1. e6 hxg4!?* 2. exf7! g3 3. f8=B! g2 4. Bxc5 Ka1 5. Bd4+ Ka2 6. Kc1 g1=Q+ 7. Bxg1 Kb3 8. c5 a2 9. Bd4 +-

1... h4!?* 2. e7! h3 3. e8=B! h2 4. Bc6 Ka1 5. b4 a2** 6. Bh1 cxb4 7. g5 b3+ 8. Kxb3 Kb1 9. Be4+ Ka1 11. c5 h1=Q 12. Bxh1 Kb1 13. Be4+ Ka1 14. c6 fxg5 15. c7 +-

* Black sets a trap: if White queens or rooks instead of bishoping, then Black will be in stalemate after his queened pawn is taken. If 1... fxe6, then simply fxe6-e7-e8=Q as the f6 pawn will be able to move and there will be no stalemate.

** Black now doesn't want to play cxb4 and allow the c-pawn to queen, but White puts him in a zugzwang when it's Black's only legal move! Then White makes an anti-stalemate pawn sac on g5.

I've seen this before somewhere but can't remember the exact solution - but what if 6...Kxb3 in that line? looks like Black takes all the pawns

on the other hand, it's not necessary, because 5. b4 and White wins. But there is also another possibility to consider in the h4 line, I think, let me not give away anything more

contact checks (except for the clearance with e1Q) ftw 1.-n@d4+ 2.exd4 nxd4+ 2.kf1 p@e2+ 3.kg1 e1Q+ 4.nxe1 ne2+ 5.kf1 q@g1+ 6.rxg1 ng3+ 7.fxg3 p@e2+ 8.kf2 n@d1 mate

Wow, that's an awesome dual! I've missed the Ng3+, P@e2+, N@d1# finish! The problem should be reformulated as 'mate in 9' to make it the only solution.

My idea (after 4. Kg1) was 4... Nxf3+ 5. Ngxf3 Bxf3 threatening N@h3+ & Q@g2#, or e1=Q+ & N@e2#, or 6. Nxf3 Q@f1#, or 6. B(Q)@f1 exf1=Q+ 7. Kxf1 P@e2+. This solution is longer because its 5th move is not check and allows White to prolong the agony by dropping with checks.

There's one more question: how can Black mate in four if the king goes to d3 instead of f1? I see two twin ways to do so, one of them doesn't require dropping the second knight. That line is certainly very easy but cute enough to be mentioned imo.

I'm gonna' try writing out my full stream of consciousness on this one.

First thought is: a6 Ra1 d7, but our king on the 7th rank proves problematic after Rd1 a7 Rxd7 is check and the a-pawn falls (rather than queening as it often would with that pattern, otherwise).

There could be Bg1+ or Bf4+ deflection shenanigans, gotta' keep an eye out for those.

Does a6 Ra1 Bg1+ work? Idea being that if the king moves then a7 and the pawn is guarded, and the d-pawn has a chance to run because the rook is tied to the a-file. And if Rxg1 then maybe the pawns are free? Like, going back to the beginning, a6 Ra1 Bg1+ Rxg1 d7 Rd1 d8=Q (not a7 Rxd7+ of course) Rxd8 and that doesn't work because the rook stops the a6 pawn no problem. So instead, is there something to a6 Ra1 Bg1+ Rxg1 a7 Ra1? No, if d7 Rxa7 the d-pawn is pinned and falls.

So while I'm not forgetting about that idea, in some other more complex variation, it's not so simple as just immediately sacrificing the bishop. There is also maybe an idea of trying to get the rook to take on a7 (and it not be check), so that Bg1+ can win the rook, but I don't see a way to force it yet.

There's also a threat we probably can't just completely ignore of Rh1 pinning the bishop.

Let's see. Back to the top. Possible moves are a6, d7, king move, or bishop move. It seems d7 Rf7+ must be hopeless (Kg8/g6 Rxd7 and there's certainly no way to win - at the very least the rook will always be able to sac itself for the a-pawn and the bishop can't mate on its own. And black may have winning chances there as well, but for now I'm still assuming it's white to play and win (not white to play and draw) so I won't even worry about that part.)

1. a6 Rh1 doesn't seem to work for black (2. a7 Rxh2+ 3. Kg6 Rh8 {3. ... Ra2 4. d7 Rxa7 5. d8=Q} 4. d7 and actually now I'm not so sure white does win this... neither pawn can queen immediately and I don't think the king can drive the rook off the back rank. Can the king maneuver to e7 and force a trade on d8, then queen on a8? Yeah, I think so. Can't quite calculate it with complete certainty, but I *think* so. So the point is 1. ... Rh1 does appear to fail.)

Therefore 1. a6 Ra1 is forced. 2. Kg8? Maybe? What happens there... 2. ... Rxa6 3. d7 Ra8+ and I don't think we have anything. The only idea would be to try to squeeze the bishop in to the back rank somewhere to provide interference and let the pawn queen, but that only makes sense if it's a light-squared bishop that could get to c8 I think...

Gah, I have no idea. And this, gentlemen, is what it looks like when a Class C player completely does not understand a position, and bangs his head against it for 15 minutes. Hopefully it was amusing, since it sure wasn't successful.

No that is not bad at all BJJ, very good job! That was pretty much my thought process too and and this point you can only make progress if you see the driving idea of this etude, which can take any amount of time really, took me a whiiiiiiiile. Also, even though Ra1 would be forced after a6, you have to consider Rf7 and Rf8, too

I was looking at 1.Bg1+ but couldn't figure out how's it better than 1.a6 yet. maybe it's not the solution after all

Gah! I was so close!

3. Kh8! And black can't gain a tempo

I find the efficiency so beautiful in this. No matter which file black rook tries to use to get to the 8 rank, he is cut off by white, either by a7 or d7 or the king. Pretty cool Also, there is a fun sideline: 1. a6 Rf8 2. a7 Ra8 3. d7 Rd8 4. Bc7!

1. Nc5

1. - Qg6
2. Ne6+
2. - Ke8 3. Bh5
2. - K somewhere else 3.Nf8+

1. - Qb5.
2. Ne6+ followed by Nd4 after anything but Ke8 which allows Nc7+.

1.. Qc6 should be the main line I think. So far I can't find it..

Hm, if this is it it's quite nice indeed

So after the initial 1. Nc5 Qc6 2. Ne6+ Ke8, black is actually almost in zugzwang. There is just one square where black queen can go to without getting lost on the next move, and that square is d6. So I think we have to play 3. e5! and now if 3.. h6/h5 we just take gxh6 and it doesn't change anything and if 3.. a6 then maybe 4. Kh2 (gotta cover that h1 square from the black queen, hence h2) and I think it's gg.

Found exact same line as YKW

Yeah I didnt bother giving the obvious lines

move 438 isn't as obvious as it looks

This problem came from chess.com's tactics trainer.

1. Rg8+ Kf5 2. Rg5+ Ke6 3. Re5+ Kd6 4. Bg5!! (the point) d1=Q 5. Bf4 Qg4+ 6. Bg3 Qb4 7. Re4+ resigns, 1-0

Bg5 is really a great move.

I quickly got to the position after Kd6 and felt like this has to be it - the whole setup of pawns and all, Bg5 was not that hard to find after that