How many distinct chess games are possible, and which is the longest?

Sort:
SeanEnglish

Kilonewton, heres a way to get there on the first move of move 18:

it seems like there might be a faster way(especially since white has to make that repeat with the knights to get everyone back home) but I cant see it.

SeanEnglish

ahh, nevermind, here's a way to do it in less:

here it ends on the first move of  move 16

So 16 is an upperbound... for a lower bound, 8 pawns have to be captured, so that takes at least 8 moves. The first move, no capturing can happen, so thats 9 moves. It is obvious that a piece must be used, and when they return to their original spot, that won't be a capture, so at least 10 moves. It's rather obvious that it will take more than 10, but not easy to prove, so I'm going to stick with 10

So the shortest way is somewhere between 10 and 16. I think it would be hard to get more precise then that. 

ButWhereIsTheHorse

Hehe !

SeanEnglish

ahh, heres a way to increase the lower bound:

using only a knight would take forever(obviously), and using a bishop would require a lot of pawn moves to put them all on one color. Using a rook, they can't capture the kings pawn without causing check, which adds at least one move, so if using a rook, two pieces are necessary, or a few extra pawn move an neessary, Using a queen, there are three pawns that being taken in standard position cause check, so that would create extra problems. So, it seems that you must use two or make extra pawn moves. The second piece must also return home, which won't be a pawn capture, so at least 11 moves. Furthermore, there is no way that both pieces first move will be a capture without a bunch of extra pawn moves, so at least 12 moves.

So between 12 and 16 moves.

I defnintely can't see a way to do it in less than 16 though. 

kiloNewton

hehe, its an exciting problem because none can use engine here Laughing

SeanEnglish

yeah, I mean, you definitely could use a computer here, but you would have to completely reprogram it, but 16 moves deep is within current computational power.

kiloNewton
SeanEnglish wrote:

 16 moves deep is within current computational power.

reachable to 16 moves if you use top chess engines in std chess. but if you program from scratch (and you have to, because this problem has completely different rules/objective than std chess) its a difficult task.

top engines can reach 16 moves because they can wipe out weak moves (which are more than 80% of possible legal moves)

 

 

so, its true its programable for this specific problem, but tremendous mental work have to be done there. (certainly no-one is going to invest that amount of time)

 

thats why, we can say that, solution posted here will be genuine. what you can't expect in normal chess puzzles.

SeanEnglish

Yeah, I don't think anyones going to reinvent chess programming in order to post to this forum...

Btw, did you have a better solution than mine? 

kiloNewton

Queens Tour, 14 moves



SeanEnglish

ah, very nice, I thought of using a queen, but the idea of checking the king seemed foolish.

Funny thing though, the position you reached and the position I reached are two different ones, and there are two other distinct positions that also look idential to our two... I wonder how the minimum number of moves to reach each one differs... 

kiloNewton

9 moves if illegal moves allowed (?!...Cool)

8. Rxh2 Rxh7  9. Rh1 Rh8      last two can't be inserted in figure - chess.com doesn't allow!! however it allowed own pawn capture Tongue OutLaughing
SeanEnglish

two moves, one move to rake your hand across the board and grab all the white pawns, and second move to rake your hand across the board and grab all the black ones... unless using two hands counts as one move(because I just see no way to grab all 16 paws with one swipe of a hand)

kiloNewton

one move by two hands! seen that in clubs - some use two hands in castling.

however for convenience you can make a steel structue like this:

please, dont go MAD before pulling.

shell_knight
SeanEnglish wrote:

two moves, one move to rake your hand across the board and grab all the white pawns, and second move to rake your hand across the board and grab all the black ones... unless using two hands counts as one move(because I just see no way to grab all 16 paws with one swipe of a hand)

Try using a smaller set :p

SeanEnglish

does it have to be steel? could I use a lighter material, or would that be insufficent for my One-Move Pawn ScraperTM ?

I suppose some lighter materials would break if you were using pure osmium pieces... 

SeanEnglish

Shell_knight, I think you're on to something... if you use one of those cheap magnetic folding travel sets, I bet you could do it in one move!

kiloNewton
SeanEnglish wrote:

Shell_knight, I think you're on to something... if you use one of those cheap magnetic folding travel sets, I bet you could do it in one move!

you need amazing craftmanship to do that, as you cannot displace other pieces.

SeanEnglish

I never said it would be easy, merely possible.

shell_knight

Make pawns out of e.g. solid nitrogen, pieces out of regular stuff.  Then set the board in a regular room (temp and pressure) i.e. you can remove the pawns with no moves.

SeanEnglish

ahhh, good thinking Shell_knight. If you require that all pieces be made of the same material, with the exception of differences required to make two different colors, could you still get away with no moves?