1. Bc2! 1-0. White has a winning attack.

I gave up and peeked at the computer. solution is awesome but now I feel I should have got it

My computer is listing three different lines as equal (all evaluating at exactly 0.00) at depth 20. Houdini 1.5. How deep does it have to go to find a clear solution?

Edit: The answer is depth 21, lol.

Edit 2: What the **** are these moves I don't even...

I didn't move any pieces ( except in my head; had the position etched in memory ) so my guess is:

my line as before but 3. Ke1 Re4+ 4. Kf2 Rf4+ 5. Ke3 Rf6 6. Qxg6+ Rxg6 7. Kd4 etc. and now there is a winning pawn ending; there's a lot of little details like 4...Re2+ 5. Kf3 Rf2+ 6. Qxf2 etc.

1. Ba4!! ( the point is to get the c5 square for the Queen! ) b5 2. fg6+ Qxg6 3. Bc2 Rxf4+ 4. Kg1 Rf6 5. Qg5! Rxe6 6. Kf1 Rf6+ 7. Ke1 8. Re6+ Kd1 9. Rd6+ Kc1 10. Rc6 Qc5! etc.

I was led astray by all those drawn pawn endings...

EDIT: looked at RoundTower's attempt; yeah, I think if he persisted, he would have found the same line.

Awesome tries everyone! A couple people were very close and mangler got it!

RT, that's exactly how I felt when I was shown the solution

To me, the most fun comes from the actual process of trying to solve this, not the solution itself.

So naturally you begin with the obvious fxg6+. You discount Qxg6 at first sight, because hey, Bc2. You check Kxh6, you see that it's winning. You check Kg8, you see that it's easily winning too. You start questioning your sanity. Kh8? No way? Just for the sake of it you check Kh8 and feel a little bit ashamed. Back to square 1 - why is this such a hard etude if white wins in a couple moves in all the lines? You realize that you are looking for something as black. There is nothing in Kh6 and Kg8 lines.. ok, let's put that black queen on g6 for a bit.. Ooh, Rf4+! Maybe it is something. OK, but Ke1 just wins easily! Aaaaah, Re4+. OK, but then Kg1 instead of Ke1 wins easily! Aaaah, Rf6.. God damnit.

So, you are sitting at the position after 1. fxg6 Qxg6 2. Bc2 Rxf4+ 3. Kg1 Rf6. You are three moves deep but it seems like you started ages ago. What is happening in this position? All the pawn endgames are drawn/lost. You triple check that and start going dizzy. Uuugh, I DON'T UNDERSTAND ANYTHING! We couldn't have made a mistake earlier, this HAS to be the position. (10 minutes later) OK so I think we have to escape to the queenside with the king, there is nothing else to do in this ****ING ****TY POSITION. But Black is always threatening Kxh6, we absolutely can't give it up. Hm, I guess 4. Qg5 looks good then. Okay, we are maybe making progress. 4.. Rxe6 seems the only move, 5.Kf1 Rf6+ 6. Ke1 Re6+ 7. Kd1 Rd6+ 8. Kc1 WE DID IT BABY. Oh wait... 8...Rc6... God ****ing damnit.

So now you feel with your entire heart that you are stumbling somewhere close. You start backtracking one move at a time and see that there is just nothing there. Nowhere. WHERE is that damn mistake? You have removed basically everything possible by process of elimination. And to quote Arthur Conan Doyle's Sherlock Holmes: "when you remove the impossible, whatever remains, however improbable, is the truth". And this is exactly when you have to come up with 1. Ba4 to force the b6 pawn to b5, so that in that long main line you have 10. Qc5 after Rc6. I don't think it is possible to solve this without agonizing over those other lines for at least some time

Anyway, props to everyone who have tried, some very interesting and fun lines.
Hope you enjoyed it!

Yeah I think on another day I see the ba4 idea first, so it's always in my mind that I can play the position with the pawn on b6 or b5. (it did cross my mind that maybe it needed to be there so some pawn ending is winning). Instead I spent time fruitlessly wondering how to get the WHITE pawn to b5

bc6 be8 nd5 nf4. Cute

whoops maybe that doesn't win. So bc6 kf7 nd5!

ok if black doesn't go f7 we get the first line, but the ending is winning. But then there is a nasty dual, so am I missing something...

the first try is correct, guess the dual on move 2 if black goes kf7 isn't that terrible

what's the purpose of the a-pawns?

I'm not usually a study purist, despite how it must sound here, but it's often fascinating to see a detail in a study why an irrelevant-looking piece had to be there. Here they don't seem to make a difference.

bleh they're unnecessary. I lift these from chess.com tactics trainer when I come across good ones. They way the trainer works is that Black makes a move and then it's your turn. Guess whoever added this tactic fudged with the original in order to make a7-a6 Black's first move

okay I'll lead a little: 1.Rf5+ Kg6 (1.-Ke6 2.Re5+ or 2.h3 win eventually) and now?

Inspired by Rei Ayanami's joke about chess with 16 queens on each side, I tried to run a game between engines (two Houdinis 1.5a) - with one king on each side because otherwise there would be no mate opportunity - but alas they kept making illegal moves, confusing queens for knights, trying to castle etc.

However, one of my attempts, when I left knights and rooks as they were, replacing only bishops and pawns by queens in the initial position, led to a interesting spot. (The beginning of the game was -4. Qdd4 Nf6 -3. Nc3 Qdd5 -2. Qcf4 Qee6 -1. Qee3 Qfd6 0. Nf3 O-O??, and White saw no tactics and proceeded with 1. Qfd3?! Qdxf4 and tried to make an illegal move - Qexf4, forgetting that the pieces on e3 and e6 were queens, not pawns, and the e3 queen was pinned, by which it terminated the game.)

1.Rf5+ Kg6 2.h4 Bf2 3.h5+ Kh6 4.Kf3 Bh4 5.Re5 e2 6.Rxe2 Kxh5 7.Rh2 Kg5 8.Rh1 and the bishop is trapped