Chess.com's Weekly Study: September 11th 2016

Hello Chess.com!

Starting with 2016, we will be posting a Weekly Study, courtesy of Yochanan Afek, Grandmaster for composing endgame studies. These challenging positions are designed to stimulate (and improve upon!) your creativity, depth of calculation, pattern recognition and pure imagination.

Z. Birnov, 1957

Enjoy! 

The variation in the comment (after 4. .. Nh4+) is easy to foresee but what was really troubling me was the defense 5. ... Ng6. Though white wins in several ways as per tablebase, none of them is obvious to a human solver.

It shows that the issue of 'tablebase wins' existed long before the tablebases. Some analytical lines in endgame studies were provided by authors but ignored by solvers due to complexity. Sometimes you just have to 'feel' your way through parts of a solution to arrive at the other end. Which is exactly what I did.

Possibly Birnov's analysis after 5...Ng6 below in white.

6.Nd3 (stops black knight escaping via e4)

Knight has now only h4 or f8.

If immediate 6...Nf4 then 7.Bc6+ and 8.Be4 immobilises the knight and king can capture.

If immediate 6...Nf8+ then 7.Kf7 Nh7 8.Nf4. Now if 8...Ng5 9.Kg6 and the knight has nowhere to go. But if the knight stays put then after 9.Ne6 it is immobilised and king can capture.

If Black moves king or pawn first, then 7.Kf6. If 7...Nh4 then 8.Ne1 immobilises the knight. If 7...Nf8 then 8.Nf4 and either 8....Nh7+ 9.Kg6 Nf8+ 10.Kf7 Nh7 11.Ne6 or 7. K or P (any) 8.Ke7 Nh7 9.Ne6.

White then captures the pawn and wins the bishop and knight ending (or possibly just wins the bishop and knight v pawn ending which can often be quicker)..

In fact 5...Nc8 (mate in 2) makes it seem as if Black had a death wish, when compared to 5...Ng8 (mate in 40).

The global truth on studies is that all defenses together are the solution. Since they all lose, none is worse than the other. In publication only the interesting variations are picked. Ignored are usually the boring 'tablebase type' variations no matter how long they resist. They used to be in the addendum in smallprint - when print was the standard. Today, they are hardly even mentioned since 'everyone' knows that all studies are engine-checked and are therefore OK. The deathwish in endgame solutions is equivalent to the attitude that "if life ain't exciting then life ain't worth living at all".

In the future when there are decent puzzle interfaces these will be capable of filling the holes in the pruned solution tree. They will play you on any move you pick for either side in any problem or study.

I guess that Na5 as the first move on the study, threatening Bc6# next is not good, since black would play a6.  Is that correct?  I will set this study OTB to see the possible moves in this study.  Very interesting.  

5...Ng8

6.Bd7

If immediate 6...Nh6 then 7.Kf6 and either 7... Ng8 8.Kf7 Nh6 9.Kg7 or 7...Kb8 8.Kg7 Kc7  9.Bh3.

If 7...KB8 8.Kf7 Kc7 9.Bh3

Think 5...Ng4 is better try.

Note on the format. It continues to surprise me that so many puzzles are posted under the author "AAA vs ?" heading. This is not clever since the author isn't playing anyone. The proper way to do it is to enter the author name under the "event info".

It shows how many people are programmed to docilely fill in forms: enter 'names' in 'name fields', 'dates' in 'date fields' etc. Remember, we are boss, not the forms or the computer programs. 

I think you're right it's not good, but Black would probably attempt to connect his king and knight so I would guess 1...Kb8.

Er, just noticed the "#" - don't think that's right either. 

But we just learned from Arisktotle that no move (by the losing side) is worse than any other. It follows that no move is better than any other.

I always thought that the best defence (or one of the best) was traditionally given as the mainline with other defences shown as "(If ... then ...)".

Here I would have expected 5...ng4 to be the solution (though without playing out the bishop and knight ending) and 5...nc8 to appear in the comments.  

I will accept then that all moves are created equal in studies (though I'm still inclined to think that some are more equal than others).

What I meant was that if it occurred in a game Ng4 would set the opponent the harder problem and thus be more likely to actually result in a draw. 

"Harder for the (human) opponent" sometimes counts as 'more interesting' but sometimes it counts as the precise opposite 'more boring'. On the average good endgame studies are hard but not extremely hard. More important are adjectives like: thematic, beautiful, unexpected, sacrificial, paradoxical. Or in general all the things you are ill-advised to do in the first order approach to playing chess correctly!

In the current study the 'unexpected' aspect is white chasing the black knight to the black king where it is best capable of defending it. And then the surprise 'checkmate in two'! What is also in the background of this study is 'domination' of the black knight. 'Chasing pieces' is a somewhat unusual strategy in 'first order chess' since they usually outrun their attackers and find shelter, but it is quite common in studies especially against knights.

Ok. In the context my comment was misplaced.

But the solution as given still leaves one thinking, "Why did he do that?" at the end. I think some analysis of the alternatives in the comments might have been appropriate.

(You mention this yourself in your first post.) 

I am not aware of any misplaced comments. The solution given by chess.com is (usually) below minimal. I suppose that Birnov's own solution is about right for peer composers and good solvers - in the perspective of that time, 1957. Today it might be further reduced since all have access to some TB plus a good game playing engine.

In general an author contributes what he believes is the content of a composition, which is mainly centered around his 'themes' and 'surprises'. Everything else is checked against the computer as well but is generally left out. Especially left out are variations where white has more than one move to win (duals) since they have no value for the composition. Personally I might give one or two example moves and then assert that 'white wins'.

See for instance the comments I gave in the endgame I composed and posted in 'ChessReina's X 24' puzzle (last post 4 days ago). I had some lines checked by Stockfish and Houdini but included none of their analysis. They were simply off-theme.

Note: The strategy for solving an endgame is always to detect the ideas (themes) of the study. That means ignoring what looks like hard to analyze variations. They probably don't matter and they probably have precisely the outcomes they must have in order to have a correct study. That's is the 'reverse engineering' equivalence for study and problem solving. It's one of a bag solving tricks. Another important one is: if 'x' would be a solution then 'y' would be a solution as well, therefore 'x' can never be a solution!

Arikstotle wrote:

for instance the comments I gave in the endgame I composed and posted in 'ChessReina's X 24' puzzle (last post 4 days ago).

This one, Problemas y Finales (X 24) ?

Certainly no misplaced comments there! (Liked it by the way - already did it a few days ago.)

Arikstotle wrote:

The strategy for solving an endgame is always to detect the ideas (themes) of the study.

Sounds like a good tip, I'll try to apply it. (I suspect without too much success. I usually just look for forcing lines.)

@MARattigan:  Like you, I have trouble overlooking the "boring" lines. In the current study, is was interesting to me that after 5...Ng8 6.Bd7 Kb8 7.Kf7 Kc7, out of 22 legal moves White has only one move that wins. It didn't bore me to discover that. Maybe it means I'm titillated by the wrong things.

@9531I: Interesting. I saw it was the only place for my bishop in the line I was taking but didn't check out the other 21.

I really liked 5...Ng6 6.Nd3 Nf8+ 7.Kf7 Nh7 8.Nf4 Ng5 9.Kg6, which is a very problem like line. Totally failed to notice 9.Kg6 did the trick first time through the analysis.

Yes, you got the right endgame.

'Forcing lines' is another good tip. But as you see in the current study,  'forcing lines' do not always exist.

And of course, it is OK to have as a hobby the research of the margins. Some decades ago, composers would spend a lot of time researching these themselves as well. Now, they feed them to an engine while underway with the composition and if they fail they simply change the setting a bit until they get one that works, even hundreds of times. What is vital to a composer is that his study does have an idea with a 'logical justification basis', implying that a correct study 'might exist'. When there is a logical flaw, then every setting will fail! It gets us a good separation between man and machine. We do the ideas, the machine does the verifications. Most composers like it.

Note: When you participate in a solving contest, you get a piece of paper with a small white area (for at most 2 or 3 short variations) per solution. It teaches you to focus on the essentials, both for the white and the black moves.

The only solving contest I have participated in, about 56 years ago, when I was a relative chess beginner, was arranged by a classmate among a group of high school students. The positions were studies, and only the correct key move was needed to get credit. On the final position, I needed to solve it to win the contest, but I had no idea what to do. When my time was about to expire, I just guessed the silliest move I could see, and it turned out to be correct.