< Earlier Kibitzing · PAGE 95 OF 97 ·
Later Kibitzing> |
Jun-05-25 | | stone free or die: Ah <beat>, that's a little too bad and a little surprising it's not sufficient. C'est la vie. (I admit I just relayed it over as I was formatting my post, and didn't really look into it.) |
|
Jun-05-25
 | | beatgiant: <stone free or die> The counting problem looks tricky, and I haven't actually tried solving it myself yet. In fact, I had overlooked the solutions with intervening check when I said it was <a lot more than just one> above. |
|
Jun-06-25 | | stone free or die: <beat> here's the solution Laver gave, with Akihiro Kanamori's original solve: <
1.f3 Nf6 2.e4
({Kanamori} 2.Kf2 Nh5 3.d3 Ng3 4.Be3 a6 5.Qe1 Nxh1# )
2...Nxe4 3.Qe2 Ng3 4.Qxe7+ Qxe7+ 5.Kf2 Nxh1# *
>
Clearly there are many permutations, and even superfluous waiting moves (4...a6 in Kanamori's solve). I think adding the condition that both sides check each other before the final mate restricts the solve quite a bit. But I'm not sure it's so interesting, beyond being a counting problem. You have the White pawn move permutation, and the various paths the Black knight can take to h1. . |
|
Jun-06-25
 | | beatgiant: <stone free or die> Now that you've posted the solutions, I can post about the counting problem without spoiling it. With intervening check, the only solutions I found were the one given, and the same with White's first two moves reversed. I agree that's boring, except for the cleverness of the solution itself (like I said, I had missed that until you said it was possible). The other set of solutions is a bit tricky to count. White has two series of moves (f3, Kf2, Qe1) and (d3, Be3) that can be interleaved, and black has many choices for the throwaway move, which can be done at different turns, and some of which would block the knight's path and so can only be played after it has passed. That seems like a pretty good exercise in combinatorial counting, not something one could do quickly in one's head. |
|
Jun-06-25
 | | beatgiant: The other tricky thing about the counting problem is the dependency between the king and the knight (if the knight is taking the route via e4, White has to play Kf2 only after it has passed, and that also affects the timing of Black's throwaway move). |
|
Jun-06-25 | | stone free or die: <beat> don't forget White has the other option of blocking e3 with the QB: < 1.f3 Nf6 2.d3 Nh5 3.Be3 Ng3 4.Kf2 e5 5.Qe1 Nxh1# >  click for larger viewI'm being trying to see if I could do it with the QN on the Black side, but the best I've gotten is this M-3: <1.e3 Nc6 2.Ne2 Ne5 3.g3 Nf3#>  click for larger viewCould have finished with <3.c3 Nd3#> too. |
|
Jun-06-25
 | | beatgiant: <stone free or die> <don't forget>, not sure what you mean by that, my post above did mention Be3. The QN can reach h1 in 5 moves, for example: 1. f3 Nc6 2. Kf2 Nd4 3. Qe1 Nf5 4. d3 Ng3 5. Be3 Nxh1# |
|
Jun-06-25 | | stone free or die: Ah sorry. I'm the one who apparently forgot. |
|
Jun-06-25 | | stone free or die: (as in - forgot to read!) |
|
Jun-06-25 | | stone free or die: I like the QN solve. |
|
Jun-06-25
 | | beatgiant: I may post an attempted answer to the counting problem this weekend, unless someone beats me to it. Although after missing the solutions with the checks before, I can't be as confident that I've actually found all the solutions.... |
|
Jun-07-25 | | stone free or die: <<beat> Although after missing the solutions with the checks before, I can't be as confident that I've actually found all the solutions....> Ironically, learning of those solutions actually increased the probability you've found all the solutions. |
|
Jun-09-25
 | | beatgiant: Now I'll start sharing my findings about a game ending in 5...NxR mate. The hardest part is showing that we've found the complete set of solutions. Although I'm not sure we have, I'll make the case by posting a series of claims about possible solutions. Some of these claims are ironclad, others more like "there's no such solution because I couldn't find one after trying hard" so I won't be shocked if I got some of them wrong. So it's a new side puzzle for other kibitzers to try to refute these claims. If nobody refutes these, I'll later show how I count the solutions we found. |
|
Jun-09-25 | | stone free or die: (You don't really need my concurrence, but "sounds reasonable") |
|
Jun-09-25
 | | beatgiant: Continuing discussion on the problem "game ending in 5...NxR mate." Claim: The rook has to be captured on its original square. It would take at least two moves to develop the rook to a different square. If the king also has to be moved, that's another two moves and there won't be enough
time to set up conditions for the mate.
So we have to consider cases with the rook on d3 or f3, which takes three moves to reach. We also need to prevent one of White's pawns from recapturing the rook, which can be done either by moving them or pinning them. With the king on e1, only the e-pawn can potentially be pinned but it takes at least three moves to get Black's queen in position for that, because White is too busy to set up a Qxe7+ Qxe7 type exchange here. If Black has to play three knight moves to reach d3 or f3 and three moves to get the queen to the e-file, that's already more than five moves. Or with the pawn moves, it's three moves to get white's rook to d3 or f3 and two to move the pawns protecting the rook, but the rook would still be guarded by another White piece and it would take a sixth move to prevent that. Examples of failed attempts: 1. a4 Nc6 2. Ra3 Ne5 3. Rd3 d6 4. c3 Qd7 5. h3 Qe6, and now Black needs one more move to play 6...Nxd3#. Or 1. a4 Nc6 2. Ra3 Ne5 3. Rf3 Nf6 4. g3 Nd5 5. e3 Nc3, and now White needs one more knight move to unprotect the rook and allow 6...Nxf3#. |
|
Jun-09-25
 | | beatgiant: Still on "game ending in 5...NxR mate."
Claim: The mate has to take place on h1 and not on a1. It takes white four moves to clear a path to c2 and move the king there, or five to reach b3, and that doesn't leave enough time to cover the squares to set up a mate. Examples of failed attempts: 1. d3 Nc6 2. c3 Na5 3. Kd2 Nb3 4. Kc2 g6 5. Nf3 when White needs one more move to play 6. Nd2, allowing 6...Nxa1#. Or 1. d4 d5 2. Kd2 Nc6 3. c3 Nxd4 4. e4 Nb3+ 5. Kc2 dxe4, and Black needs one more move to play 6...Nxa1#. Or 1. d3 Nc6 2. c4 Nd4 3. Kd2 Nc2 4. Kc3 c5 5. Kb3 Qa5, and Black needs one more move to play 6...Nxa1#. |
|
Jun-09-25
 | | beatgiant: "Game ending in 5...NxR mate" continued.
Claim: The mate must be given by an original knight and not a pawn promoted to knight. It would take 5 moves to promote a pawn. Assuming we stretch the problem statement to allow the last move to be 5...PxR=N mate, the last move must be 5...g2xh1=N, but that cannot give mate because White would have 6. Kg2. Example of failed attempt: 1. f3 h5 2. Kf2 h4 3. Qe1 h3 4. d3 hxg2 5. Be3 gxh1=N+, which isn't mate because of 6. Kg2. |
|
Jun-09-25
 | | beatgiant: Further on "game ending in 5...NxR mate."
Claim: The white king can only be mated on f2 and not on g3. It takes White three moves to clear a path to g3 and move the king there, leaving two spare white moves plus one spare black move, and that doesn't allow enough time to cover all the squares around g3. Examples of failed attempts: 1. f3 Nf6 2. Kf2 Ng4+ 3. Kg3 Nf2 4. Nh3 g5 5. e3 Nh1+. There wasn't enough time to cover g4. Or 1. f3 Nf6 2. Kf2 Ng4+ 3. Kg3 Nf2 4. d4 d5 5. Bf4 Nxh1+. There wasn't enough time to cover h4. Or 1. Nf3 Nf6 2. Nh4 d5 3. f3 e5 4. Kf2 Ng4+ 5. Kg3 Nf2 and Black needs one more move to play 6...Nxh1#. |
|
Jun-09-25
 | | beatgiant: "Game ending in 5...NxR mate" continued.
Claim: If the QN gives the mate, its only possible path is Nc6-Nd4-Nf5-Ng3-Nh1#. If the king is on f2 as we claimed above, the last move has to be Ng3xh1#. The knight can't reach g3 via e2 because that would remove the e-pawn and break the mate. It can't go via e4 because of interference with White's king on
f2. It can't go via h5 because the pawn on d7 blocks the 3-move path there. So only f5 remains, and the only path there within the
available time is the one claimed.
Examples of failed attempts: 1. f3 Nc6 2. Kf2 Nd4 3. d3 Nxe2 4. Be3 Ng3 5. Qe1 Nxh1+ is not mate. Or 1. d3 Na6 2. Be3 Nc5 3. f3 Ne4 blocks 4. Kf2. Or 1. f3 Na6 2. Kf2 Nc5 3. d3 Ne4+ and now White has to lose moves responding to the check. Or 1. f3 d6 2. Kf2 Nd7 3. Qe1 Ndf6 4. d3 Nh5 5. Be3 Ng3, and
Black needs one more move to play 6...Nxh1#. |
|
Jun-09-25
 | | beatgiant: "Game ending in 5...NxR mate" continued.
Claim: If the KN gives the mate, its only possible paths are Nf6-e4-g3-h1, Nf6-h5-g3-h1, Nh6-f5-g3-h1. As before, the last move has to be Ng3xh1#. To reach g3 in three moves, the knight has to come via e4, f5 or h5. Each of those is reachable in two moves only along the paths listed. |
|
Jun-09-25
 | | beatgiant: "Game ending in 5...NxR mate" continued.
Claim: Our known solutions with the queen checks are the only ones that don't require White to play Be3 to block this escape square for White's king. Black has at most one spare move and that move will have to cover at least the e3 square unless White is doing so, and ...Qxe7 is the only way to do that in just one move. Alternate solutions for accomplishing this all take more moves than our known ones. Examples of failed attempts: 1. d4 Nf6 2. e4 Nxe4 3. Bg5 Ng3 4. f3 h6 5. Bxe7 Qxe7+ and it takes one more move to play 6.Kf2 Nxh1#. Or 1. f3 Nc6 2. Kf2 Nd4 3. Qe1 Nxe2 and now White isn't able to play 4. Qxe7+ yet so the mate will again take a move longer. Or 1. d4 Nh6 2. Bg5 Nf5 3. Bxe7 Ng3 4. f3 Qxe7 5. Kf2 and it will take another move for White to play 6. Qe1, allowing 6...Nxh1#. |
|
Jun-09-25
 | | beatgiant: I'll probably revisit here in about a week to see if anyone has posted counterexamples or refutations of any of my claims about "game ending in 5...NxR mate" and based on what we've learned, post my accounting. |
|
Jun-13-25
 | | beatgiant: I found a counterexample for my first claim above: 1. a4 Nc6 2. Ra3 Nd4 3. Rd3 Nxe2 4. c3 Nf4 5. Ne2 Nxd3#. The key point I previously missed is we can save a move for White by having the knight capture the e-pawn. I think solutions with capturing the c-pawn or g-pawn still won't work because that would be check. So we have to consider solutions with a rook not captured in the corner. This is already my second oversight on this problem, so the invitation is still open for anyone to break any of the other claims above. |
|
Jun-16-25
 | | WannaBe: Saw this one online, so...
I have two ropes, different length, material, thickness, (but the ropes do burn...) and it takes an hour/60m to burn the ropes. With this information, can you measure out 45 minutes? |
|
Jun-16-25
 | | beatgiant: <WannaBe> In your puzzle, are we allowed to measure and cut the ropes? |
|
 |
 |
< Earlier Kibitzing · PAGE 95 OF 97 ·
Later Kibitzing> |