Clearing the board (VERY HARD)

Starting from the standard position, find the minimum number of moves it takes to clear the board so that only the kings remain. Can you do it in 17?

That sounds like a Sam Loyd puzzle. Here's an easier one to warm up with. With all 16 white units on the board in their starting positions, and no black units, find a square to place the black king so that White to move can force mate in 3.

Nice! I hadn't heard of Sam Loyd before and he has loads of cool puzzles, thanks a lot! He did attempt this problem but suprisingly a shorter solution than his exists!

This is the best I can do. :P

Specialists have tried to create a unique SPG (shortest proof game) leading to a position with only 2 kings but to my knowledge they have failed so far. What the current shortest (non-unique) proof game is I don't know but I doubt it is 17 moves.

In 1895 Sam Loyd published a 17-move solution, but his final position had many 17-move proof games, so it didn't meet the requirement of being a unique SPG. In 1994, Wilts & Geissler published a 17-move solution that wasn't unique, but was unique if Black's second move had to be a capture. Some unique 16.5 move solutions have been found requiring that the FIDE Dead Reckoning rule be suspended. As of June 2012, the best unique result with no conditions required 19.5 moves. Maybe since then a 17-move unique solution with no conditions has been found, and Chess_Dogg will soon show it to us.

Here is Sam Loyd's solution.

Massacre proof game in 17 moves (non-unique)

Nice, here was my 17 move solution.

I have a 16.5 move solution that does not require dead reckoning but I have no idea whether it's a uniqu final position. I'll post it in a few days but I'd be interested to see whether anyone else has another 17 move or 16.5 move solution.

@n9531l h4 seems to work for the checkmate. 1.d4 and then either 1...Kh5  2.Qd3 Kg4 3.Qh3# or 1...Kg4 2.e4+ Kh4 3.g3#. Cool puzzle and definitely trickier than I thought it would be!

Your (non-unique) game is very good. It may have been published previously by someone else; I'm not sure about that.

I do know of a non-unique 16.5 move game by G. Ponzetto (January 2000) with the white and black kings ending on f2 and e7. So far I haven't found any evidence of a unique solution with no conditions shorter than the 19.5 move game I mentioned. I'm looking forward to seeing your 16.5 move game.

Perhaps surprisingly, the only legal positions for the black king that don't give White a mate in 6 are a4(5), e4(7), f4(7), and h4(3). So you found the only solution.

@Arisktotle: As you would probably expect, François Labelle is the man to go to for computer-aided compositions. wKe5, bKe2, unique PG 19.5 (EDIT: I am dreaming) (François Labelle, StrateGems 58 April-June 2012)

EDIT: the one n9531l mentions.

Labelle calls his game a shortest proof game, and it looks like one to me. Why do you say not SPG?

...because I appear to be dreaming? Thanks.

Clearly my information was outdated. With the help of the unpleasant 'massacre' keyword I have now been able to find the current state of this subject.

Good to have you back, Remellion!

What is the current state of this subject? Is there a newer result than Labelle's 2012 game?

Wow, that SPG is really cool, I never knew so much research had been made into this topic. @n9531l I think my 16.5 is the same one you quoted (sadly). It would be an interesting problem to try to find shorter SPGs although I'm guessing it would be incredibly difficult without some code. Problems like this really shows just how diverse and complex a game like chess is!

Solving the 19.5 SPG is already kind of a headache. I've looked at it several times but can't even decide how to attack it. I don't know what it'd take to compose it without a computer-aided search.

nice! I prob can't do any better

@Remellion:  Short of a revelation from an omniscient being, I don't think a human can solve it without computer help. Clearly, Labelle had a great deal of computer help.

Not that I know of. My intention was to state that the use of the proper search term had uncovered the same data as relayed by the other posters, Remellion and you in particular. No news.