Reviews | August 19, 2010 18:03

Review: Nunn's Chess Endings 1

Review: Nunn's Chess Endings 1Almost as soon as I opened Nunn's Chess Endings Volume 1, I realized the book was probably an 'instant classic', ignoring J.M. Coetzee's definition of a classic as 'that which survives'. My intuition proved to be correct in that I think this really is a fantastic book. The only problem is: who plays endgames anymore these days?

Not me, to be sure - at least not frequently enough by far to devote an awful lot of time to all the subtleties of the practical endgames Nunn describes and analyses in his book. This makes my task as a reviewer a rather difficult one: though I think the book is absolutely brilliant, I feel sorry for the author already because I fear his audience is becoming smaller and smaller.

Back in the good old days, when games were still adjourned and FIDE hadn't introduced the 'blitzing out' even of officially rated games on all levels of play, endgame manuals were absolutely essential for anyone wanting to reach a decent level in chess. Nowadays, it seems to me, things are rather different - at least for the majority of club players (and I suspect even beyond that level).

As an experiment, recall the last time you've seriously had to play a complex endgame and was able to actually invest some time in it (say, at least an hour). Next, recall the last time you've seriously analysed a complex endgame, say for at least an hour (preferrably over the board). I even suspect many strong players have never done this at all in recent years. And who can blame them?

But then again, reading Nunn's Chess Endings, published by Gambit, is such a joy that it almost makes me want to introduce a new rule, once suggested by a club member of mine, Wim Nijenhuis: instead of 'blitzing out' games (endgames, that is), why not 'blitz in' the opening phase (which everyone knows by heart anyway) and actually allow some thinking time for endgames, which are infinitely more interesting?

As exhibit A, consider the following position, which Nunn took from the award winning Van Perlo's Endgame Tactics:

O'Kelly - Forintos
Bordeaux 1964

Van Perlo comments that the position is quite simple, and perhaps the New in Chess editors thought that it was so simple that it didn't need checking with Fritz or with the tablebases, but if so they were mistaken. Although White's knight is as far away from the pawns as is possible on an 8x8 board, it can still make it back in time to save the game.

1.Nc7 Kd4!? This odd-looking move at least sets a trap for White.


Van Perlo considers this to be the losing move, but he is wrong. One suspects that this error is derived from the analysis given by Ugrinovic in the Encyclopedia of Chess Endings, since Van Perlo repeats the faulty ECE analysis move for move. 2.Nb5+ Kd3 3.Nd6 is another way to draw, but not 2.Ne6+? when Black wins by 2...Ke3 (this is a position of reciprocal zugzwang) 3.Kg1 g2 4.Ng5 f2+ 5.Kxg2 Ke2.

2...Kd3 3.Nd5? This is actually the losing move. White can still draw by 3.Ne6 (not mentioned by Ugrinovic or Van Perlo) 3...Ke3 4.Kf1! reaching the reciprocal zugzang mentioned above with Black to move, and after 4...g2+ 5.Kg1 Ke2 6.Nd4+ Ke3 7.Ne6 the draw is clear. There was even a second draw by 3.Nb5 Ke3 4.Kf1!

3...f2+ 4.Kf1 Ke4 0-1 After White's knight moves, Black wins by 5...Kf3.

Many things could be mentioned about this fragment, such as Nunn's well-known tendency to write in a polemic style, or his insistence on computer-checking all lines (which in fact is why he brings up the example in the first place), or his erudition regarding other endgame classics. But what struck me most in this passage is the fact that, apparently, nobody had noticed these drawing lines before, despite the fact that the game is over 45 years old and has been published in at least two major endgame manuals. There's so much still waiting to be discovered!

But of course, to me the real question seems to be: why has nobody noticed before? I fear in part it has to do with the sad fact that almost nobody actually analyses these textbook examples to such depth anymore. From my own team, mostly filled with 2200-something players, I think it's fair to say nobody (including me) really analyses this type of endgames anymore. We may stumble upon, and even appreciate their innate complexity and beauty - but to actually analyse it and try to understand what's really going on, that's something else altogether.

The book's introduction and the first chapter (The Three Key Endgame Skills) are some of the best endgame-related chess prose I've read in a long time. In it, Nunn tries to prepare the reader for the abundance of practical endings that are waiting for him in the following 300+ pages. In fact (and this will no doubt come as a disappointment to the author), I think Nunn's prose is by far the best part of the book - at least for people like me who simply don't have the time and energy to really immerse myself in all these examples, however fascinating they are. Here's how Nunn introduces pawn endings with outside passed pawns:

The strength of the outside passed pawn in king and pawn endings is drummed into chessplayers by every textbook on the endgame. Using such a pawn, it is possible to deflect the defender's king away from the main mass of pawns, leaving a rich harvest for the attacker's king when it gets among the pawn mass. There is a good deal of truth to this accepted wisdom, and in the first section below we shall look at cases in which the outside pawn lives up to its reputation.
Yet this reputation is to some extent overstated, and in the following section we shall look at cases in which the outside passed pawn proves a hollow threat. The third section deals with an interesting practical case: one side has a powerful outside passed pawn, but the opponent has an extra pawn.

This seems like a pretty meagre bit to quote from such a richly illustrated book, but the point I want to make is that even if you read only this small introduction, you might learn something subtle from it - perhaps it will even prove to 'stick' better than when you analyze all Nunn's examples in-depth. I've often seen discussions about pawn endgames being cut off by someone simply declaring 'outside passed pawn', implying an obvious and straightforward win. While only introducing his real material, Nunn dismisses such simplistic notions while making the reader curious for more. Such as, for instance, this:

Kirov - Ermenkov
Sofia 1973

Black has an outside passed pawn and at first sight the win should be simple. He pushes the h-pawn, deflects the white king, marches with his king to take the e3- and b3-pawns and then promotes his a-pawn. However, one aspect of the position favours White: he only needs to take the relatively close d6-pawn in order to create a passed pawn of his own. Another factor which is not obviously relevant in the diagram position, is the weakness of the b6-pawn. These compensating factors mean that White is just able to hold this position, although accurate play is necessary.

1.Kh3!! It was quite an achievement for White to find the only move to save the game. Other moves lose (...).

1...Kf5 (...) 2.Kh4 Ke4 3.Kg5 Kxe3 4.Kf5! It takes too much time to go for the h-pawn, so White must create his own passed pawn as quickly as possible.

4...h5 This is the only chance, since if Black runs for the b-pawn, White promotes first.

5.Ke6 h4 6.Kxd6 h3 7.Kc7 h2 8.d6 h1Q 9.d7

It is perhaps surprising that Black cannot win here, but this is the point at which the weakness of b6 enters the picture. Black cannot force the white king in front of the d-pawn and the best he can do is transfer his queen to e7 with gain of tempo. Then he has a free move before he has to exchange queens on d8. If Black's pawn were on a7 instead of a5, then the resulting king and pawn ending would be winning for Black, but as it is, White is in time to take on b6 and create a passed b-pawn (...).

One of the funny things about this example is that my computer evaluates the positions in the beginning position and the final diagram as a simple win for Black (-4.90). I'm confident most chess players wouldn't bother to analyse the endgame even without this evaluation, let alone seeing the machine agrees with their intuition. But chess is not a matter of intuition only, but also of concrete analyses, which is exactly why it's such a pity endgames are so rarely (seriously) played these days.

Another thing the attentive reader has perhaps noticed already, is that Nunn almost always takes his examples from either new or relatively obscure games, rather than drawing on existing and heavily analysed material from World Champions and other greats. The book doesn't contain a single example from Karpov's, Fischer's or Capablanca's games, and only one from Smyslov's. (Of course, this makes cross-reference checking with endgame books pretty much impossible for reviewers!)

Nunn's Chess Endings Volume 1 is not a book with elementary endgames. It's a book dedicated to practical, concrete examples requiring precise analysis. Technique is another part of endgames, outside the scope of this book. (Indeed this is why Karpov and Capablanca are so conspicuously missing!) Clearly, in this format lies its greatest trump and its greatest risk. Who will go where the author went? In the introduction, Nunn optimistically notes that "I believe that anyone who works their way right through the books will see beneficial results."

This is surely true (though isn't it true for almost every serious chess book?), so I prefer to go one step further: anyone who casually reads this book will see beneficial results. Perhaps this is the chess-definition of a 'classic': it survives even in the light of modern laziness. I'm sure Coetzee would be pleased.


Share |
Arne Moll's picture
Author: Arne Moll


Ardjan's picture

Hi Arne,
funnily I would draw the opposite conclusion in this era without adjournments: it has become much more useful and important to study the endgame! I still play them a lot, also in our team matches.
By the way, I notice part 2 of these series has also just been published.

chandler's picture

fantastic review!! your best in my recollection...

jussu's picture

People in chess are remarkably different. This review makes so many (implicit) pointsa that go against my feelings that I cannot resist the temptation to show my humble opinions.

"But of course, to me the real question seems to be: why has nobody noticed before?"

To me this is an obscure detail in chess history - maybe interesting for someone else, good luck, don't mind me passing by and shrugging.

"As an experiment, recall the last time you’ve seriously had to play a complex endgame and was able to actually invest some time in it (say, at least an hour)".

Around 1990 when I played correspondence chess. But why the question? If we had an hour to analyse an endgame before making a move then we would not need books like Nunn's Chess Endings. Alright, this was an exaggeration, but as Ardjan mentioned, the need to play endgames well in any given moment with very short time is precisely why endgame books are useful.

"Next, recall the last time you’ve seriously analysed a complex endgame, say for at least an hour (preferrably over the board)".

12th June 2010, last time I played in an OTB tournament.

Teplitz's picture

Studying the end game improves your whole game, it is shortsighted not to study them just because you rarely get to play end game positions.

It reminds me of the remark made to Mark Dsvoretsky when he asked a strong female player if she studied the games of Alekhine. She replied "no I dont need to, its not like I'm ever going to play him"

Tom's picture

Good review and one that makes me think that if I had time to study chess, I might start at the end.

Doubt Coetzee would be pleased at the idea of a classic chess book given the fleeting, negative mentions chess gets in his books!

Felix Kling's picture

When I look at my last tournament games, 3 games ago I had an endgame, actually I calculated a rook endgame with a lot of pawns down to a winning ending queen vs. king and pawn (I reached it with a forced series of trades but spent quite a lot of time on it, maybe an hour). Also if I look at my games in the database I've got quite a number of endgames, also because of my openings with black, but some with white, too (I remember a nice pawn ending where I used 45 minutes for calculation just after the time control, but I didn't find the most precise moves, although I managed to get a draw since my opponent was also not able to find the tricky win. It was actually one of the first positions in the endgame manual by dvoretsky that arose (pawn b3 vs. a5 and c5, kings on c3 and b5)).

I like the endgame manual by dvoretsky, as you get the most important ideas and tricky examples for their application, although the presentation of the material is not perfect (but didactics in chess books is another case :) ).

So I can't share the view that endgames are not important anymore, what I would like to see is a concrete comparison of the book with the other books (like dvoretsky or the more practical book by silman).

Btw., chess engines, including Rybka which you may have used (looks like the author mentions Fritz in his book O_o) have the most problems in endgames. They simply don't understand fortresses of any kind, so if you have a position without being able to make progress, the chess engine won't get it. This is true for positions with opposite coloured bishops or drawn rook and pawn vs. queen endings for example. But sometimes you can show the engine the draw (like in this example if you play some more moves) and then go back, after the engine saw the draw.

Arne Moll's picture

Ardjan, I agree working on your endgame intuition has probably become more important, but I can't help feeling that really delving deeply into positions and studying all the subtle details (which is what this book is about!) has mostly become something of the past for many players. Sure, general technique and the basic essentials can be extremely useful, but what use is it trying to find some deeply hidden win in a highly complex pawn-endgame when you've only got 5 minutes left to finish the entire game, or 30 seconds per move?

noyb's picture

Thank you for the excellent review Arne. Reviewing books is a difficult task and you do many good reviews. Despite already having around 40 endgame books in my library, I may well have to spring for two more!

Eric's picture

Arne, Thanks for discussing that which makes this endgame book different from other endgame books; you've helped me make an intelligent buying decision.

Suggested improvements to your text:

this type of endgames
should be
this type of endgame
[think: 'this endgame type' rather than 'this endgames type']

Here's now Nunn
should be
Here's how Nunn

drummed into chessplayer
should be
drummed into chessplayers
(but check the book itself and use [sic] if the word chessplayer is verbatim, or does Nunn say 'drummed into a chessplayer'?)

by someone simply declaring
should be
by someone's simply declaring
[gerund 'declaring' takes a possessive pronoun to make a complete nounal object of the preposition 'by']

René Olthof's picture

There is little arguing that John Nunn again has turned in a great job.
However, he used the 2006 first edition of Van Perlo's prize winning Endgame Tactics to quote from - see Example 1 in the review above. That's quite unfortunate because in the third edition issued in 2008 the text has been completely revised and brought up-to-date with the latest 6-men endgame database technology. That's a bit like using a telephone directory or encyclopaedia, not from 2010 but from 1996, and then complaining that some numbers or items are outdated.

Volume Two is about rook endings.

If you want to get an autograph by the Doctor himself - go to Hotel Krasnopolsky in Amsterdam before the NH Hoteles TOurnament ends there!

Rick Vaughn's picture

Dr. Nunn was making a point about how much one can trust previous analysis merely using Van Perlo's as an example of a book that was "computer checked." The example is one that could easily have been checked in 2005 using a 5-piece database since it is a 5-piece EG. One doesn't need a 6-piece TB. But even without an EGTB, the solution is quickly found by top chess engines circa 2005-2006 which was the point — caveat emptor. Or perhaps, trust but verify.

Arne Moll's picture

Hi Eric, thanks for your corrections. I can't say that I understand what's wrong with the first and the last though. Seems to me both are correct in terms of grammar and style, though perhaps the alternatives are more to your personal taste?

René Olthof's picture

Ricky Vaughn:
you would be right if this was one of a small number of examples.
In the first edition we wrote that a computer was used to check material. This information was misleading in that people took it as a confirmation that all material was 'computer checked', which was clearly not the case. Also some cases had been checked.
In the third edition this mistake had been straightened out.
What Nunn wrote is completely true en acceptable - I only wanted to point out that it is unfortunate (and frankly) incomprehensible that he didn't use the third edition but rather the first. The reason for this is clear: when he started the project in 2006 the first edition was there was.
However, when you transfer the whole thing to opening analysis I think you would agree with me that when you start writing a book in 2006 on any given opening line and publish the book in 2010, it makes sense to update your sources in the working process and not stick to the original 2006 sources.
ECO writes this or that implies that you have been using the most recent ECO, not ECO 1 or 2.

René Olthof's picture

'Also some cases had been checked.' should be 'Only some cases had been checked.'

René Olthof's picture

when he started the project in 2006 the first edition was there was. -->
when he started the project in 2006 the first edition was all there was.

Arne Moll's picture

I don't know, René, the analogy with opening books is a bit unfortunate in my opinion. Opening theory changes not because of technological improvements or because authors discover mistakes in their books, but because of practical experience and home analysis by thousands of individuals from all over the world and of all levels of play. On the other hand, endgame positions and their analyses, are much more static. There's much less urgency to use the latest edition of an endgame manual than for an opening book.

Besides, I can't really blame Nunn for taking the 2006 edition. After all, this was the award-winning edition, not the later versions of the book.

Also, even though the later editions do mention that improvements were made compared to the first print, I assume most readers still think of improvements in terms of minor corrections and additions rather than multiple corrections and even radical changes of analysis, wouldn't you say?
Especially in relatively simple positions, and in this particular example this was even explicitly stated by the author himself! Finally, in defence of Nunn's somewhat polemic choice of words, I think it's fair to say that in this example, the author didn't just 'improve' on his analysis, but in fact totally changed his initial assessment of the position.
That said, I enjoyed Van Perlo's book a lot!

Latest articles