Wednesday, October 01, 2008

This is the end

It must be nearly a quarter of a century since I came across a report of an exchange between Karpov and some amateur chess players that went something like this…

Amateurs: What should we do to improve?

Karpov: I don’t know what you do now.

As: Play a lot … look at openings.

K: But not endings? Do the opposite. Study endings.

Karpov's words have been lodged in my head ever since, though needless to say I've never actually paid them any heed. Sadly the level of my endgame knowledge is not much higer now than it was back in the mid eighties. I suspect, as far as club players go at least, that I’m not alone in the lack of attention I give to this area of the game.

Where to start in turning things around? In principle there may be no end to what we could learn about endgames but with limited time and energy available how best to focus our attention?

I suppose it’s worth knowing why and how White wins this

and Black draws this

but is there really any need other than as learning for its own sake to know, for example, how to deliver mate from here?

(19 moves at most with best play apparently)

So I'm left with two questions for you dear readers of the S&BCC blog:-

(a) How would you evaluate your own endgame knowledge?
(b) What is most productive way to study the endgame?


ejh said...

Bishop and knight - yes, you need to know.

I've been on the wrong side of it and my opponent performed it successfully: I've seen a friend survive it because their opponent didn't know what to do (this was in the 4NCL).

I'm also aware of it occurring in an Olympiad and the mating method not being known to the player with the pieces - they managed it on something like the 48th move.

Learn it.

Other than that - endings are played incredibly badly by club players (including me) because we don't practice them. We also don't take any notice of basic things that we've known since we were kids - for instance, lower-ranked players keep losing drawn ending against me because they insist on putting the pawns on the same colour as their bishop. Not unusually, they then offer a draw, just at the point when in fact, they've made it really easy for me to win....

How to learn better ending play? I don't know, precisely. But it'll do you a power of good if you try.

Mike G said...

There are two methods of mating using knight and bishop from the position you show. The first is the "W's" method usually taught in the UK where the king is driven along the side of the board towards h8 and the knight moves c7, d5, e7 etc. The 2nd is the "Russian method" where the white pleces form a cage around the black king (initially the WB is put on g3 and the WN is put on e3 and the WK completes the cage on files with the BK on the right and in the cage. Progressively the cage is shrunk forcing the BK nearer h8. The 2nd method requires is in the book "a pocket guide to endgames" by David Hooper (the first book I bought on endings and now sadly out of print).

(Not answering the questions, just showing off my endgame knowledge.)

(Not answering the

ejh said...

I may have mentioned on here before that I was actually taught B&N mates by Nigel Povah, at a training school for Hertfordshire juniors when I was (I think) 17.

He explained the first of Mike's methods (you do have to know this, by the way, since at the crucial point it looks like the king is escaping) but prior to that, you need to know how to drive the king towards the "wrong" corner, which method is a little less tangible.

I think he referred to the technique as the "barrage", though as I say, it was a long time ago.

Anonymous said...

Wales international Nigel Saunders says he always teaches his pupils how to do the bishop and knight mate. Not so much, he says, for the knowledge of the mating pattern itself, but because it helps them realize how to make bishops and knights work together, which has much more wide-ranging application.

Anonymous said...

Answer to Number 2 required


Anonymous said...

My grade is 150-160 and have never studied endgames. I just think for myself when it comes up for real in a game.

I have never had bishop/knight endgame before so I haven't lost any sleep over not knowing anything about it.

ejh said...

My grade is 150-160 and have never studied endgames. I just think for myself when it comes up for real in a game.

Do you do the same in the openings?

If not, why not?

George - if I could answer your question, I would. But it's possible that the answer may alter depending on how old you are, how strong you are, what resources (boooks, computers, clubs) you have access to, how much spare time you have etc.

Anonymous said...

I've recently found Silman's Complete Endgame Course: From Beginner to Master very useful to work with, as it structures the endgame knowledge around what he feels is useful to know at various playing strengths.

He doesn't cover Bishop & Knight v King though...

Jonathan B said...

My own answer to (b) for what it's worth is that starting with some basic king and pawn endgames must be a pretty good idea.

After all a lot of those drawn or won rook and pawn endings are only drawn or won if you know when you can or can't exchange rooks depending on what you're trying to achieve.

What counts as basic/essential king and pawn ending knowledge and where we should go thereafter, however, are questions I don't have an answer to.

Anonymous said...

Anyone who doesn't feel at ease with the knight and bishop, invest 10 mins watching this video 5 times or so and you'll learn it for sure =)

Might earn you ½ point later on.


Anonymous said...

ejh: Do you do the same in the openings? If not, why not?

No, I have stuck to the same (narrow!) opening repertoire for a few years. I haven't been back to study those openings to polish up or look at new openings for about 3 years.

I am one those who just like to turn up and play rather than staying up half the night burning the midnight oil going through 25 moves of Sveshnikov theory! As it happens I have done well against it, winning against 183 last season and have never lost against it.

ejh said...

For what it's worth - last night I had insomnia and at around five in the morning I went and tried to make with B&N against the computer. I began with Kd5, Nd6 and Bb6 against Kd7, i.e. quite a favourable starting position, and it took me forty moves.

All right, I can't have been at my most alert, but on the other hand I didn't have a clock running.

Now I know how it's done (though obviously not perfectly). Can anybody tell me they could work it out from scratch with a clock running - quite likely close to running out, these days - and be sure of getting it done inside fifty?

Anonymous said...

In case anyone's interested, I just found this great tutorial on B&N by

I'd never come across the "Method of Three Triangles" before, but it seems reasonably straightforward!

Phaedrus said...

Studying this endgame is probably a waste of time. The chance that you get this endgame on the board in one of your games is very small.

And not only that, You only have to know how to play it when you are the one that is trying to win. So the very small chance that you have to use the skill to mate is even decreased by another 50%.

The defending side has to know only one thing: keep your king away from the corner square that has the same color as the bishop.

I take my chances and use my time to study things that are more likely to be useful in my games.

ejh said...

Does it occur to you that if you have not studied it, you wouldn't know the technical aspects of it that you've just advised?

Among the many resaons that you're wrong is that it's not just a question of being in a certai nending. it's about other endings that may lead into it - in which, if you do not know the (easily learned) technique involved, you'll be obliged to avoid lines that actually led to an easy win.

Robinson said...

I have to second the recommendation of Silman's Complete Endgame Course. Excellent book with a ready-made approach to studying the endgame -- start at the beginning and work up through your current rating (the endgames are arranged by what-you-need-to-know at each level).

I believe Silman generally dismisses the usefulness of studying the N+B ending as a too-rare happening (despite your experiences). I would think most people would better spend their time on various R+P endings, as they seem to come up so often and are so-little understood by average club players.

Anonymous said...

How to get better at endings? My 3 pennies ...
(1) Play for the ending in your games. In particular: (a) When clobbering your opponent, there will come a time when you can continue the king hunt, or cash in for the material plus endgame. Mate is mate, but the published brilliancies have all kinds of lines that end with "an extra pawn". Take the pawn and grind it out. Tip#1: There is difference between playing for the endgame when it is objectively best, and being a self-styled endgame "specialist". Don't be one of those. (b) When getting clobbered, you get the opposite choice. Either you can lash out with an unsound combination in the hope that your opponent goes astray, or you can grimly hunker down in a lost ending and make both players suffer for the point. The second method gets better results in three ways. (i) Your opponent might not know the ending (our current topic). (ii) You will learn something you didn't know before (see #2 below). (iii) You will get a rep as one tough cookie. Even when you lose, you can make your opponent hate winning. Tip#2: If you do eke out a draw, be sure to begin the post-mortem with the words "Of course you were completely winning."
(2) Study your games the way they were meant to be studied. Backwards, and no computer. Try to identify the *very* last moment when the game tipped from a draw to a win, and the *exact* sequence of moves that would have drawn. Do this before you even look at the opening. (If you do this right, you will probably never get around to the opening.) Then look in an endgame reference to check your result. Only then turn on the computer to check your analysis. Do this for your wins and losses. If you never reached an endgame, oh well, you will have to study the middlegame for a change. If the game was a draw, try to find the last moment when the "chances" vanished. Tip#3: The next time you are at the chess club, instead of setting up some opening position and asking everybody what they would do in that position, set up one of your "last moment" positions and ask the same question. Be prepared to revise your assessment of the position.
(3) Come up with a study program, and do it. My recommended program would include (a) the common pawnless endings (yes, including K+B+N v K) (b) piece v pawn(s) (c) king and pawn endings (d) K+R+P v K+R. Of course this means a book. A good outline of endgame topics is in Tarrasch's The Game of Chess, but it's only a chapter. Then again, this one chapter and some Reinfeld stuff carried me from USCF 1200 all the way past 1800. For king and pawn I recommend Sutherland & Lommer's 1234 Modern End-Game Studies. Reached USCF 2200 on that plus Znosko-Borovsky's book. Of course I have lots of endgame books in my library, but the ones mentioned were the ones I actually studied, along with the occasional Mednis and Benko column. Eventually I buckled down and worked through Fine's BCE start to finish, in ChessBase, with Fritz *off*. Only took a couple months of all-nighters. That coincided nicely with my peak USCF 2375. Those were the days. ... Here is a good test for an endgame book that you are considering: Look at all the diagrams that have K+R+P v K+R and count up the various files for the pawn. A good book will total in this order, most to least: center-P (d/e), BP (c/f), NP (b/g), then RP (a/h). This reflects their relative importance in practical play. Tarrasch has *zero* diagrams with a rook-pawn. Some books (Nunn's Secrets comes to mind) have the exact opposite ordering. I have Nunn's book but wouldn't dream of trying to learn from it. I did scan Silman's book at the bookstall and had a favorable impression, but these days I buy computer books. I agree with the idea that there are grade-appropriate endings to study. For example, queen endings are all tactics, so there is not much point to studying K+Q+P v K+Q below candidate master; although some general observations about where to put the defending king might be helpful. As a counterpoint, though, if you are going anywhere in life you sometimes have to get out of your depth, just to see what you are made of. So from time to time you might want to look at "difficult" endings just for grins. Tip#4: Pick one of the tough theoretical positions from your study program, and offer to play *both* sides against all comers at the club. See if you can hold the draw for the inferior side. Then turn the board and see if you can win. Then try to draw again now that your opponent has had a chance to see some ideas. I am not particularly recommending the Szen position for practice but, for those who know the story, the principle is the same.

An Ordinary Chessplayer