Best UnderPromotion Endgame Compilation

Very good one and very difficult but it sort of "goes out with a whimper rather than a bang". I have a feeling the composer wanted more of this composition but couldn't get there. But he did get the already awesome task of AUW!

Nah I love that one, I don't care if these don't directly end in mate, but just an obviously winning position instead. I'm surprised there are actually so many examples of bishop promotions. The last bishop promotion example can be summarized as this:

To avoid stalemate but simultaneously defend be4 mate at the same time. And the white pawn has to promote because otherwise ra1+ defelcting the rook off the 4th rank and the be4 and black wins. The additional underpromotions in the original example make it even better.

Why 1...Be4+?

I assume because white can position his rook to defend checking squares and use his bishop to gobble up the black pawns eventually, but good point, I'm not sure. Maybe at some point white can force the king away from f2 and free his king from any checkmate threats, but good point.

This example of why it's important to master the 3 knights endgame.

But Black's not playing for White. Seems to just lose a bishop.

Not at all obvious that 1.a8=B doesn't draw as well as 1.a8=Q. I managed to reach a TB draw against SF15.1.

You're right, stockfish says black should play bf7 and letting it play itself it exchanges down to a drawn position. I keep forgetting that every single piece alters the position, just go by the OP, not my simplification, was just trying to explain the general concept.

Another amazing underpromotion to each of the 3 pieces puzzle. However, there might be a conflicting line where stockfish can hold a draw, I'm not sure exactly what moves it occurs at though.

You and @MARattigan have been cheating on this endgame. What happened to the knight on b8?

Which endgame? I was talking about the one in #46.

When I looked there was no b8 knight.

Just looked again - still no b8 knight. (Wish the bl**dy thing would go away.)

As far as I know this famous study by Liburkin is correct. But I am not an expert on archives and histories

Supposedly there's a stockfish line which shows that at some point black can reach a position which the tablebase shows is a draw, but have to look it up again, I just remembered the position, not all the variations.

Here is a variation derived from a different study with a surprising outcome! Quite a disappointment when you have just mastered the 3 knights endgame

.

#46 was an incorrect quote from #44.

I know that stalemate pattern with 3 knights, but it also depends on what the alternative is. In my initial example, if white didn't check the black king, the black queen would hit the king on the back rank forcing white to give up one of the knights. King and 2 knights and a queen vs a king and queen may be a win depending on the position, but definitely not king and queen and 1 knight vs a king and queen. A more solid example is one where the alternative is getting checkmated next move:

Anything other than a knight, white loses here.

Actually, my version has the same result - just as in the original composition where the black queen is a rook. After the pawn promotes to queen, black can sac his queen resulting in stalemate! Not the only way to a draw though!

Don't know about that. (You can always check that kind of assertion on the Syzygy site.)

So KQNvKQ also depends on the position. (Yours is indeed a draw but your "proof" is not adequate.)

Here is one of my favorite under-promotions, the "Excelsior" by Sam Loyd. This one ends with the under-promoted piece delivering the mate!:

But white can promote to a queen in that one, so an underpromotion isn't necessary. By underpromotion endgame puzzles I mean where underpromoting is necessary, and promoting a queen either draws or loses the game when it could be won or drawn, respectively.

Here you are then, 4 under-promotions that lead to a FORCED mate in 6: