paul: Verified by Jacobi with this input:

stip dia 6
forsyth rnbqkbnr/pppp1ppp/8/8/4K3/5Q2/PPPP1PPP/RN3BNR
cond cavaliermaj (2023-06-16)
Mu-Tsun Tsai: Wait, 4.Ke2 walks into check by nightrider b8, isn't it? (2012-07-15, hidden)
Henrik Juel: You are right, Mu-Tsun, but this seems to work:
1.Nxe7 Nxe2 2.Nb1 Nxc1 3.Df3 Nca5 4.Ke2 Ngc6+ 5.Kd3 Nb8 6.Ke4 Ng8+
where the black nightriders switch places (2012-07-15, hidden)
Mu-Tsun Tsai: Yeah, I came up with the same thing. Computer check is still running to see if that's the only one. (2012-07-15, hidden)
Mu-Tsun Tsai: Hum... Probably have to give up computer check. The estimated run time is simply ridiculous. I'll just believe that this corrected solution is unique. (2012-07-15, hidden) more ...
comment

Henrik Juel: Without the condition the solution also works, but there are many cooks (2020-11-01)
Frank Müller: Diagramm prüfen: wKd1, wDe1 (2011-03-28, hidden)
paul: Checked by Jacobi (with relay condition) (2020-10-31, hidden) more ...
comment

Die Autoren schreiben:
Correction:
Problemkiste, Issue 91, Februar 1994, Seite 128, Abteilung Bemerkungen & Berichtigungen
"Nr. 3212 [sic!] v.Gruber/Müller: wie HG mitteilt muß die Bedingung "KAMIKAZE REX INCLUSIV" heißen, nicht Kamikaze-Circe rex inclusiv. Letzteres wäre nicht nur ein Zuviel an Bedingungen, sondern auch noch UL, weil da auch 31-steinige Stellungen möglich sind."
Dieser Eintrag wird also mit der Korrektur aktualisiert!
Erich Bartel: Das war ein schöner Tag am 21.8.1993 bei Hans Gruber zusammen mit Frank Müller und ich bedanke mich recht herzlich für die Widmung, die mir ausgezeichnet schmeckt, fast so gut wie der super delikate Original-Zwetschgendatschi von Hansens Mutter!!!(\eb) (2007-03-04)
A.Buchanan: I don't get this: please can someone explain (2020-12-09)
Henrik Juel: Under the Kamikaze-Circe condition, where both captured and capturing man are reborn after a capture, the initial game array is legal, but removing one man gives an illegal position, so the IGA is a legal cluster
I have no info about the super delicate original Zwetschgendatschi from Hansen's mother (2020-12-09)
A.Buchanan: Hi Henrik, please clarify the rules for the middle part. Why is e.g. 1. a4 b5 axb5[sBb7] with the white pawn disappearing because it can't be reborn on occupied b2? Is it because the rebirth square for capturer is taken to be the start square not the end square of the move? That seems unlikely because I read in PDB that promotion counts before capturer's rebirth. The other possibility is that capture is illegal if the capturer cannot be reborn, but that seems inconsistent with the spirit of Kamikaze. Normally in Circe, a captured piece doesn't come back if the rebirth square is occupied. That surely would still apply here. Glossary is silent on all this. Thanks! (2020-12-09)
Henrik Juel: I think you are right, Andrew, so the IGA is not a legal cluster under Kamikaze-Circe, and the problem is cooked
In the more serious matter at hand, Zwetschgendatschi means plum cake... (2020-12-09)
A.Buchanan: I can't believe that serious fairy heavyweights such as were engaged in this problem would have made something cooked. And I still don't know what K-C actually is. I will ask HG. Maybe he has some comments on plum cake too? :-) (2020-12-09)
A.Buchanan: Henrik has also been copied on an email from Hans Gruber which states in part:
"The problem by Frank and me (PDB: P1013968) is cooked - the intended solution does not work, because the initial game array is NOT a legal cluster under Kamikaze-Circe rex inclusive. Of course this was unavoidable, as dedication problems always are unsound. We should have known that before but did not see in-how-far we were wrong. It is good to learn about that a bit more than a quarter of a century after. I dare to state, without consulting Frank, that there will be no correction to the problem. The basic idea "initial game array = legal cluster" just does not work." (2020-12-09)
Arnold Beine: Perhaps the problem can be fixed by changing the fairy condition into "Platzwechselcirce rex inklusiv". Then every position with 31 pieces is illegal and the IGA should be legal. (2020-12-09)
Henrik Juel: The problem can probably be saved just by changing the condition to Strict Kamikaze-Circe RI (2020-12-09)
A.Buchanan: What is the exact rule for Kamikaze-Circe please? And how does strict differ? (2020-12-09)
Henrik Juel: In Kamikaze-Circe, as part of a capture, the capturer is reborn if its rebirth square is empty (otherwise it vanishes), and the captured is reborn if its rebirth square is empty (otherwise it vanishes). The rebirth squares are the standard Circe ones (based on the capture square).
Strict Kamikaze-Circe additionally stipulates that a capture is possible only when both rebirth squares are empty; so the piece count never changes (2020-12-09)
A.Buchanan: Thanks Henrik: that doesn't sound very "strict" Kamikaze, sounds like "no-sharp-edges video-game" Kamikaze, where no-one actually can get hurt :) Regular Kamikaze RI is trivially a legal cluster, no? And more interestingly, ordinary Kamikaze is a legal cluster too, because if just one piece is removed, that must have been captured by the *other* king, who is still stuck in his cage. That would be my suggested fix (2020-12-10)
Henrik Juel: Yes, any condition where 31 men cannot occur in a game would do
I chose strict K-C, because that was the most simple change to K-C (2020-12-10)
A.Buchanan: Strict K-C RI the number of pieces for each side must always be the same. K RX you can certainly have 31 pieces on the board but not in game array position. It’s still very simplistic but it’s a bit more subtle (2020-12-10)
comment