The Chess Mind

We started here, with this deceptively simple position; the task White to move and win:



I also requested that readers who attempted to solve it let me know their (attempted) solutions and their ratings. Between readers, friends and students, it seems that for players 2000 and up, it's pretty straightforward; around 1700-1800 most of the key ideas are spotted fairly quickly but some detail or other is usually missed; below that, only dogged effort will succeed (and occasionally did, as in the case of regular reader sbb1cpa).

So thanks to everyone for their feedback; I might try something like this in the future, too, to build up a collection of positions that can be more precisely ratings-indexed than "easy", "kinda tough", "migraine-inducing", etc.

And now, the solution - click here.

Here's our starting position, originally presented in this post:



It's Black to move and draw, from the game Golod-Seeman, Cappelle la Grande 2006. Think you've got it? Click here and find out!

Finally, here's a third exercise for your solving pleasure:



This exact position is my invention (at least to the best of my knowledge), but I've seen the key idea at least twice before: once in a rapid Jay Bonin game, and once in Alexander Baburin's excellent Winning Pawn Structures. (Cf. his analysis of Kholmov-Kremenietsky, USSR Trade Unions Ch. 1981, pp. 105-108. A note about the book title: it's inapt (though it wasn't Baburin's fault), as 90% covers the isolated queen pawn and the remaining 10% the affiliated hanging pawns and isolated pawn couple structures. It's not a general treatise like Andy Soltis' Pawn Structure Chess or Hans Kmoch's old Pawn Power in Chess.)

This background out of the way, the task is to evaluate 1...g5. The solution will be given on Wednesday.

Courtesy of reader Joseph Amaral, our second king and pawn ending study is a reasonably well-known effort by F. Teed (1885):



White to move and win.

For this problem, I'd like to make a special request of my readers: please write me, via this link, to let me know two things: (1) your attempted solution and (2) your rating. (But please do not give the solution in the comments!)

The solution will be given on Tuesday.

Solving king and pawn endings is among the best exercises available to a chess player.

For starters, the reduced material makes calculating everything in one's mind relatively easy, even when the variations are extremely long. Second, pawn endings are fundamental, so skill in solving studies is very likely to translate into improving one's playing skill in a pretty direct fashion. And finally, they're just plain interesting! (At least I think so, and I hope some of the exercises I present will leave you feeling the same way if you don't already.)

Here's the first one:



From an actual game, Black to move (and draw). (Solution tomorrow; quasi-hat tip: Chess Today-1936)

A problem many young chess players suffer is an inclination to what's often called "hope chess". Hope chess works like this: one sees an idea for himself - often a cheapo - and then considers only second-rate defenses for the opponent. In short, he calculates for his opponent only what he hopes he will play.

Most of us overcome this difficulty quickly enough; understandably, as the path of hope chess leads to destruction. Human beings are good learners, but we're not always good at meta-learning. What do I mean? Just this: we may a learn a lesson in one particular context, but not do such a good job of applying it to broader contexts.

For example: when players start working on studies, they'll often find some good idea for White (studies are almost invariably White to move and win or draw), make natural replies for Black, and consider the problem solved. Unfortunately for them, that's not how studies work! Any study worth its salt will have at least one of two features: a seductive false lead, refuted by a brilliant Black resource, or a brilliant Black resource in the main line topped by an equally brilliant White rejoinder.

Thus if the inexperienced solver thinks she's done the job by means of the first good idea she finds for White, she's almost surely mistaken: she's guilty of another version of hope chess.

Here's an example from a study I solved earlier today.



(M. Liburkin 1949) White to move and draw.

The problem is clear: Black has two pawns ready to queen and White isn't stopping either of them yet. That makes the first two moves obvious (or at least apparently so):

1.Ng3+ Kh4 2.Kb2.

Both pawns are stopped, and now if Black plays 2...Kxg3, White uses his bishop to take the knight's place: 3.Bc6 with a draw (3...c1Q+ 4.Kxc1 Bf5).

Problem solved? Hardly - it's too easy! White didn't have to find a single difficult move, and there wasn't anything aesthetically pleasing about the solution, either. So for all you budding solvers out there, remember to look for at least two things in a successful solution: (1) real resistance, and (2) a meaningful aesthetic component (flow, paradox, sacrifice, humor - something will grab your attention and hopefully make you smile).

With that in mind, I wish you success in solving this puzzle (and others); you can find the answer here.

We start with this innocent-looking position



and try to figure out how White (to move) is going to save it. I presented it on Saturday, and if you think you've succeeded in solving it and desire confirmation, or if you've tried and want to see the answer, enlightenment is just a click away.

Here it is; White to move (and draw, obviously, as no other goal makes sense):



J. Fritz, 1965

The solution will be given in a day or two; as usual, I ask those who have solved it to be courteous and not spill the beans in the comments section.

John Nunn is not only a very strong OTB (over the board) grandmaster, he's also a real connoisseur of studies and problems: both as a composer and especially as a solver - he is the current world champion at that discipline!

Having recently turned 50, a study competition was held in his honor, and you can find the winning entries, with solutions, here in both .pdf and .pgn formats. In my opinion, solving studies is one of the best ways to improve, and I'm sure these will prove worthy of one's efforts. (Warning: while I think practically anyone will benefit from trying to solve these, these are likely to prove challenging even to those on the higher end of the rating spectrum.)

Yesterday, I introduced a position tough for humans and really tough for computers to solve, and in case it's not tattooed in your minds yet, have another look:



M. Matous, Szachy 1975


Did you solve it? Apparently at least one IM and one NM were stymied, and none of my software came even remotely close. Fortunately for my sanity, I solved it, and I hope some of you did as well. Of course, you're welcome to keep trying, but if you want relief, the answer is just a click away...

Here's a puzzle that will make you work and give your computers heartburn: White to move and win.



M. Matous, Szachy 1975


A few weeks or months ago, I came across a slightly different version of this puzzle, with the Black pawn on a6 (rather than a7, as mistakenly given earlier tonight, or a5, as Andrey has since informed me and as given above), and after a few minutes' thought found the right idea.

Unfortunately, checking it with the computer way back when, it found a defensive try I hadn't considered at first, and as it hadn't yet detected a win for White, I assumed I was wrong and left the puzzle to languish until today.

After watching the computer's futility this time around and after getting the right position at last, I went back to work by myself. I quickly convinced myself that my initial approach had to be right, and then it was just a handful of minutes until I overcame the final difficulty. It's a nice problem, but I have to say I'm not nearly as impressed by it as I was initially inclined to be, as (a) I solved it relatively easily and (b) there's an aesthetic flaw: the Black pawn can be placed on a5 or a6 without any substantial difference.

A word of encouragement in closing: while it's true that it's a real pain for current chess engines, a diligent human who has solved studies in the past should be able to work it out. You can do it!

Again, out of consideration for other readers, please don't state the solution in comments (or even hint at it). If, however, your chess software can solve this study from the starting position, I would like to know about that!

The solution will be given later today (Wednesday).

Further update:

If Black's a-pawn is on a6, a5 or a4, or if Black's a-pawn is missing altogether, White wins. (The latter scenario also has the drawback of being solvable by Mr. Chips - so don't test it until you're 100.000000% sure you've got it right!) If Black's a-pawn is on a7 or a3, then White's usual winning idea fails nicely. Finally, Black is just winning if the pawn is on a2.