Shortest-proof-game challenge

Baibikov down to 51.5 moves:

Down another three moves. This could be the shortest.

#1061
Shorter

Well, maybe now it's the shortest. How will we ever know for sure?

Congratulations. That's some impressive chess golfing!

By breaking down all the necessary black moves.

Ke8 needs 4, Qd8 needs 1, Bc8 needs 2, Bf8 needs 2, a7 needs 1, d7+e7 need 5+6, h7 needs 4.

With some more effort, one should be able to show that Nb8+Ng8 need 3+7, Ra8 needs 2, Rh8 needs 3, and g7 needs 8. That's already 48 altogether.

Has this position been golfed yet?

This is Troitzky's retrograde problem from 1915 shown at the top of https://www.janko.at/Retros/Masterworks/Part1.htm

where one must show that it's not possible to reach the position without g7-g5 as the last move.

Our champion here at making proof games shorter is tygxc, who always finds an improvement over any proof game I post, but he sometimes has a problem finding the first proof game for a new position. He can make me eat my words by posting proof games for #1049 and #1052.

#1063
Shorter

@tygxc just disproved that by swapping the roles of Rh8 and the promoted g7, to end up with Rh8 making only 2 moves (and g7 still making 8).

On the white side he saved a move by leaving Nb1 unmoved.

A nice side effect is that all 3 pieces in the final position, Rook, Bishop, and Knight, must be promoted in the shortest proof game!

Congratulations are in order again. Bravo!!!

Here's a slightly different Baibikov proof game with under promotions only

Let me know if you're working on a proof game for #1049. If no one is, I'll post my unshortened 74-move proof game tomorrow.

69 moves (might be shortened)

They say tomorrow never comes, but here it is. This is my proof game for #1049, ready for someone to shorten.

#1071
Shorter

this should be the shortest

Holy crap! Just when you thought there was nothing left to improve...

Instead of sending Ra8 to b5 in 2 moves, and the g7 pawn to promotion and on to a5 in 5+2 moves, this game sends Ra8 to a5 in 1 move and g7 to b5 in 5+2 moves.

And then somehow finds an inefficiency in white's moves as well... Bravo once again!

I agree this should be the shortest. In fact I would happily bet $100 on it...

#1077
This is shortest indeed. I got the idea from your underpromotions to rooks.

Here's a variant where the knights behave quite differently. In addition I tried to make long sequences of consecutive moves by the same piece, which I find aesthetically pleasing.

For a proof of optimality, one cannot improve on these move counts for pieces in the final position:

0 moves: b7@b7 c7@c7 b2@b2 c2@cs
1 moves: a7@b6 a2@b3 d2@c3
2 moves: Ra8@a6
4 moves: Ke8@a5 h7@h3
5 moves: e2@a7 Ke1@a3 e7@a2
6 moves: d7@b1B
9 moves: h2@b8N

and not on the total count of these captured pieces:

0 moves: f7@f7 Nb1@b1
1 moves: Qd1@d7 Bf1@b5 Ra1@a3 Qd8@d3
2 moves: Rh8@a6 Bc1@c5 Bc8@c4 Bg7@c3
3 moves: Ng1@c6
4 moves: Rh1@b6 Ng8@b3
6 moves: Nb8@g7
7 moves: g7@b5 g2@a4 f2@b4

Total: 43+50 = 93 = 2*47-1.