninus Interruptus
2014.05.01
Joined
11 Apr '07
Moves
92274
17 Jun '13 16:06
Originally posted by hamworld
Let's say hypothetically you're studying a particular endgame(3 pieces including king each side and 7 pawns each side). Let's say you set it up with Rybka and play a real match against Rybka from that position, and end up drawing it in all 10 games.

Is it safe to say that you've mastered that endgame or is it possible that a human might try something you've never seen before(since humans play differently than computers) and just outplay you?
Well, you might be missing a win.
hamworld
Joined
31 Oct '05
Moves
47
17 Jun '13 16:48
Originally posted by SwissGambit
Well, you might be missing a win.
If you're going for a draw, would that be a bad thing?
SwissGambit
Caninus Interruptus
2014.05.01
Joined
11 Apr '07
Moves
92274
17 Jun '13 18:02
Originally posted by hamworld
If you're going for a draw, would that be a bad thing?
Going for a draw when you have a win is always a bad thing.
sundown316
The Mighty Messenger
The Wood of N'Kai
Joined
13 Dec '03
Moves
156184
17 Jun '13 18:52
I would say against most human players you would have very good chances to convert some of those positions to wins,because no human can play like a computer,and in my 40+ years of playing this here game,I can state that one constant has remained:most players below the rank of master can't play an endgame worth a damn.
greenpawn34
e4
Joined
06 May '08
Moves
42492
18 Jun '13 01:291 edit
..."most players below the rank of master can't play an endgame worth a damn."

And even Masters have not mastered the endgame....
...Apparently no matter how good you are you never master the endgame.

(by coincidence I have an incredible endgame blunder ready for the next blog
played by someone who was at the peak of his career and claimed by
some to be one of the best endgame players in the world.)

"....most players below the rank of master can't play an endgame worth a damn."

(part II)
I would not say that.
White found a move here that Capablanca, Rubinstein....or even Rybka
would never have played. I would put £100 down on the table and
say none of those would even consider it.

Two non-masters from last seasons White (Somerset) v Black (Middlesex) Match.

Mike Richardt (2026) - Black: Nevil Chan (2122)

White to play

Infact I'd go as far to say that if you showed this position before the game
to the player who played the actucal move he would even see it. It's that deep!

White played...

Nf5!!

OK White went on to lose, but you won't find moves like that in your endgame books.
hamworld
Joined
31 Oct '05
Moves
47
18 Jun '13 02:37
Originally posted by SwissGambit
Going for a draw when you have a win is always a bad thing.
The annoying thing was, I got into a winning position against the player even I barely knew what I was doing and messed it up and drew.

The only reason I got into that winning position was because my opponent barely had any time to think...he offered a draw and I accepted.

I seriously doubt the next time I get into that endgame I will end up lucky and in a winning position(even though the position is good for me) as it seems tricky to convert to a win in 4/0 chess(5/0 but we had 4 minutes by that endgame)
greenpawn34
e4
Joined
06 May '08
Moves
42492
18 Jun '13 09:442 edits
I would not kick yourself too much for mis-playing an ending in blitz.
If you screwed up a K & R v K failing to mate in 20 seconds then that
is serious but missing a subtle tempo waster, a triangulation or even
failing to find the correct plan within a few minutes knowing you might
not have the time to complete can be forgiven.

In my experience most just thrash our the moves in blitz to gain clock
time leaving the player winning no time to convert the win.

You don't study endings to be able to re-produce all the technique in blitz.
Is there a book out there called. 'Endgames for Blitz Players.'

(and if you have seen vid's of the good guys bashing out Rook endings
then look again. They are often loaded with errors.)

The good guys say when the middle game is over STOP! take 5 minutes
out to adjust your train of thought.
Suddenly you have this extra piece that has spent the entire game hiding.
Your King.
You do not have time for that in blitz.

This is about as close as I get to an ending in blitz and even then I managed
to get a mating attack.
Played on Game Knot a week or so ago. I'm White. (5/0)

Once again I sink a Knight in on d5 in my anti Najdorf line.

Invert Board PGN Comments

1. e4 c5 2. Nf3 d6 3. d4 cxd4 4. Qxd4 Bd7 5. Bc4 Nc6 6. Qd3 Nf6 7. O-O e6 8. Nc3 Be7 9. Bg5 Na5 {This always happens in this line but I like the open a-file. Many a tiime I've swung that a1 Rook over to the Kingside via a4.} 10. Bb3 Nxb3 11. axb3 Bc6 12. Rfe1 Qc7 13. Nd4 O-O 14. Nxc6 bxc6 {Ah the diea. Hit the Black a-pawn. Don't know what blitz ghost the lad saw. Possibly did not want to leave e7 Bishop unprotected with Qxc6, Perhaps angling for a healthy d5.} 15. Ra4 e5 16. Rea1 h6 17. Bxf6 Bxf6 18. Ne2 {This little chappe is looking at Nf5.} 18... g6 {Which is prevented.} 19. Qe3 {Sniffing h6 and a7 - which one do I get? I let my opponent decide.} 19... c5 {Wow! Now it's Nd5 time.} 20. Qxh6 Rfd8 21. Nc3 Bg7 22. Qe3 f5 23. Nd5 Qb7 24. f3 {It's important to defend the Knight with a pawn.} 24... f4 {This I liked, now everyhting happens on the Queensidde.} 25. Qc3 Rd7 26. Qc4 Kh7 27. Ra6 g5 28. h3 {No open files on the Kingside.} 28... Kg6 29. Qa4 Rdd8 30. Qc6 {I'll be needing my King on the Queenside so Queens off.} 30... Qxc6 31. Rxc6 Bf8 32. Rc7 a5 33. Ra4 Ra6 34. Kf2 Rb8 35. Rd7 Rc6 {He's going for counter-play. Cannot let my King get caught on the Kingside, there is a mating net there.} 36. Ke2 c4 37. Rxa5 cxb3 38. c3 {Now remember to watch out for Rxc3 and b2 ideas. The d5 Knight is stopping that at the moment.} 38... Rbc8 39. Kd3 {OK that's all the Rxc3 ideas stopped. Now the d5 Knight is free to create some tricks.} 39... Rb8 40. Re7 {If BxR then NxR + wins a piece.} 40... Kh6 41. Re6+ Kh7 {The wrong way. He had to go h5. Now it's King mating time and the Knight is the star performer.} 42. Ra7+ Kh8 43. Nf6 {Threat Rh7 mate.} 43... Bg7 44. Nh5 Bf8 45. Rg6 {My oppoent resigned here, well let his clock run out, there is not much he can do and I had a nice finish planned.} 4...Rcb6 { Played with idea of Rb7 blocking out a Rook. What else?} 46. Rg8+ {You have got to know you mating patterns. Boys. You must know your mating patterns.} 46... Kxg8 47. Nf6+ Kh8 48. Rh7
hamworld
Joined
31 Oct '05
Moves
47
19 Jun '13 07:02
Originally posted by tvochess
I think it depends on the complexity of the endgame. If the endgame is a theoretical draw, your computer will not find a usefull strategy and just play moves that keep its position level. As long as you don't blunder, you'll manage to hang on till a draw. However, in a complex endgame, a human opponent may steer the game into positions which are harder to s ...[text shortened]... that's still quite a lot of options, and probably complex. So possibly a human may do better.
What if you beat the computer in a 10-game match? If it was relatively simple like a K + B + P vs the enemy king, that might be easy to beat Rybka in, but what about this complex endgame? If I regularly beat the computer in that endgame, is it fair to say that I've mastered that particular endgame?
tvochess
Joined
08 Apr '09
Moves
19509
19 Jun '13 08:02
Originally posted by hamworld
What if you beat the computer in a 10-game match? If it was relatively simple like a K + B + P vs the enemy king, that might be easy to beat Rybka in, but what about this complex endgame? If I regularly beat the computer in that endgame, is it fair to say that I've mastered that particular endgame?
The same argument applies. When the engine knows it is lost, it will play moves to delay checkmate as long as possible. This may or may not be the thoughest reply. A human player could set-up some tricks in a lost position, which either get him mated earlier, or turn it into a win if you fall for it.

A K+B+P vs. K is safe, because there is no counterplay. This means you can still win, despite inaccuracies. However, if these remain in your play, I wouldn't say you mastered it.

By the way, I think humans can only master very simple endgames.
hamworld
Joined
31 Oct '05
Moves
47
19 Jun '13 08:49
Originally posted by tvochess
The same argument applies. When the engine knows it is lost, it will play moves to delay checkmate as long as possible. This may or may not be the thoughest reply. A human player could set-up some tricks in a lost position, which either get him mated earlier, or turn it into a win if you fall for it.

A K+B+P vs. K is safe, because there is no counterplay ...[text shortened]... uldn't say you mastered it.

By the way, I think humans can only master very simple endgames.

Just so I understand you, you're saying that if I regularly beat Rybka *here*, it does not mean I've mastered this endgame? I wouldn't call this position "lost", but I guess you're right in terms of a human maybe setting up some traps here. But I'm not sure how many traps he could set up though, in this relatively simple position.

Humans can only master very simple endgames? Well, that's depressing... Surely it is possible to increase one's strategic understanding of an endgame?

I would like to think that Capablanca has mastered some complex endgames.
tvochess
Joined
08 Apr '09
Moves
19509
19 Jun '13 09:50
Originally posted by hamworld
[fen]r3k3/pppppppp/8/8/8/8/PPPPPPPP/R3KB2 w Qq - 0 1[/fen]

Just so I understand you, you're saying that if I regularly beat Rybka *here*, it does not mean I've mastered this endgame? I wouldn't call this position "lost", but I guess you're right in terms of a human maybe setting up some traps here. But I'm not sure how many traps he could set up though, ...[text shortened]... n endgame?

I would like to think that Capablanca has mastered some complex endgames.
What I mean is that if you practice your endgames against a computer, you will only have gained experience with the "best" moves/strategies/plans by your opponent. However, there may be other ways your opponent could act, which introduce difficulties not encountered while playing the computer.

So, my point is that seeing only a fraction of the possible replies, you couldn't have mastered the whole endgame (in general). Nevertheless, if you consider a particular endgame simple, practicing with the computer may be sufficient.

I don't find it a depressing thought. Otherwise, chess would have been solved, making it work/study rather than a game.
hamworld
Joined
31 Oct '05
Moves
47
19 Jun '13 10:27
Originally posted by tvochess
What I mean is that if you practice your endgames against a computer, you will only have gained experience with the "best" moves/strategies/plans by your opponent. However, there may be other ways your opponent could act, which introduce difficulties not encountered while playing the computer.

So, my point is that seeing only a fraction of the possible r ...[text shortened]... g thought. Otherwise, chess would have been solved, making it work/study rather than a game.
So the only way to master that endgame is to analyze everything to death and learn how to win?
Paul Leggett
Chess Librarian
The Stacks
Joined
21 Aug '09
Moves
113572
19 Jun '13 13:39
Originally posted by greenpawn34
[b]
In my experience most just thrash our the moves in blitz to gain clock
time leaving the player winning no time to convert the win.

You don't study endings to be able to re-produce all the technique in blitz.
Is there a book out there called. 'Endgames for Blitz Players.'
This may sound a little strange, but Ocean64 and I (along with a 9 year old student) meet every Monday morning for a chess study session, and one of our techniques is "blitz endings"!

We don't do it for many endings, but one example is that we will play the bishop, knight, and king vs king ending at a blitz time control. At this point, all three of us can mate someone with the pieces distributed randomly on the board with 5 minutes on the clock.

All that said, we don't really do it for the ending. None of us have had the ending in a real game, and don't expect to.

Our real idea is to practice using the knight and bishop in tandem, as a form of tactical practice. It's like a martial arts kata, where we go through the forms and motions of the pieces to warm up our brains for the session.

We'll also sometimes practice queen vs rook, queen vs rook and pawn (which is a draw if you know what squares your pawn needs to be on), and some other positions just to keep them fresh.

We don't spend too much time on it (I just made an unintended joke...), but it's a good warmup to start the day just to get the juices flowing, and it's fun.

I will also add as an aside that I think this technique cured my blind spot for backward knight moves. That alone is worth the price of admission for me.

Only Chess Forum

studying endgames