lubaza wrote:
Very well, and most of it in line with my solution. But it is not the shortest possible.
Improved by 1 move, but I’m unsure if white can get the pieces in time before black has to waste a tempo with Ra8-d8. 109 moves:
Shortest-proof-game challenge
Sort:
BishopTakesH7 wrote:
Reposting this one.
Could probably be shortened, but a try:
daStrwbrry wrote:
BishopTakesH7 wrote:
Reposting this one.
Could probably be shortened, but a try:
Mar 8, 2023
0
#3605
daStrwbrry wrote:
lubaza wrote:
Very well, and most of it in line with my solution. But it is not the shortest possible.
Improved by 1 move, but I’m unsure if white can get the pieces in time before black has to waste a tempo with Ra8-d8. 109 moves:
Impressive. 109 is also my best. I said we had very much the same ideas, but I realize now the differences are there, and quite interesting.
First, there are two ways to let the two Kings swap positions over the g3-area, you use the one with WK over f2-g3 while BK is on h1. But you can also do the opposite, as in my example. The advantage is that it requires one move less for Black since WK can step on h3, disadvantage is that the WK needs more moves and so less freedom for White to do other maneuvers.
Secondly I let the BK have an extra move on its way from e8 to g3, by going via h4. That brings me down to only two moves for the WB from g4 to f1.
There's also a question about which N should pass first over g6 before playing g7-g6. It will cost a move or two for one of the sides, so it's a matter about where to put that burden. If these ideas can be combined with yours and lower the limit to 108, I'd be happy to see how.
BishopTakesH7 wrote:
A random position I just made. (#3606)
Here's a legality proof. How much shorter can it be?
daStrwbrry wrote:
lubaza wrote:
Here is the example that brought me to this thread. It's a slight modification of a retro chess puzzle you might know and which I really liked. I am pretty sure I have found the length of the SPG games, but I leave it to you first.
That position reminds me of one I posted before in another thread (posts #6521 and #6522 there), although it is reflected and the black rook does not give check. It has to do something with the 75-move rule, although I have yet to make a proof game of it.
Here is an SPG for that, minus the 75 moves after: https://pdb.dieschwalbe.de/P0008399
Mar 9, 2023
0
#3609
RewanDemontay wrote:
daStrwbrry wrote:
That position reminds me of one I posted before in another thread (posts #6521 and #6522 there), although it is reflected and the black rook does not give check. It has to do something with the 75-move rule, although I have yet to make a proof game of it.
Here is an SPG for that, minus the 75 moves after: https://pdb.dieschwalbe.de/P0008399
That is not the shortest for the given position. First of all, the capture h3xBg4 happens when the WK is at c8 and BR at e7 if you want to only spend 75 moves afterwards. That happens after move 36 in your example, move 37 is just a waste. But it can happen already at move 32:
The placements of BKc2 and BBa7 compared to your example does not mean anything, there is plenty of time for Black to rearrange those pieces while the WR moves to f8. This game continues with the well-known 75 moves to give the position of the original puzzle in 107.5 moves. I claim this is an SPG, argument will follow.
Mar 9, 2023
0
#3611
So what is the fastest way to reach the position at https://pdb.dieschwalbe.de/P0008399? Playing the game backwards (preferred when analyzing retro puzzles), the 75 "first" moves are well-known and lead to the same as the final position, except with the WK on c8, WR on h1, WB on b4, BB on a7 or b8, and BK somewhere in its reachable area. We could either uncapture h3xBg4 now, or after 76 moves with WK on b7 and BR h8, or after 77 moves with WK on a6 and BR on h8 (the last two are preferred since the BR ends up at home). In either case we are able to proceed to get the following position after 81 retraction moves where the last was the retraction of b7-b6, and with White to make next retraction.
It is possible to have the BK at a1 and WB at a5 instead if that is preferred.
The minimum number of retraction moves needed to get the pieces home from here is:
White: Ra1=2 (uncaptured at b6/c5), Nb1+Ng1=7 (2 to b1, 5 to g1 no matter which N goes where), Bc1=2 (only if going a5/b4-d2-c1), Qd1=2 (uncaptured at b6/c5), Pa2=2, Pc2=1, Pd2=1, in total 17 + King moves from b5 to e1.
Black: Ra8=2 (uncapture somewhere), Nb8+g8=7 (uncapture at b3/c4), Qd8=2 (uncapture e3/c4), Bf8=4 (a7-b6-a5-b4-f8, visit to a5 necessary to retract Pc5), Pa7=2, Pe7=1, in total 18 + King moves from a1/c2 to e8.
From here, White is to make one more retraction move than Black.
First assume the BK moves over a4. There are basically 2 possible King arrangements from here.
- WKb5-a4-a3-a2-a1 while BK waits at c2. With no castling, White/Black needs 8/7 King moves, in total 25 for both. But either Na8-b6 goes first, then White waits two moves and is up to 27, or Na4-b6 goes first, Black waits one move and is up to 26, so White here also is up to 27. White will save one move by castling, but then WR must go c5-d5-d1, and in all my attempts it rests too long on d5 denying BK access to the 5. rank.
- WKb5-a4-b3-c2-d2 while BK waits at a1. Black needs 8 King moves, totally 26, and again 27 for White.
So 27 White moves are needed, and that gives a total of 107.5 moves for the entire puzzle.
Now, BK would save a move by going b3-b4-c5-d6-e7-e8. But from all my attempts, that requires so much to be arranged for the WB and WN to get away in time, that I quickly run out of moves. But if anyone can prove me wrong, I would be happy to see it.
EvinSung wrote:
#3619
EvinSung wrote:
#3617
Legality proof.
n9531l1 wrote:
EvinSung wrote:
#3617
Legality proof.
Here’s shortest, if I counted correctly.
lubaza wrote:
So what is the fastest way to reach the position at https://pdb.dieschwalbe.de/P0008399? Playing the game backwards (preferred when analyzing retro puzzles), the 75 "first" moves are well-known and lead to the same as the final position, except with the WK on c8, WR on h1, WB on b4, BB on a7 or b8, and BK somewhere in its reachable area. We could either uncapture h3xBg4 now, or after 76 moves with WK on b7 and BR h8, or after 77 moves with WK on a6 and BR on h8 (the last two are preferred since the BR ends up at home). In either case we are able to proceed to get the following position after 81 retraction moves where the last was the retraction of b7-b6, and with White to make next retraction.
It is possible to have the BK at a1 and WB at a5 instead if that is preferred.
The minimum number of retraction moves needed to get the pieces home from here is:
White: Ra1=2 (uncaptured at b6/c5), Nb1+Ng1=7 (2 to b1, 5 to g1 no matter which N goes where), Bc1=2 (only if going a5/b4-d2-c1), Qd1=2 (uncaptured at b6/c5), Pa2=2, Pc2=1, Pd2=1, in total 17 + King moves from b5 to e1.
Black: Ra8=2 (uncapture somewhere), Nb8+g8=7 (uncapture at b3/c4), Qd8=2 (uncapture e3/c4), Bf8=4 (a7-b6-a5-b4-f8, visit to a5 necessary to retract Pc5), Pa7=2, Pe7=1, in total 18 + King moves from a1/c2 to e8.
From here, White is to make one more retraction move than Black.
First assume the BK moves over a4. There are basically 2 possible King arrangements from here.
- WKb5-a4-a3-a2-a1 while BK waits at c2. With no castling, White/Black needs 8/7 King moves, in total 25 for both. But either Na8-b6 goes first, then White waits two moves and is up to 27, or Na4-b6 goes first, Black waits one move and is up to 26, so White here also is up to 27. White will save one move by castling, but then WR must go c5-d5-d1, and in all my attempts it rests too long on d5 denying BK access to the 5. rank.
- WKb5-a4-b3-c2-d2 while BK waits at a1. Black needs 8 King moves, totally 26, and again 27 for White.
So 27 White moves are needed, and that gives a total of 107.5 moves for the entire puzzle.
Now, BK would save a move by going b3-b4-c5-d6-e7-e8. But from all my attempts, that requires so much to be arranged for the WB and WN to get away in time, that I quickly run out of moves. But if anyone can prove me wrong, I would be happy to see it.
Excellent analysis.
lubaza wrote:
So what is the fastest way to reach the position at https://pdb.dieschwalbe.de/P0008399? Playing the game backwards (preferred when analyzing retro puzzles), the 75 "first" moves are well-known and lead to the same as the final position, except with the WK on c8, WR on h1, WB on b4, BB on a7 or b8, and BK somewhere in its reachable area. We could either uncapture h3xBg4 now, or after 76 moves with WK on b7 and BR h8, or after 77 moves with WK on a6 and BR on h8 (the last two are preferred since the BR ends up at home). In either case we are able to proceed to get the following position after 81 retraction moves where the last was the retraction of b7-b6, and with White to make next retraction.
It is possible to have the BK at a1 and WB at a5 instead if that is preferred.
The minimum number of retraction moves needed to get the pieces home from here is:
White: Ra1=2 (uncaptured at b6/c5), Nb1+Ng1=7 (2 to b1, 5 to g1 no matter which N goes where), Bc1=2 (only if going a5/b4-d2-c1), Qd1=2 (uncaptured at b6/c5), Pa2=2, Pc2=1, Pd2=1, in total 17 + King moves from b5 to e1.
Black: Ra8=2 (uncapture somewhere), Nb8+g8=7 (uncapture at b3/c4), Qd8=2 (uncapture e3/c4), Bf8=4 (a7-b6-a5-b4-f8, visit to a5 necessary to retract Pc5), Pa7=2, Pe7=1, in total 18 + King moves from a1/c2 to e8.
From here, White is to make one more retraction move than Black.
First assume the BK moves over a4. There are basically 2 possible King arrangements from here.
- WKb5-a4-a3-a2-a1 while BK waits at c2. With no castling, White/Black needs 8/7 King moves, in total 25 for both. But either Na8-b6 goes first, then White waits two moves and is up to 27, or Na4-b6 goes first, Black waits one move and is up to 26, so White here also is up to 27. White will save one move by castling, but then WR must go c5-d5-d1, and in all my attempts it rests too long on d5 denying BK access to the 5. rank.
- WKb5-a4-b3-c2-d2 while BK waits at a1. Black needs 8 King moves, totally 26, and again 27 for White.
So 27 White moves are needed, and that gives a total of 107.5 moves for the entire puzzle.
Now, BK would save a move by going b3-b4-c5-d6-e7-e8. But from all my attempts, that requires so much to be arranged for the WB and WN to get away in time, that I quickly run out of moves. But if anyone can prove me wrong, I would be happy to see it.
Here is the 107.5.
Mar 10, 2023
0
#3619
A random position I just made.