checkmate in 3

Originally posted by hedonist
ah, smokescreen.

mephisto has left the building...

I'm still here ... There wasn't anything I could say until at least someone mentioned the draw after Qf1+. As usual, SG has done more than just 'mention' it.

As SG has proven, the mate-in-two fails because after Qf1+ black can claim draw.

Even then, it is amazing how Rxg6! (the only move!) solves the problem. Especially since after Rg8+ RxR black has no solution against the mate-in-one Qf1+. Bij taking on g6, the rook on g3 uncovered the bishop that now covers pawn h2.

OK, I can see that, and it is very clever, but what I don't understand about puzzles like this is how you can be sure it is mate in three. I always thought it was a succession of moves that opponent had to follow. Here though, it is dependent on black whether it is mate in 2 or 3...
If black decides that suicide is the best way forward (and to get the board into this position, they would both have tendencies that way), what is to stop black doing:

Rxg6, RxR,
Qf1++

and stopping the mate in 3 criteria?

Originally posted by morgski
OK, I can see that, and it is very clever, but what I don't understand about puzzles like this is how you can be sure it is mate in three. I always thought it was a succession of moves that opponent had to follow. Here though, it is dependent on black whether it is mate in 2 or 3...
If black decides that suicide is the best way forward (and to get the boa ...[text shortened]... way), what is to stop black doing:

Rxg6, RxR,
Qf1++

and stopping the mate in 3 criteria?

There are four phases in this puzzle.

The first phase is that the 'obvious' mate-in two solution as opposed to the posted 'mate-in-three' should lead to the conclusion that there is something in the posted position that needs to be analysed (and not the belief that the puzzle itself is flawed). Often it is en passant, promotion or underpromotion, castling, ... that makes the position special, in this case it is the 50-moves rule.

The second phase is to find the fastest way backwards from the posted puzzle position to the critical position (which is the position in the game after the last pawn move or capture) and AND to prove that the fastest way to reach the final position takes 49.5 half moves. The only criterion is that the moves must be legal, not wether they are good or bad from quality point of view. The critical position and the moves backward from the posted puzzle position have been posted by SG.

The third phase is to prove that you can reach the 'critical' position in a legal chessgame, again regardless of the quality of the moves. In this case it is pretty obvious that such games can easily be constructed.

The fourth and final phase is the puzzle as was posted, mate in three against the best possible defense (as in a standard puzzle) but with the added information obtained by the previous phases that there is no room for a first move without capture or pawn move. In this case, it is a relatively easy solution.

Originally posted by hedonist
The problem with problems is....

they always look like their set up wrong 🙄

Ok, thanks.

Sorry for doubting you guys.