KBN versus K

[Arshad Ali]

I note Wiki says this ending occurs about once in 5000 games. Does that seem right? I’ve only witnessed it 4 or 5 times.

Do people also practice, say, Q v R or R + B v R or some of the R+P endings?

It occurs very rarely. I don't know the stats but once in a thousand games (or even less) seems plausible. I practice KNB vs K from time to time. I've practiced R+P vs R in the past. I've found Nunn's Secrets of Rook Endings very useful in this regard. And in practice I've found this has stood me in good stead as I've known instinctively what positions to aim for, whether a pawn up or a pawn down in complex rook endings. Skill in these basic endings seems to translate into greater skill in earlier parts of the game.

The reason I started the thread was I found a masterful 8-page discussion of KNB vs K in Pandolfini's book, Endgame Workshop.

[Roger de Coverly]

And given there is chat of adjournments elsewhere, would you adjourn K+B+N v K and would you expect it to be resumed?

With the caveat that the last time I played a game which might be adjourned was some years ago, if you had the B+N you would likely adjourn as soon as possible, as indeed you would in any ending you knew to be theoretically winning, but were unsure of the exact method. Equally if you were the defender and the technique once looked up relatively easy to apply, you would resign particularly if it involved travelling to an away venue. "Relatively easy" would apply to KBN unless you had considerable contempt for your opponent's endgame ability.

The 1 in 5000 statistic seems about correct. One recent British example was a draw in last year's 4NCL, but that could have been influenced by time shortage.

[Andrew Collins]

I have had 2 opportunities where I could of ended up playing out KNB v K, the first was a rapid play game where I thought I had made an inaccuracy and had no choice but to go in to this ending, however it wasn't necessary, so of course I didn't bother.
The second occasion was at an e2e4 event in Brighton, my opponent had blundered a piece and subsequently lost another piece due to it being trapped on the side of the board, he did have chance to force off my remaining pawns, though in the event he did not do so, however it was certainly the more practical choice to test my skill with that endgame. I practice with the computer, but like the idea of occasionally practising this sort of ending at the chess club with human opponents.

I also practice a couple of other endings with my table bases just to keep certain I can do them under time pressure in practice. Rook and connected pawns v Rook and King. Defending Rook vs Rook and Bishop is not very good with table bases on the computer because the comp doesn't set any tricks or traps and will just happily repeat the position, I had this once online in a blitz game which I lost.

Q Vs Rook never had it on the board, had opportunity for it to arise in a league individual match the other week, but my opponent missed the chance and I had an easier win. I do practice with my computer though, I think this is one example of an endgame where the likely practical human plan of keeping the rook close to the king all the time, makes the task simply easier, rather than having to play a few accurate moves when the King is forced for example on the Back rank and the rook shuffles from side to side on the 6th rank

I like endgames! I'm going to see if anyone at my club wants to practice a few of these endgames with the clock, perhaps rook+connected passers, KKB and lucena postion etc...

[Angus French]

For what it's worth, my database of 5,310,410 games has 1,254 KBN v K endings (so roughly 1 in 5000 also), with 299 drawn.

[Andrew Collins]

is there a way for me to search my database with Fritz 13 like this?

[Jonathan Bryant]

Q Vs Rook never had it on the board, ....

I wrote a blog post a year or two back asking if anybody had ever had queen v rook. A friend wrote back saying that he and I had played it in the last round of a rapidplay tournament. I have no idea how it went other than I lost (with the rook). Perhaps that's why I forgot it!

I had KRB v KR on the defending side in an internet correspodence game. The position on the board was favourable and at three days a move it turned out to be pretty easy to defend (using Nunn's Secrets of Pawnless Endings as I guide).

Angus and I quite often meet up to look at endgames. We tend not to play them out as much as analyse them. Today's session was dedicated to Vancura.

[Angus French]

is there a way for me to search my database with Fritz 13 like this?

Andrew, I think you need chess database software such as ChessBase or SCID. I use SCID which is free and accepts PGNs as input (though you're better off converting the games to SCID's native format if you want to do extensive searches).

[Andrew Collins]

thanks for that angus

[Martin Benjamin]

Probably putting the mockers on myself here, and it will occur in my next game, but I am not 100% sure I would win with K+B+N v K these days, if I had only a little time. I remember learning how to do it when I was young, but I have never had the position on the board. Apart from a little embarrassment that someone of about 2000 - 2100 ELO really should be able to do it blindfold, it does not worry me though. I think that if I had the time, desire and self-discipline to study chess, I would practise many other aspects of the game, and when it came to endings, I would prefer to hone my technique in basic rook and pawn ones. I may throw away half a point every 5,000 games on average by failing to deliver checkmate with bishop and knight, but that would be nothing compared with the hundreds or thousands of half points and full points I am still in the process of throwing away in the other 4,999 games through misunderstanding many other middlegames and endgames.

[David Blower]

I think there is another reason for practicing KBN v K endgames. Yes first and foremost there is a slight chance that you may have to do it in a game.

But perhaps more importantly and more realistic is the fact that you can see how well the bishop and knight work together at forming walls and blocking out squares.

[James Toon]

I had the KBN a few years ago and failed to deliver mate within 50 moves. I was short of time, but that's not an excuse – I learned the technique a long time ago (in one of Averbakh's books I think) but I hardly ever revised or practised it.

That led me to look up the technique in a few endgame books. The best explanation I found was in "100 Endgames You Must Know" by Jesus de la Villa. He uses the W technique rather than Triangles, but explains the procedure very clearly so that you understand the logic behind it. (There is more to it than the W technique, but it comes in at the most difficult point).

Jesus de la Villa says that with perfect play the mate takes a maximum of 33 moves starting from the most unpromising position and meeting the best defence.

I've also failed to win with Q v R, and in several R+P endings I've had the nagging feeling that I missed a win somewhere. All those dropped half-points do add up, so this summer I'm going to put away the openings books and concentrate on endings instead.

[Mike Truran]

Number 2 son can now do this in less than a minute from any position. That's a few hours better than his dad's best time.

[David Robertson]

But why bother? Whole populations of chess players go through entire lives never coming close to this ending. The probability of ever encountering it is vanishingly small. It would be more useful to learn to clip your toenails while standing on one hand.

(added) OK, I've now read the wiki on the matter, and find illustrious others have asked the same question. I read that the ending occurs once in 5000 games. Assume an active chess career of 50 years and one quickly sees the colossal improbability of ever meeting it.

Other benefits? I'm sure Mike's No. 2 son, if he enjoys solving this puzzle, will go on to enjoy much more about the game. So that's good.

Can I solve it? Yes I can - and have - but not in live play. Over dinner a few years back, Nigel Short and I were discussing the ending. He challenged me to solve it. So I did. At table. Blindfold (using the 'W' and 'cage' method)

[Paul McKeown]

@Dave R

I think David Blower gives one good answer to your question regarding why one should know how to mate with KBN-v-K, how to get that rather clumsy combination of pieces to wotk together. Another is that if you actually know how to do it, you no longer go out of your way to avoid it as the stronger party. I have done it once in a rated/graded game, subsequently I have probably done it at least half a dozen times in blitz games, you just let it happen instead of wasting time trying to save rubbishy pawns. In any case, it certainly is a much more frequent occurence than KBB-v-K for instance.

[Mike Truran]

Paul

That's pretty much where number 2 son's coach and I got to:

1. Improves your dexterity with a couple of pieces that don't always work well together.
2. Means you don't play sub-optimal moves just to avoid the ending.

Moreover, having him do it with five, then four, then three, then two minutes on the clock certainly seemed to be helpful as a time trouble management exercise.