Die Schwalbe

2348 problem(s) found in 3568 milliseconds (displaying 100 problem(s)). [COMMENTDATE>=20200919 AND NOT S='Buffalo Sunday Times' AND G='h#'] [download as LaTeX]

1 - P0000036
Valery Liskovets
Nikita M. Plaksin

5528 Die Schwalbe 99 06/1986
P0000036
(10+11)
h#1.5 (AP, RV)
1) 1. ... cxd6ep 2. 0-0-0 Dxc7#
Nicht 2. Kf8 Dg8? da der ep-Schlag AP bewiesen werden muss.
2) 1. ... gxf6ep 2. 0-0-0 Tg8#
Nicht 2. Kd8 Tg8#? da der ep-Schlag AP bewiesen werden muss.
play all play one stop play next play all
VL: Solution:
I 1... c*d6 e.p. 2.0-0-0 Q*c7#
II 1... g*f6 e.p. 2.0-0-0 Rg8#

NOT:
2.Kf8? Q/Rg8#?? no AP-justification!

Castling implies e.p. 'a posteriori' and is necessary only to this end. (2002-04-03)
A.Buchanan: Solutions & tries (Popeye v.485) are:
1. ... cxd6ep 2. 0-0-0 Dxc7#
1. ... cxd6ep 2. Kf8 Dg8#
1. ... gxf6ep 2. d4 Tg8#
1. ... gxf6ep 2. 0-0-0 Tg8#
1. ... gxf6ep 2. Kd8 Tg8#
So after 1. ... gxf6ep:
2. Kf8 Rg8 is not even a mate (f7 flight)
and is 2. d4 a kind of dual thematic AP try? No-one's mentioned it. It would be easily fixable by shifting wBd2 to d4.
However, the main thing issue I don't understand PRA+AP logic (maybe I did once?). Under normal PRA algorithm, the "parts" for cxdep, gxfep & 0-0-0 respectively are YNY, NYY, NNN. Others are either impossible (3) or dominated by application of castling convention (2). OK so far.
There is no solution in the NNN part, so I suppose we must use AP in another part to eliminate it. But castling in another part where we are explicitly *assuming* that castling is ok seems to me to prove nothing. And if this inference did work somehow (which seems to me incredibly unlikely), why would we need to do it in *both* of these other parts? Surely one would be enough, but that means that the other part would be cooked.
Or maybe somehow AP "proves" that there are only 2 dimensions for PRA space, cxdep & gxfep, but that is so incredibly circular and horrible.
This kind of problem needs a proper step-by-step solution. Currently the solver can at best sort-of-vaguely-see-what-the-composer-might-be-alluding-to: i.e. it's just a joke.
Actually, I don't think that it can be PRA at all. I think it must be RS with two solutions supported by AP. But then there's still the question of why we need to castle in *both* solutions. I am sure I have seen an AP problem with two solutions where there is 0-0 in one solution and 0-0-0 in the other, which suggests that we *don't* need to castle twice in this problem. Which interpretation is correct, or do we have such a vague interpretation of these things (in order not to make independent spirits sad) that both are allowed to "work". (2020-12-30)
VL: Generally I agree with Andrew's analysis. The presented solution needs some corrections and clarification.

At first, the extra try 1.gxf6 e.p.? d4. I don't remember details but indeed, at present I'd prefer to exclude it by shifting wPd2 onto d4. One more thematic try enriches a problem but it is better to have an equal number of tries after both e.p.'s. In general, in AP-problems, thematic tries are highly desirable but not necessary formally for soundness. Dual avoidance in tries is here just due to the flight f7. Yes, 2.Kf8? Rg8#?? is indicated wrongly (unfortunately, H.103 with the published solution isn't available for me).

Now the most doubtful point: dual avoidance in the solution. Do we need castling in both partial solutions or at least one would suffice for justification? In my opinion, both approaches make sense in principle. In this problem implemented is the former, what means that every partial solution is permitted to be considered separately and, thus, should contain legalizing castling. This is similar to ordinary AP-problems, in which every line of the solution must contain castling. There is a distinct much subtler and rarer approach, which I call "Total AP", that considers all lines as a whole and thus requires to contain castling at least once (with subtle soundness issues). So, the current problem is treated (implicitly) not in the spirit of Total AP.

The problem implements a certain kind of triples of mutually retrodepending special moves (w/b castlings and e.p.). Namely, it is a "hybrid" AP-problem, where castling justifies a posteriori the typical pRA-choice between two e.p. captures. This is one of lucky triple kinds that admit (controversially) sound implementations based on hybrid logical conventions (logics for short). One more such a triple is represented in my earlier P0002474. Totally 37 distinct kinds of interacting triples can be identified, and only for few of them I know (or at least expect) reasonable implementations under special logics (whichever artificial and arguable they can be but consistent internally). Under an implementation I mean (loosely) a problem with an orthodox stipulation (#n, h#, etc.) supplemented by a suitable retro-convention like an ad hoc combination of pRA, RS or AP such that the position of the problem contains the corresponding triple of moves AND its full solution depends on all their combined (il)legalities. An extremely subtle matter. (2021-01-29)
more ...
comment
Keywords: a posteriori (AP), En passant as key (2), Partial Retro Analysis (PRA), Castling (sg), Volet Pawn
Genre: h#, Retro
FEN: r3k2n/1pp4P/N6p/2PpKpP1/2Q1p1R1/4P1pB/3P1p2/8
Input: Gerd Wilts, 1995-06-03
Last update: A.Buchanan, 2022-03-23 more...
2 - P0000047
Nikita M. Plaksin
Faat Fatchullin

5646 Die Schwalbe 101 10/1986
2. Preis
P0000047
(11+10)
h#2*
*) 1. ... 0-0-0 2. Txf2 Dxg1#
1) 1. cxd2+ Kxd2 2. Txf2 Dxg1#
R: 1. ... Da8-h8 2. d3xTe4 Da2-a8 3. c2xSd3 Sc1-d3+ 4. b4xLc5 Db1-a2 5. b3-b4 Ba2xSb1=D 6. Sa3-b1 und weiter z.B. (Beispielauflösung mri)
6. ... Te8-e4 7. Sc4-a3 Sg4-h2 8. Se5-c4 Th2-g2 9. Sf3-e5 g2-g1=L 10. Sg1-f3 Sh6-g4 11. Sf3xDg1 Th8-e8 12. Sd4-f3 f3xSg2 13. Se3-g2 f4-f3 14. Sc6-d4 Tg2-h2 15. Sa7-c6 Kh2-h1 16. Sc8-a7 Dh1-g1 17. Sb6xLc8 Tg1-g2 18. Sf5-e3 Dc6-h1 19. Sh4-f5 Kh1-h2 20. h2-h3 a3-a2 21. Sa4-b6 Sa2-c1 22. Sb6-a4 Sb4-a2 23. Sa4-b6 Sd5-b4 24. Sb6-a4 Se3-d5 25. Dh3-f1 Sf1-e3 26. Dg4-h3 f5-f4 27. Df4-g4 a4-a3 28. Db4-f4 a5-a4 29. Da3-b4 Kg2-h1 30. Sf3-h4 Kh3-g2 31. Da2-a3 g2-g1=T 32. Db1-a2 Kg4-h3 33. Dd1-b1 h3xTg2 34. Tg1-g2 Kg5-g4 35. Sh4-f3 Kf6-g5 36. Th1-g1 Se3-f1 37. Sf3-h4 h4-h3 38. Sg1-f3 e6xLf5 39. Lh3-f5 Ke7-f6 40. Lf1-h3 Sc4-e3 41. g2-g3 Ke8-e7 42. Sa4-b6 Sg8-h6 43. a2xTb3 Tb6-b3 44. Sh3-g1 Ta6-b6 45. Sb6-a4 Ta8-a6 46. Sg1-h3 a7-a5 47. Sh3-g1 Sa5-c4 48. Sg1-h3 Sb3-a5 49. Sc4-b6 Sc1-b3 50. Sa3-c4 Sb3xLc1 51. Sb1-a3 Sa5-b3 52. Sh3-g1 h5-h4 53. Sg1-h3 c4-c3 54. Sh3-g1 Lf8-c5 55. Sg1-h3 Dc7-c6 56. Sh3-g1 Dd8-c7 57. Sg1-h3 c5-c4 58. Sh3-g1 h7-h5 59. Sg1-h3 e7-e6 60. Sh3-g1 Sc6-a5 61. Sg1-h3 Sb8-c6 62. Sh3-g1 c7-c5 63. Sg1-h3
play all play one stop play next play all
Anton Baumann: Auszeichnung Informalturnier 1986: 2.Preis
Preisbericht: 'Die Schwalbe' 06/2011 S.124 (2023-01-02)
Henrik Juel: How is the SE corner released, without ruining the castling? (2023-01-02)
Mario Richter: Good question, Henrik! I first thought that releasing the SE corner without ruining White's castling right is impossible, but the trick is to uncapture a black Queen in the SE corner at the right moment.

Perhaps Theodore Hwa can use ths problem as a test case for his latest improvement to Retractor 2 ... (2023-01-02)
Henrik Juel: Thanks, Mario
In view of the prize I suspected that the problem was correct, but I did not find the uncapture trick (2023-01-02)
Henrik Juel: C+ Popeye 4.61, because with Black to move White may not castle (2023-01-02)
comment
Keywords: Castling (wl)
Genre: h#, Retro
FEN: 7q/1p1p1pp1/8/2P5/4P3/2p3PP/1P1PPPrn/R3KQbk
Input: Gerd Wilts, 1995-06-03
Last update: Mario Richter, 2023-01-02 more...
3 - P0000058
Leonid M. Borodatow
5758v Die Schwalbe 103 02/1987
P0000058
(13+9) C+
h#3
b) sBa6 statt sLa6
a) 1. ... exf6ep 2. 0-0-0 Lxf4 3. Td7 a8=D#
b) 1. ... Kxf4 2. Te7 e6 3. Tb8 axb8=D,T#
play all play one stop play next play all
Anton Baumann: Mattdual in b): 3.Tb8 axb8=D,T# (2022-12-16)
A.Buchanan: So Borodatow got it all working! In a) Black might have captured hxgxf and axPb. So castling rights might still be maintained with the ep. In b) on the other hand, it must be axb, bxa, exf and wPgxh6, so there was the cage. Promotion to TD is tolerated in the final move, although it may not be puristic, without the convention, too many mates would be excluded (2022-12-16)
comment
Keywords: En passant as key, Castling (sg), Cant Castler, Valladao Task
Genre: h#, Retro
Computer test: HC+ Popeye v4.87 & basic retro thinking
FEN: r3k3/P7/b3r1pP/4PpBP/3nnpKR/5PRB/5PP1/7N
Input: Gerd Wilts, 1995-06-03
Last update: A.Buchanan, 2022-12-16 more...
4 - P0000101
Leonid M. Borodatow
Evgeny V. Kharichev

6209v Die Schwalbe 110 04/1988
P0000101
(12+10) cooked
h#3
b) sBh5 nach h6
a) 1. 0-0-0 Sxc7 2. hxg3 Sd5 3. g2 Se7#

Beispielauflösung mri:
R: 1. Kf1-e1 Ta2-a1 2. Kg2-f1 Tb2-a2 3. La2-b1 Tb1-b2 4. Kf3-g2 Tf1-b1 5. Ke3-f3 g5xDh4 6. Lh2-g1 Tg1-f1 7. Dd4-h4 Tg2-g1 8. Lg1-h2 Th2-g2 9. Kd3-e3 Tg2-h2 10. Kc3-d3 Th2-g2 11. Kb2-c3 Tg2-h2 12. Kc1-b2 Th2-g2 13. Kd1-c1 Tg2-h2 14. Ke1-d1 Th2-g2 15. Da1-d4 Tg2-h2 16. Dd1-a1 d6xLc5 17. La3-c5 Th2-g2 18. Lc1-a3 Tg2-h2 19. b2-b3 Th2-g2 20. Ld5-a2 Tg2-h2 21. Sc5-a6 Th2-g2 22. Lg2-d5 h6-h5 23. Lf1-g2 Tg2-h2 24. Lh2-g1 Tg1-g2 25. Sa4-c5 g2-g1=T 26. Sb6-a4 f3xTg2 27. Tg1-g2 f4-f3 28. g2-g3 f5-f4 29. Le5-h2 h7-h6 30. Lg7-e5 f6-f5 31. Lf8-g7 f7-f6 32. e7xLf8=L Lg7-f8 33. e6-e7 Lf8-g7 34. d5xTe6 Tg6-e6 35. Sc8-b6 Tg8-g6 36. Sb6xLc8 g6-g5 37. c4xDd5 Dg5-d5 38. Sd5-b6 Dd8-g5 39. Sg3-h1 e7xTd6 40. Ta6-d6 Th8-g8 41. Ta1-a6 g7-g6 42. b3xSc4 Sa5-c4 43. a2xSb3 Sd4-b3 44. Sc3-d5 Sc6-a5 45. Sb1-c3 Sb8-c6 46. Th1-g1 Sf5-d4 47. Se4-g3 Sh6-f5 48. Sg5-e4 Sg8-h6 49. Sf3-g5 Sh6-g8 50. Sg1-f3 Sg8-h6 51. h2-h3


b) 1. Kf7 f4 2. Kg6 c3,c4+ 3. Kh5 g4#
play all play one stop play next play all
in b) fehlt Schwarz ein Retrotempo zur Aufrechterhaltung des Rochaderechts, es kann maximal eine Stellung wie z.B. r3kB2/pppp1p1p/3p2p1/8/8/7P/1PPPPPP1/1NBQKBRN erreicht werden, in der entweder sTh8 oder sLf8 nicht als Schlagobjekt für die notwendigen weißen Bauernschläge zur Verfügung stand.

Der Dual 2. ... c3+,c4+ in b) war von den Autoren in Kauf genommen worden., die (schon bei der Lösungsveröffentlichung bekannten) NLs aber natürlich nicht. Ein Schwalbe-Löser ("PS") kritisierte den "völlig unnötigen Vorwärtsballast" und schlug implizit vor, das Vorwärtsspiel komplett wegzulassen und stattdessen direkt zu fragen: "Ist die 0-0-0 erlaubt?"
Cook: a) 1. hxg3 Sxg3 2. c6 Sf5 3. 0-0-0 Sd6#
b) 1. Kf7 f4 2. Kg6 c4+ 3. Kh5 g4#
Neufassung 5949.
A.Buchanan: It would be great if someone can retrieve 6209v from Die Schwalbe, please, to confirm the diagram and intended solution. If castling right remains, then sBf must promote on g1=T. There is then some jiggling around because Tg1-h2 to unpin Lg1 and let it go b1. In (a) this is doable, but in (b), Black runs out of tempi however there is an alternative no-castling mate with sKh5. However that's just a rough sketch, and I don't want to spend more time until the actual published form is confirmed (2022-01-08)
Ladislav Packa: Is everything allright here? Because solution a) is not a solution - after 3 .... Se7 # ?? 4.Kc7 / b8! (2022-01-09)
Mario Richter: Indeed, the given solution only works with wLh2 instead of g1. I only have access to the solution, but not to original diagram. Perhaps someoe with access to it can give us the correct details? (2022-01-10)
A.Buchanan: Yes I didn't check the solution to a) matches one of the actual ones. With wLh2, it's still cooked, as is b) in any case. Possible to rearrange the free pieces to make b) sound, but I don't see how to fix a) at the same time. (2022-01-10)
more ...
comment
Keywords: Castling (sg), Cant Castler (sg)
Genre: h#, Retro
FEN: r3k3/pppp4/N7/2p4p/7p/1P4PP/2PPPP2/rB2K1BN
Input: Gerd Wilts, 1995-06-03
Last update: Mario Richter, 2022-01-10 more...
5 - P0000416
Thomas Brand
4496 Die Schwalbe 83 10/1983
P0000416
(1+2+1) cooked
-1(w+s), dann h#1
Imitator e4
R: 1. ... Kg8xBf7,(If5-e4) Kb4xSb3,(If6-f5), dann 1. Ka4,(Ie6) Sc5#,(If8)
play all play one stop play next play all
Cook: NL
R: 1. Ke7xLf7,xSf7,xBf7,(Id4-e4) d6xDe5,(Ic5-d4), dann 1. Kc2,(Id4) Db2#,(Ia1)
R: 1. Kg6xTf7,Ke6xTf7,(If3-e4,Id3-e4) d6xDe5,(Ie4,Ic4), dann 1. Te7,Tg7,(Id4) Db2#,(Ia1)
Anton Baumann: 'Die Schwalbe' 08/1984 S.308: der Autor korrigiert: wBe5 statt sBe5 (2022-12-23)
comment
Keywords: Help retractor, Kindergarten Problem
Pieces: i = Imitator (I)
Genre: Retro, Fairies, h#
FEN: 8/5K2/8/4p3/4-I3/1k6/8/8
Input: Gerd Wilts, 1995-06-03
Last update: A.Buchanan, 2021-12-13 more...
6 - P0000583
Andrey Frolkin
Günter Lauinger

3468v Die Schwalbe 72 12/1981
P0000583
(10+12) C+
h#2 (AP)
Satzspiel:
*1. ... 0-0! 2. cxd4 Tc1#

Verführung:
1. exd3ep? 0-0 2. cxb4 Tf4# aber zuletzt R: 1. d2-d4?? ist illegal, da es den wLc1 als schwarzes Bauernschlagobjekt ausschließt

Lösung:
1. ... 0-0! (AP-Legalierung des weissen Anzugrechtes) 2. cxd4 Tc1#
play all play one stop play next play all
Alle Versuche, das Satzspiel 1. ... 0-0! 2. cxd4 Tc1# durch einen schwarzen Vorschaltzug aufrecht zu erhalten, scheitern.
Mario Richter: 1. exd3ep 0-0 2. cxb4 Tf4 ist nur Verfuehrung (scheitert an schwarzer Bauern-Schlagbilanz).
Loesung: 1. ... w0-0! (AP-Legalierung des weissen Anzugrechtes) 2.cxd4 Tc1# (2009-02-10)
A.Buchanan: Supersedes P0000553 (2022-01-08)
Mario Richter: Hi Andrew, I only quoted from the "official solution" ('Die Schwalbe' Heft 72, 12/1981, p.399).
Perhaps a better way to get anwers to your interesting questions is to ask the authors directly ... (2022-01-08)
A.Buchanan: AP Type Petrovic is a try, because wLc1 was captured at home, yet Black still made 6 pawn captures.
AP Type Keym is the actual solution, because if it was really BTM, then White would have lost castling rights.
So there are two kinds of AP here. (2022-03-21)
more ...
comment
Keywords: Castling (wk), a posteriori (AP) (Type Keym), En passant as key, a posteriori (AP) (Type Petrovic)
Genre: h#, Retro
Computer test: BC+ Popeye v4.87
FEN: 8/2p5/1pPp4/bRpP4/BPkPp3/qp2p2p/rP2P3/4K2R
Input: Gerd Wilts, 1995-06-03
Last update: A.Buchanan, 2022-03-21 more...
7 - P0000615
Andrey Lobusov
4088 Die Schwalbe 77 10/1982
14.Lob
P0000615
(15+6) cooked
h#2* (AP)
1. ... Kg5 2. 0-0?? Se7#
1. ... Txh7 2. Kf8 Txh8#

1. cxb3ep Txh7 2. Kf8?? Txh8#
1. cxb3ep Kg5 2. 0-0! Se7#
play all play one stop play next play all
Wh captured 10 by pawns (including hxPg-g8) but not on the last move. wPa3 came from a2: wPa5 captured bPa.
Bl took only b3xa2.
In set play, Bl has no last move except for K or R.
For the main play, the intention is that if Bl retains castling rights, last moves were R: 1. b2-b4 b3xa2 e.g. 2. Bg1-d4 b4-b3 3. d4xc5 etc. However there are retro cooks e.g. R: 1. Bg1-d4 b3xa2 2. d4xc5 etc, with wPa5 the original wPb. wPb4 & wPb5 come from further East.
Cook: Retro logic for main play is unsound.
A.Buchanan: Not trivial to fix as shifting wPa5 to b6 leads to promotion cooks. Adding black material must be carefully done to avoid disrupting uniqueness of e.p. as tempo. (2022-03-11)
A.Buchanan: OK just to say I have found a form of this idea that works. This one was the very devil to fix - the final form includes different mates and retro logic, so becomes an "after" rather than a "correction", and I will find somewhere else to publish first. (2022-03-19)
more ...
comment
Keywords: Castling (sk), a posteriori (AP) (Type Petrovic), En passant as key, Tempo Move, Superseded by (P1409841)
Genre: h#, Retro
Computer test: Popeye v4.87 & simple retro-logic
FEN: 4k2r/5p1p/5K1R/PPP2N2/1PpB4/P1P5/p1RNB1P1/Q7
Input: Gerd Wilts, 1995-06-03
Last update: A.Buchanan, 2023-05-20 more...
8 - P0000649
André Hazebrouck
2261 Die Schwalbe 47 10/1977
4. Preis
P0000649
(8+14)
h#2.5
b) AP (2 Lösungen)
a) 1. ... 0-0-0 2. 0-0 Tdg1+ 3. Kh8 Txh6#
b) also 1. ... cxb6ep 2. La6 0-0 3. 0-0-0 a8=D#
is clearly the intention
play all play one stop play next play all
Can count 1+7 visible pawn captures, leaving 1+1 unexplained. Need to resolve wPfgh & bPg. Suppose bK never moved, then one of:
1) wPf waylaid, wPhxPg=, wPg=. Here w00 right can remain.
2) Or similarly, wPh waylaid, wPfxPg=, wPg=. Again w00 right can remain.
3) bPgxPh=, wPfxg= (or wPfxe then captured by bPf), wPg= w00 right lost.
1+1 captures explained in either case.
On the other hand if bK moved, then maybe wPf=, and only requires one more capture to resolve g&h files. The w00 right can remain. So there is a captured unit unaccounted for, and we can't validate the ep.
On the other hand if we are in case 3 above, then we can't know that wPbxa captured dark bishop, so White might retract a6-a7.
We also need to know that w00 rights remain. s000 rights (i.e. prior movement of sTa8) are not relevant.
If we know that w00, w000 & b000 remain, then we are in case 1 or 2 above with bPg captured unpromoted. Therefore wPb6xLa7 due to bishop shade. Therefore R: 1. b7-b5 b6xLa7 to unblock.

Only one candidate solution does not begin with ep: 1. ... 0-0-0 2. 0-0 Tdg1+ 3. Kh8 Txh6#. So this is the solution for a).

Let ????? denote validity of w00,w000,b00,b000,ep. Possibilities are: YYY?Y and all of YYN??, YNY??, NYY??, YNN??, NYN??, NNY??, NNN??. So there are 30 possibilities.
Under PRA, the solution parts would be YYYYY, YYNYN, YNYYN & NYYYN. For the first there are 20 solutions, while the second and third have 0 solutions. So this is not the right paradigm.
Under SPRA, there would be a single solution part YYYYY with 20 solutions. So this is not the right paradigm either.

Under RS with AP, the solution of a) still works. This comes from ?YY??. On the other hand, if the first move is ep, then we are in YYYYY. So all castlings are valid. But when we get to perform the mate, we need to know that based on castlings actually performed, the ep is valid. So the solution must include w00. So is 1. ... cxb6ep 2. La6 0-0 3. 0-0-0 a8=D#.

By combining the information of both of these solutions, we know we are in YYY?? so YYY?Y is the only possibility, and ep is legal. However, the combination of these two solutions would also validate any of the other 18 ep solutions that do not include 2. ... 0-0. They contribute no new evidence, but how to exclude them?

This is one of the issues with "AP Consolidation". Normally the number of solutions is not a constraint. But here perhaps we should insist that *only* two solutions are allowed?
Cook: 1. ... cxb6ep 2. La6 Tb1,~ 3. 0-0-0 a8=D#
18 different possibilities for W2 which are validated by the same logic that validates 2. ... 0-0.
Henrik Juel: [I don't follow the silly convention of writing black moves first]. 1.0-0-0 0-0 2.dTg1+ Kh8 3.Txh6#. b) If Ke1,e8 and Th1 never moved, White captured f/hxPg and b6xLa7; if Ta1 also never moved, last white move was b6xLa7. 1.cxb6ep La6 2.0-0/0-0-0 0-0-0 3.a8Q#. The a) solution also works in b). (2003-12-18)
more ...
comment
Keywords: a posteriori (AP) (Type Petrovic), Castling (wkwlsk), En passant as key
Genre: h#, Retro
Computer test: Forward logic Popeye v4.87 & basic retro-logic works, but then AP Consolidated protocol seems to cook the problem.
FEN: r1b1k2r/P2p4/2p1n2p/ppPq4/1np5/p7/P2PP3/R3K2R
Input: Gerd Wilts, 1995-06-03
Last update: A.Buchanan, 2022-05-24 more...
9 - P0000653
Eliahu Fasher
2311 Die Schwalbe 48 12/1977
P0000653
(12+14) cooked
h#2 (wer?)
Kees: possible fix: Lb1=Sb1 -De1 +Ld1
White begins: 1.Kxb7 Lxe2 2.Kc8 La6#
(1.Txd8+ Kxd8 2.Kf8 Th8# illegal for white has no last move) (2023-06-07)
A.Buchanan: Your fix is good, Kees. It removes the cook, and sLd1 denies R: 1. c2xb3 as well as sLb1 did. Note R: 1 Sc6-d8 0-0+? as black castling rights were lost to let wK enter the back rank. (2023-06-08)
more ...
comment

Genre: h#, Retro
FEN: 1bKN1rk1/1ppn1r1R/5p1P/4pP1p/3p1p2/1PP3P1/PP2P3/Rb2q3
Input: Gerd Wilts, 1995-06-03
Last update: A.Buchanan, 2023-06-08 more...
10 - P0000674
Leonid M. Borodatow
2475 Die Schwalbe 51 06/1978
P0000674
(12+11) cooked
h#3
1. ... Kgxf4 2. Tf6 e6 3. Tf8 Sg7#
play all play one stop play next play all
Cook: 1. ... exf6ep 2. 0-0-0 gxf4 3. Td7 a8=D
Failed try, as Bg4xBh5 allows bRh8 to have escaped the cage prior to ep.
Anton Baumann: Sollte eine Verbesserung von P0000777 sein.
Autorabsicht: "letzter Zug f7-f5 sonst retropatt! Jetzt ist aber die Rochade illegal, da der sTh8 nur über e8 aus seiner Ecke kommt. Daher Verführung 1. ... exf6 e.p. 2.0-0-0? Lxf4 3.Td7 a8=D#. Lösung: 1. ... Kxf4 2.Tf6 e6 3.Tf8 sg7#"
Aber: Der s h-Bauer kann auch auf h5 geschlagen worden sein; so konnte der sT über h6 aus seiner Ecke entkommen. Die vermeintliche Verführung ist also NL! (2022-12-14)
comment
Keywords: Castling (sg), Superseded by (P0000058)
Genre: h#, Retro
FEN: r3k3/P7/p1r3pP/4PpBN/3ppnKn/6P1/1PP2PPq/7N
Input: Gerd Wilts, 1995-06-03
Last update: A.Buchanan, 2022-12-16 more...
11 - P0000760
André Hazebrouck
1120 Die Schwalbe 23 10/1973
P0000760
(9+10) C+
h#2 (AP)
1. bxc3ep Sa6 2. 0-0-0 Tc4#
play all play one stop play next play all
Henrik Juel: -1.c2 c5xLb4 -2.Lc3 c6 -3.Lb2 c7 -4.Lc1 a6! -5.b2 b2xDSa2 etc. White captured sDTSS by fxexd, gxfxe, allowing Black to capture f7xPe6xPd5 etc. (2003-12-18)
A.Buchanan: Why not e.g. wSb1? (2022-03-04)
Henrik Juel: That also seems to work (2022-03-04)
A.Buchanan: I really like the motivation for ep. I guess the motivation for Rb1 is to be inside the cage that forms around it, but it’s still a bit loose. Note there is no retro try. (2022-03-05)
more ...
comment
Keywords: En passant as key, Castling (sg), a posteriori (AP) (Type Petrovic), Volet Pawn
Genre: h#, Retro
Computer test: HC+ Popeye 4.87 + simple retro reflection
FEN: rN2k3/1p1pp1pp/8/p7/RpP5/PP6/p2PP3/KR6
Input: Gerd Wilts, 1995-06-03
Last update: A.Buchanan, 2022-03-05 more...
12 - P0000777
Leonid M. Borodatow
1424 Die Schwalbe 30 12/1974
P0000777
(12+10) cooked
h#2.5
1. ... exf6ep 2. 0-0-0? Lxf4 3. Td7 a8=D# try
1. ... Kxf4 2. Tf6 e6 3. Tf8 Sg7# solution
play all play one stop play next play all
Cook: 1. ... Kxf4 2. Te7 e6 3. Tb8 axb8=D,T#
See P0000674
Henrik Juel: The stipulation should probably read h#2.5: 1.exf6ep 0-0-0 2.Bxf4 Rd7 3.a8Q# (2003-08-18)
Anton Baumann: Letzter Zug war f7-f5, sonst retropatt. Die weissen Bauern haben 6x geschlagen. Der s a-Bauer wurde so ebenfalls als Schlagopfer benötigt und musste 2x schlagen, um sich umwandeln zu können. Weiss hat also auf h6 den s Bauern geschlagen. Unabhängig davon ob der sTe6 ein Umwandlungsturm ist oder nicht, muss der sTh8 über e8 aus der Ecke gekommen sein. Also ist die Rochade illegal!
Autorabsicht: 1. ... exf6 e.p. 2.0-0-0 Lxf4 3.Td7 a8=D# ist die Verführung!
Lösung: 1. ... Kxf4 2.Tf6 e6 3.Tf8 Sg7#
leider geht auch: 1. ... Kxf4 2.Te7 e6 3.Tb8 axb8=D,T# (2022-12-13)
A.Buchanan: I agree Anton. I suggest as a fix -bBa6 +bDc2. I think retro, solution and try all work correctly now. Do you folks agree? (2022-12-13)
Anton Baumann: Der Vorschlag von Andrew ergibt eine korrekte Lösung, aber nicht im Sinne des Autors: ohne einen s Stein auf a6 ist f6-f5+ als letzter Zug möglich, d.h. kein Retropatt weil Weiss a6-a7 zurücknehmen kann! 1. ... exf6 e.p. ist deshalb nicht zulässig. LB wollte aber dass f7-f5 der letzte Zug war, damit aber die Rochade illegal ist! (2022-12-14)
Anton Baumann: Mögliche Korrektur: a6=sL, g2=sD (2022-12-15)
A.Buchanan: Yes I agree. I had overlooked the unmove of wPa7 (2022-12-15)
more ...
comment
Keywords: Castling (sg), Valladao Task, Superseded by (P0000058)
Genre: h#, Retro
FEN: r3k3/P7/p3r1pP/4PpBN/3nnpKP/6PB/1P2PPb1/8
Input: Gerd Wilts, 1995-06-03
Last update: A.Buchanan, 2022-12-16 more...
13 - P0000846
Klaus Weisert
181 Die Schwalbe 04/1970
P0000846
(3+7) C+
h#1
wZug also
1. ... b3!#
1. d1=S?,d1=T? b3#
play all play one stop play next play all
Two indistinct retro tries.
Originalforderung: 'Hilfsmatt' in 1 Zuge

Die 'Schwalbe' (Heft 9, Juni 1971, S. 230) zum recht trivialen Inhalt des Problems: "Warum bringt der Redakteur so etwas? Erstens, um einen Verfasser, der sich zum ersten Mal an den Rand des Schachbretts [Name der Urdruck-Rubrik, in der solche Probleme publiziert wurden] begibt, zu ermuntern; und zweitens (indem er das Stück an den Anfang stellt), um Löser, die sich sonst nicht an den Rand des Schachbretts trauen, mit einigen leichten Punkten dorthin zu locken!"

Nicht 1. d1S(T) b3#, denn Schwarz ist nicht am Zuge; daher ganz simpel 1. b3#
A.Buchanan: I suspect I don't get the full idea here. (2021-10-20)
Henrik Juel: I do not either
The two possible promotions do not help (2021-10-20)
comment
Keywords: No legal last move for White
Genre: h#, Retro
Computer test: HC+ Popeye v4.87 & trivial retro-logic
FEN: 8/8/8/rr6/kp6/p7/KP1p4/Rn6
Input: Gerd Wilts, 1995-06-03
Last update: A.Buchanan, 2023-12-06 more...
14 - P0000899
Giuseppe Brogi
743 Die Schwalbe 06/1972
P0000899
(8+15) cooked
h#2
b) wSa1
a) 1. T2xh3 Txh3 2. 0-0 Th8#
b) 1. La4 0-0 2. Tf8 Te1#
play all play one stop play next play all
a) White cannot cross-capture, so because of sTh2, w0-0 impossible.
wBe captured sD, either to be captured on f-file or to promote without disrupting sKe8.
b) sTh2 can return to a8 only via e8 as Black cannot cross-capture, so s0-0 is impossible.
wBe was waylaid or promoted after disrupting sKe8.
Cook: 1. T2xh3 Lb4 2. Txg3 Txh8#
1. Sd2 Lf6 2. Tf8 Sd6#
(cooked by MR)
See P0003736 a companion problem.
milan: [-bPa7b6wPg3bSa1+bPg3bBa5-b4 bPc6-d6] h#2 1.2...bRh2? illegal move? M.Frelih (2023-03-30)
more ...
comment
Keywords: Cant Castler, Castling (wksk), Cross-capture (s,w), Superseded by (P1399805)
Genre: h#, Retro
Computer test: Popeye v4.87
FEN: B3k2r/pN1p1p2/1pp3p1/bb3p2/8/p1B3PP/5P1r/nn2K2R
Input: Gerd Wilts, 1995-06-03
Last update: A.Buchanan, 2022-03-16 more...
15 - P0001107
Michel Caillaud
3332 Die Schwalbe 66 12/1980
2. Preis
P0001107
(7+9) C+
h#4.5 (AP)
Circe
1. ... hxg6ep[+sBg7] 2. a1=T+ Txa1[+sTh8] 3. 0-0 Ta3 4. bxa3[+wTa1] 0-0-0 5. Td8 Txd8[+sTh8]#
play all play one stop play next play all
Wenn die beiden Könige noch nicht gezogen haben (was a.p. durch Ausführung beider Rochaden bewiesen wird!), kann der letzte Zug nur g7-g5 gewesen sein, Sonst ginge z.B. auch R: 1. Kd8-e8 c2xDb3
Henrik Juel: A possible retroplay is -1... g7 -2.f3xP(-Pg7) b5 -3.f2 c6xP(-Pb2) -4.b2 a6 -5.Rc3 a7 -6.Rc1 b6xB -7.Bf8 b7 -8.Bg7xB etc. (2003-04-22)
Michel Caillaud: The problem is not cooked as indicated, as under A Posteriori convention both castles must be played in the solution to justify the en passant key. (2022-02-04)
A.Buchanan: If wK never moved, then Black has not just moved another non-capturing pawn, as White has no prior move. The only possible exception is R: 1. g7-g5 f/h3xPg4[+Pg7]. (2022-02-04)
comment
Keywords: a posteriori (AP), Circe, En passant as key, Castling (wgsk), Promotion in forward play, Valladao Task
Genre: Retro, Fairies, h#
Computer test: HC+ all solutions begin with ep, but only one includes both white & black castling
FEN: 4k3/8/7p/p1p3pP/Pp4Pp/RP4p1/pP6/4K3
Reprints: feenschach 61, p. 534, 08/1982
RA152 diagrammes 72 07-08/1985
Input: Gerd Wilts, 1995-06-03
Last update: A.Buchanan, 2023-06-28 more...
16 - P0001117
Michel Caillaud
3872 Die Schwalbe 74 04/1982
12. Lob
P0001117
(11+12)
h#2
b) sBh7 nach h6
a) 1. 0-0-0? Sg6 2. Lb8 Sxe7! illegal castling
1. Kf7! Sxh7 2. Ke6 Sg5#
b) 1. 0-0-0! Sg6 2. Lb8 Sxe7#
1. Kf7? Sh7 2. Ke6 Sg5+? 3. hxg5!
R: 1. Ke3-e4 Sd6-b7 2. h2xDg3 Dg2-g3+ 3. Kd2-e3 Db7-g2 4. Ke1-d2 Dc8-b7 5. e4-e5 Dd8-c8 6. f3xLe4 Lb7-e4 7. Kf1-e1 Lc8-b7 8. Ke1-f1 b7-b6 9. Kf1-e1 Lb6-a7 10. ... La5-b6 11. ... Ld2-a5 12. ... Lc1-d2 13. ... d2xTc1=L 14. ... e3xTd2 15. ... e4-e3 16. ... f5xLe4
play all play one stop play next play all
Missing: Wh: QRRBB Bl: QRBB
Captures: Wh: gxf3xe (inc B), hxg + [Bf8] Bl: axbxc (for Ba7) fxexdxc/e (g1 not possible)
Assume Black can castle: so neither bK nor bRa have moved. Before bPb7-b6 (which releases QB) *all* Bl captures have been made. wPd can freely advance, all Wh units released except for rooks & Bf1. gxRf3 is forced, and now all Wh units are free and can be captured. To avoid deadlock, wB was captured on e4 not e6. Sequence must be f7-f5 Rh8-f8-f6-...-f3gxRf3.
In this position wLe4 must be played back to f1. With bPh7, wBe4 must retract either by stopping on f7, (disrupting bK) or via f5 (implying retraction of f6-f5, in which case bRa8 is an imposter).
In the alternative route via h7, wB crosses over f7 harmlessly. While wB is on g5 & h4, f7 must be occupied by a static shielding knight, but there is no tempo issue. After all this excitement, b7-b6 if followed by simple and non-unique play to reach the diagram.

(Gerd's earlier solution: Weiße Schläge: h2xg3, gxfxe, sLf8. Schwarze Schläge: a7xb6xc5; fxexdxc1=L In dieser Stellung muß der wLe4 nach f1 zurückgespielt werden. Mit sBh7 kann der wLe4 nur entweder über f7 nach f1 zurück, so daß der sK bereits gezogen haben muß, oder der wLe4 kann über f5 zurück, wozu aber der sTa8 nach h8 zurückgezogen werden müsste, um f6-f5 zurücknehmen zu können.)
Henrik Juel: To make the retroplay plausible one should uncapture bQ early on g3 and retract it to d8. The wB could also get back to f1 via f5, but this would require retracting bRa8 to h8 before retracting bPf5, so castling is still illegal in part a). (2003-04-10)
Gerd Wilts: Hello Henrik, thank you for pointing out the inaccuracy of the solution, I will make the solution more precise soon. And thank you for adding all the other solutions! (2003-04-11)
A.Buchanan: Have posted a solution based on GW&HJ ideas. More often a j’adoube of a rook pawn signals a tempo idea, but not here. The surprising motivation harmonizes with accurate and varied forward play. (2021-10-24)
A.Buchanan: This problem was featured in yesterday's Monthly International Zoom Call, and was a great success, with audible gasps of appreciation, as the significance of h7 was realized :) (2021-10-24)
A.Buchanan: Oh dear, Alfred Pfeiffer has silently reverted the German keyword to "mit Umwandlungsfigur". I'm not going to enter an "edit war" with him, but I would appreciate if he can explain his position here. This is a problematic concept to keyword, but to me "mit Umwandlungfigur" is weak and inaccurate. What is the intended distinction with the existing keywords "Umwandlung" or "Umwandlungen"? It's hopeless. We have in PDB very many poor promotion keywords, and I would like to clean up progressively. I don't know if a native German speaker would care to engage with Alfred on this point. (2021-10-30)
A.Buchanan: To be clearer: to me the German definition seems pretty good. I think the term itself should give more of a clue what's happening :-) (2021-10-30)
more ...
comment
Keywords: Cant Castler, Castling (sg), Promotion (l), Obvious promotion (l), Corridor, Retro Shield
Genre: h#, Retro
Computer test: Forward: C+ Popeye V4.87 Retro: non-trivial reasoning
FEN: r3kN2/bnppp1pp/1p6/2p1P3/4K3/3P2P1/PPP1PP2/N6n
Input: Gerd Wilts, 1995-06-03
Last update: A.Buchanan, 2021-10-30 more...
17 - P0001191
Leon Loewenton
64 Europe Echecs 38 10/1961
P0001191
(15+14) C+
h#2
b) wSd3 nach e1
a) 1. Kd8 Se5 2. Te8 Sxf7#
b) 1. 0-0 Sd5 2. Sh8 Sxe7#
play all play one stop play next play all
Originalquelle?

nicht sicher, ob Zwilling b) auch schon im Original.
Artikel von Petrovici zum "Thema Than" 03/2016 gibt als Originalquelle an: "Europe Echecs 10/1951"
hans: Counting problem
1. Kd8 Se5 2. Te8 Sxf7# !!
1. 0-0 Sd5 2. Sh8 Sxe7# ??
Black needs a tempo to get on move, and the only piece to do that is Th8, so 0-0 is illegal. (2010-06-20)
Mario Richter: The term usually used to describe this kind of problems is "Parity problem". The bRh8 might have been on h8 all the time, since the tempo move might also have been made by the black Queen, but in this case, the bK must have already moved, thereby losing the right to castle too.. (2010-06-21)
Ladislav Packa: You both are right. Black would be in this position made an uneven number moves.
This can be achieved by using Rook moves (eg Rg8-f8-h8), or by the Queen moves while she was still alive (eg Kf8-g8, then Qf8-e8-d8 and then the King back). But both possibilities preventing black castling. (2013-10-04)
A.Buchanan: Parity change could also be achieved without triangulation by e.g. SxDe8 after the black queen has moved once (2022-03-15)
comment
Keywords: Castling (sk), Parity Argument, Cant Castler, Than theme
Genre: h#, Retro
Computer test: rawbats
FEN: r1b1k2r/1ppppppp/p5n1/8/8/P1NN3P/1PPPPPPR/nRBK1B2
Reprints: 787 Themes-64 10-12/1961
(7) diagrammes 112 01-03/1995
Input: Gerd Wilts, 1995-06-03
Last update: Mario Richter, 2018-03-22 more...
18 - P0001198
Tivadar Kardos
1807 Problem 73-78 06/1961
P0001198
(8+15) C+
h#2
1. Dxc2 Tc1 2. 0-0-0? Txc2#
1. Sd7 0-0-0 2. 0-0 Tg1#
play all play one stop play next play all
Black qside castling is lost as wPa promotes but other rights can be maintained.
S: axb,cxb,dxcxb,fxexBd,gxfxe
W: a=X,g=X,fxe
A.Buchanan: wPh can be removed (2021-11-22)
comment
Keywords: Cant Castler, Castling (wgsk)
Genre: h#, Retro
Computer test: HC+ Popeye C-Version 3.54 & trivial retro-logic
FEN: r1q1kn1r/1p2p2p/bp2P3/bp1pB3/1p6/4p3/1PP1P2P/R3K3
Reprints: 71 Europe Echecs 41 01/1962
Input: Gerd Wilts, 1995-06-03
Last update: A.Buchanan, 2021-11-23 more...
19 - P0001263
Luigi Ceriani
135 Europe Echecs 78 06/1965
P0001263
(15+11)
h#2
1. bxa5 Sb6 2. axb4 Ta8#
play all play one stop play next play all
Henrik Juel: analysis
Black captured a3xb2 (not a7xb6, as the SW corner then cannot be released)
The missing white man is [Pe2], which must have promoted on b8
Now the white captures can be seen: axb, exdxcxb, and [Lf8] on f8
The only free white officer is Dd8, so the retroplay must include unpromotion of it on b8
It is impossible to shield Ke8 from check by Db8, so Ke8 has moved and Black may not castle (2022-04-27)
Henrik Juel: solution
1.bxa5 Sb6 2.axb4 Ta8#
not 1.0-0? Lxb6 2.Txa8 Txa8
HC+ Popeye 4.61 (2022-04-27)
Henrik Juel: Andrew, I should have been more precise in recommending your suggestion to introduce the new keyword Organ Pipes
I was talking about the standard meaning of Organ Pipes, which dates back to Sam Loyd, 1859: black LTTL in a problem where each L interferes with each T and vice versa, forming four Grimshaws
You have probably used a pattern search to find all occurrences of black or white LTTL, and you have neglected the interferences
This problem has LTTL in the wrong color, and there are no interferences (2022-04-27)
James Malcom: I agree with Henrik, although White organ pipes still exist, few and far. (2022-04-28)
James Malcom: Until Andrew bumbles back, I've updated the English definition to "A problem *utilizing* the pattern bishop, rook, rook, bishop in a straight line to create multiple Grimshaws." This specifies the purpose and formation of the Organ Pipes. (2022-04-28)
Henrik Juel: Thanks, James (2022-04-28)
A.Buchanan: Hi Guys. I wondered about other uses of organ pipes. In a first pass population of the 500-odd records with the current PDB interface, there’s not much time for thinking. I have multiple tabs open to eliminate the wait time associated with PDB refresh. One could go back and eliminate those which are not grimshaws. Alternatively (and this would be my preference) accept that this is a visual pattern which may occur in non-Grimshaw context. Then use the keyword Grimshaw(4x) to indicate when it’s really Grimshaw. I think it can also occur with white pattern, as a problem by Pal Benko shows. So I blitzed through the d# records, if someone wants to complete the rest that’s good. I think the existence of a few false transient positives is an acceptable price even if we take the narrow definition of Organ Pipes. (2022-04-28)
A.Buchanan: By the way, these days Deepl is good enough to give us decent translation to Feench & German and as a matter of policy whenever I make a change to a definition I try to align the other two. Other users who do not maintain the glossary are encouraged to propose definitions where there is a gap (E.g. Grimshaw). There are a lot of undefined terms, many very recent. (2022-04-28)
A.Buchanan: Hi Henrik feel free to respond to my response to your message, say thanks for the tags I have added, or add tags yourself to complete the work. The ones I’ve added were mostly the d# and it was a deliberate decision to blast through as a first pass and just add them for now anyway, not “neglected”. Now, I really don’t feel like continuing (2022-04-28)
Henrik Juel: Andrew, I believe that most PDB users appreciate your contributions to the site
I certainly do, so please continue your good work (2022-04-28)
comment
Keywords: Castling (sk), Cant Castler, Promotion
Genre: h#, Retro
FEN: N3k2r/2pppppp/1p6/B7/RP5P/RP3P2/BpPP2P1/NK1Q3r
Input: Gerd Wilts, 1995-06-03
Last update: James Malcom, 2022-04-28 more...
20 - P0001285
Jean Oudot
370 Die Schwalbe 07/1960
R. Bédoni gewidmet
P0001285
(11+11) C+
h#2*
* 1. ... dxc5 2. Dxh3 0-0-0#
1. Db2 Le2+ 2. Kc2 Ld1#
play all play one stop play next play all
Henrik Juel: the five missing black men were captured by white pawns (exfxgxh, fxgxj), so with Black to move last move was with Ta1 or Ke1, and White may not castle
C+ Popeye 4.61 (2022-11-26)
A.Buchanan: Pleasant White tempo play in both phases (2022-11-27)
comment
Keywords: Cant Castler (wl), Castling (wl)
Genre: Retro, h#
Computer test: HC+ Popeye 4.61 with simple retro logic
FEN: 8/7p/7P/2pr2pP/2bP2Pb/2pk1BRP/6qN/R3KNrn
Reprints: 553 FIDE Album 1959-1961 1966
157 Europe Echecs 90 07/1966
(10) diagrammes 112 01-03/1995
Input: Gerd Wilts, 1995-06-03
Last update: A.Buchanan, 2022-11-27 more...
21 - P0001345
F. ben Galuth
219 Europe Echecs 126 07/1969
P0001345
(10+10) C+
h#2
1. cxd3ep? g3 2. Bd4 Lg2# (wLc1 retro-blocked)
1. c5! g3 2. cxd4 Lg2#
play all play one stop play next play all
James Malcom: Solution? (2020-12-29)
Hans-Jürgen Manthey: wohl beabsichtigt: 1. cxd3ep g3 2. Bd4 Lg2#
doch auch ein Dual: 1. c5 g3 2. cxd4 Lg2# (2020-12-29)
Mario Richter: rawbats says, that f2-f4 is White's only legal last move, and 1. c5 g3 2. cxd4 Lg2# the only solution, so 1. cxd3ep g3 2. d4 Lg2# might be a try, intended to fool the solvers ...
(notice that R: 1. d2-d4?? excludes wLc1 from the set of objects available for black pawn captures!) (2020-12-29)
James Malcom: Furthermore, the h pawn couldn't have moved last, as that rook is needed for the Black pawns. wPe6 takes all remaining captures of Black pieces, and bBf8 could never escape. So the d2 and h2 pawns did not move last, and gxf3ep does nothing, meaning that all en passants are not a solution. A very fine retro "joke." (2020-12-29)
James Malcom: This means that 1. c5 g3 2. cxd4 Bg2# is the only possible solution, and as such, this should be C+? (2020-12-29)
Hans-Jürgen Manthey: Stimmt Mario
habe übersehen das der Lc1 nicht gezogen haben kann bei 1. d2-d4; aber 6 schlagfälle von Schwarz nötig sind. (2020-12-30)
A.Buchanan: I think the composer here is exactly "trolling solvers' preoccupations for the lulz", to quote Hauke Reddmann in MatPlus today, which even my allegedly fine German skills are unable to translate. But Hauke is German, so maybe he could do it? (2020-12-30)
more ...
comment
Keywords: En passant, En passant as key (Tries)
Genre: h#, Retro
Computer test: Popeye v4.85 + thinking
FEN: 8/2p1p1p1/4P3/3p1p2/2pPkPpP/4p3/1PP1P1P1/Kb3B2
Input: Gerd Wilts, 1995-06-03
Last update: Gerd Wilts, 2021-01-06 more...
22 - P0001418
Janko Furman
292 Europe Echecs 194 01/1975
2. ehrende Erwähnung
P0001418
(14+14)
Löse die Stellung auf!
h#1 wer?
White has the move, so 1. ... Sxb6! Sa6# not 1. e5? Tcxd6#
R: 1. ... Lh1-f3 2. g3-g4 h2-h1=L 3. g2-g3 h3-h2 4. d4-d5 h4-h3 5. d3-d4 h5-h4 6. d2-d3 h6-h5 7. h5xDg6 Db1-g6 8. h4-h5 Da1-b1 9. h3-h4 a2-a1=D 10. h2-h3 a3-a2 11. c3-c4 a4-a3 12. c2-c3 a5-a4 13. b3-b4 a7-a5 14. a6xLb7 etc.
R: 1. ... Th8-f8 2. g3-g4 Sf8-h7 3. g2-g3 Th1-h8 4. d4-d5 Ta1-h1 5. d3-d4 a2-a1=T 6. d2-d3 a3-a2 7. c3-c4 a4-a3 8. c2-c3 a5-a4 9. b3-b4 a7-a5 10. a6xSb7 etc.
play all play one stop play next play all
bPa & bPh promoted with wPaxb7 & bPhxg6 to clear the way. There is an inner cage containing all the White officers, and White has only 6 current retractions. So Black must unpromote something. Both White captures were on light squares, so Lf4 is original.
1) Lf3 to h1. For speed, White must uncapture Ceriani-Frolkin D on g6, which arrives on a1 in 2 unmoves. When White uncaptures L on b7, it's Phoenix of the original.
2) Tf8 to a1. (It's too slow to unpromote on h1). When White must uncapture on b7, g2 is already blocked, so White must uncapture S, and unpromote that or one of the other Black S. However, the inner cage has been breached, so the clock has stopped now. Eventually, White will uncapture Phoenix T on g6.
The tempo is exact in both cases, so White, not Black must have the move.
A.Buchanan: Very nice - exactly matching lengths across the two lines and complementary non-trivial forward play (2021-10-21)
Henrik Juel: HC+ Popeye 4.61 (2021-10-22)
more ...
comment
Keywords: Allumwandlung (ld,ts), Whose move?, Phoenix (l,t), Ceriani-Frolkin Theme (d)
Genre: h#, Retro
FEN: NnQNqr2/1PKRPprn/1pRppkP1/2pP2p1/1PP2bP1/5b2/5P2/8
Input: Gerd Wilts, 1995-06-03
Last update: A.Buchanan, 2021-10-21 more...
23 - P0001420
Branko Koludrovic
294 Europe Echecs 197 04/1975
1. ehrende Erwähnung
P0001420
(13+12) C+
h#3*
1. ... cxd6ep 2. 0-0-0 0-0-0 3. Kd7 Sa7#
1. De7 c6 2. Th8 c7 3. Tf8 Sd6#
play all play one stop play next play all
White pawn caps: axb,dxe,gxf,hxg.
Black: fxg,bxc,cxb.
wPb4 came from b3 to release wBa3, so bPb3 captured to reach that square.
All pcs accounted for means bPd never captured.
In the set play, there are 13 retro tries in which one or both players do not castle. The intention is that both castling rights are needed in order to imply the pawn double hop.
A.Buchanan: White pawn caps: axb,dxe,gxf,hxg definite.
Black: fxg and two to resolve c-file. But that may be c&d cross-capture, so in set play last move might have been c6xd5. So I think this problem is cooked. What am I missing? (2022-03-21)
Mario Richter: If Black's last move was c6xd5, how do you get white Bishop a3 out of his cage? (In that case, black pawn b7 never left the b-file). (2022-03-21)
A.Buchanan: I agree Mario thanks (2022-03-21)
Hans-Jürgen Manthey: beide Rochaden und im Satz ep. ist möglich:
R.: 1. ... d7-d5 2. Sa7-b5 b5xTc4 3. Tg4-c4 Dd8-h4 4. c4-c5 Th8-h3 5. c2-c4 c4xDb3 6. Dd3-b3 h3-h2 7. Dd1-d3 h4-h3 8. Sc8-a7 h5-h4 9. Th4-g4 c5-c4 10. Th1-h4 h7-h5 11. h2xSg3 Sf5-g3 12. b3-b4 Sh6-f5 13. Lb4-a3 c7-c5 14. Ld2-b4 g3-g2 15. Lc1-d2 a3-a2 16. a2xSb3 Sc5-b3 17. Lb3-a4 Sa6-c5 18. Lc4-b3 Sb8-a6 19. d2xLe3 Lc5-e3 20. Ld3-c4 Lf8-c5 21. Lf5-d3 a4-a3 22. Lh3-f5 a5-a4 23. Lf1-h3 g4-g3 24. g2xLf3 Lb7-f3 25. Sb6-c8 Lc8-b7 26. Sa4-b6 a7-a5 27. Sc3-a4 b7-b5 28. Sb1-c3 f5xSg4 29. Se5-g4 Sg8-h6 30. Sf3-e5 e7-e6 31. Sg5-f3 f6-f5 32. Sh3-g5 f7-f6 33. Sg1-h3 (2023-02-23)
comment
Keywords: a posteriori (AP) (Type Petrovic), En passant as key, Castling (wgsg)
Genre: h#, Retro
Computer test: HC+ Popeye v4.87 & retro-logic.
FEN: r3k3/6p1/4p3/1NPp4/BPp4q/Bp2PPPr/pP2PPpp/R3K3
Input: Gerd Wilts, 1995-06-03
Last update: A.Buchanan, 2022-03-21 more...
24 - P0001450
Henri Nouguier
324 Europe Echecs 223 07/1977
P0001450
(13+3)
h#1
hans: 1. Th2xf2 De1xf2#!
1. Th2xh3 0-0#? (Castling illegal)

R: -1. Kf1xSe1 Sg2-e1 -2. Ke1-f1 Sf4-g2+ -3. Tg1-h1 Sd5-f4 -4. Tf1-g1 Se3-d5 -5. Tg1-f1 Sf1-e3 -6. Tg1-h1 Tg2-h2 -7. Th2-h1-g1 Tg1-g2 -8. Th1-h2 g2-g1=T -9 h2-h3 h3xSg2 (2010-06-26)
HENRI: Hans solution has a tipo. Their is no D (=Queen in german) mating. One should read the solution in german : 1. Th2xf2 Ke1xf2#! 1.Th2xh3 0-0#?
or in english : 1. Rh2xf2 Ke1xf2#! 1.Rh2xh3 0-0#? (Castling illegal) (2023-10-25)
comment
Keywords: Cant Castler, Castling (wk)
Genre: h#, Retro
FEN: 8/8/8/8/8/PPP3PP/B1rPPP1r/BNk1K2R
Input: Gerd Wilts, 1995-06-03
25 - P0001453
Luis Alberto Garaza
327v Europe Echecs 225/226 10/1977
P0001453
(11+14) C+
h#1
1. hxg3ep! 0-0#! (Tf1#?)
play all play one stop play next play all
Hans-Jürgen Manthey: 1. hxg3 O-O#(auch Tf1# !) mögliche zugfolge:
1. d2-d4 f7-f5 2. Lc1-d2 f5-f4 3. Ld2-c1 e7-e6 4. Lc1-d2 c7-c5 5. Ld2-c1 b7-b5
6. Lc1-d2 a7-a6 7. Ld2-c1 Lc8-b7 8. c2-c4 Ke8-f7 9. Lc1xf4 Sg8-f6 10. Lf4-h6 g7xh6
11. Dd1-d3 h6-h5 12. Dd3-d1 Lf8-h6 13. h2-h4 Lh6-g5 14. h4xLg5 h5-h4 15. Dd1-d3 h7-h5
16. Dd3-d1 b5xc4 17. a2-a4 Lb7-e4 18. d4-d5 e6xd5 19. Ta1-a3 Le4-c2 20. f2-f4 Th8-e8
21. Ta3-g3 Te8-e3 22. f4-f5 Sf6-e4 23. f5-f6 Kf7-e6 24. f6-f7 Se4-d2 25. Sg1-h3 Sd2xf1
26. Sh3-f2 Sf1-d2 27. Sf2-g4 Sd2-e4 28. Dd1-d3 Lc2-d1 29. Sg4-f6 c4-c3 30. Sf6-h7 c3-c2
31. b2-b4 c5-c4 32. b4-b5 Sb8-c6 33. b5-b6 Sc6-a5 34. b6-b7 c2-c1L 35. g5-g6 c4-c3
36. Sb1-a3 c3-c2 37. Sa3-b5 Ta8-c8 38. b7-b8L Tc8-c3 39. Lb8-d6 La5-c4 40. a4-a5 a6xSb5
41. a5-a6 Ke6-f5 42. a6-a7 Dd8-e8 43. a7-a8S De8-e5 44. Sa8-c7 Lc1-b2 45. Sc7-e6 Lb2-a1
46. Se6-f8 Kf5-f4 47. Tg3-g5 Se4-f6 48.Ld6-a3 Te3-h3 49. La3-b2 Sc4-a3 50. g2-g4 (2020-12-11)
Henrik Juel: The white castling is needed to prove (a posteriori) that last move was g2-g4, legitimizing the ep capture (2020-12-31)
more ...
comment
Keywords: En passant as key, Castling (wk), a posteriori (AP) (Type Petrovic)
Genre: h#, Retro
Computer test: HC+ Popeye v4.87 & simple retro-logic
FEN: 5N2/3p1P1N/5nP1/1p1pq1Rp/5kPp/n1rQ3r/1Bp1P3/b2bK2R
Input: Gerd Wilts, 1995-06-03
Last update: A.Buchanan, 2022-03-22 more...
26 - P0001561
Michel Caillaud
432 Europe Echecs 304 04/1984
Lob
P0001561
(8+13)
Kürzestes h#?
1. ... bxa6ep! 2. Txe2 axb7 3. Td2 b8=D 4. De2 Db1#
1. Dc6? bxc6 2. Ld7 cxd7 3. Ke1,Kxe2 d8=D 4. Kf1 Dd1# too slow
R: 1. a7-a5 e3xLf4 2. Ld6-f4 d2xTe3 3. Ta3-e3 b4-b5 4. Ta6-a3 b3-b4 5. Tc6xBa6 a5-a6 6. Tc8-c6 a4-a5 7. Tg8-c8 a3-a4 8. Lf8-d6 a2-a3 9. e7xDf6,e7xTf6
play all play one stop play next play all
The forward play gives shortest h# in 3.5 moves, *if* e.p. is on. Otherwise the shortest is a mildly dualized 4.0 moves. So our mission is clear: assuming that Black moved last, what was that move? The position looks very open - how can one ever know? There is a knot in the south-east corner, which will only be resolved by wQ/R visiting g1.

White pawns have captured dxexf & hxg, accounting for BSP. So wPa was waylaid while wPc promoted to B. Black pawns captured dxexfxgxh, exf as well as wBf1 captured at home and wPa. That's 7 units, but bPc must also have captured to allow wPc to promote (to B on c8). So all captured units are accounted for. bPc promoted (bPh7 came from h6, so is bPd). We can't undo any of the Black pawn captures now, but we can undo d2xe3xf4, to release 2 Black units.

We can unpromote bPc, but we can't undo its capture to release a White unit, until wBh3 has unpromoted. The only White unit we can get now is from e7xf6. So this means that the two Black units we can release must be B & R, so they can retreat to f8 & g8 respectively. The timing is very tight, and there is only one way to do it. The black rook must visit a6 to unwaylay wPa, which gives White 5 more tempi, just enough.

bPc promoted on b1 to R, so we would need to undo the cage to get that. Therefore all we can do is send wQg1, then Rf1-f2 f2-f3 etc. bPa must retract immediately tp a7, so that wPa when unwaylaid can make fully 4 unmoves. Black officers are arranged to give unique retro & forward play (although with minor dual for the try) with Sh8 not just retro dressing but ensuring unique retraction Tg8-c8 not Th8-c8.
A.Buchanan: Another surprising motivation for e.p. I love Sh8. Sorry for W3 dual in the try play else the stipulation could be h#4*. I can’t fix it but it’s hard to improve on MC. Very enjoyable (2021-10-23)
A.Buchanan: Non-standard material is where the diagram contains for a player more than 1 queen or bishop of a hue or more than 2 rooks or knights. Obtrusive material is standard material, but there is some cheap reason why the unit must be promoted, most commonly that a bishop's home square remains blockaded by pawns. These categories are disjoint: no piece is ever both. Many problems in PDB do not apply this consistently, but the distinction goes back a long way in chess problem history, and is discussed by Morse.
Honestly, I dislike the word "obtrusive" whose negativity (while perhaps valid in forward problems) is inappropriate for retros. One distinguished composer objected to this tag being assigned to one of his problems. Nevertheless the concept has some interest. Renaming is a perilous business, but I am looking for suggestions... :) (2021-10-24)
A.Buchanan: Another distinction that comes to mind between "non-standard material" and "obtrusive promotion" is that normally in the former, one can't immediately point to which of the non-standard pieces was promoted: it's just that there's too many; while for the latter, one can usually point to a specific unit immediately.
"Obvious" is a candidate replacement for "obtrusive", but this might commit a cardinal sin of trying to nail down a perfectly useful and inherently vague term. Both "obvious" & "obtrusive" begin with "ob" which is helpful. What do you think? (2021-10-24)
Henrik Juel: I am curious about the award; why did this problem with good forward play and excellent retro-play only obtain a Commendation?
Maybe the judge did not like that Lh3 obviously is obtrusive...
. (2021-10-24)
A.Buchanan: Yes, and the lovely P0001117 with a similar obtrusion only received 12th Lob! But there may have been other factors. I do observe that "non-standard material" is arguably a worse defect than "obtrusive material", but the term doesn't cause offence because it's objective and non-judgemental. "obtrusive" is more subjective and inherently pejorative. To have such terms in the glossary is to put curators in an invidious position. I think the concept has its place, but I would like to replace it with something less scornful. I'm up for "obvious". It's an easy and reversible change: let's do it and see if mobs of protesters form outside the gate :) (2021-10-24)
A.Buchanan: Have changed "obtrusive material" to the non-pejorative "obvious promotion". It may still be regarded as a defect. As a placeholder, I have also changed the unclear German "mit Umwandlungfigur(en)" to "augenscheinlich Umwandlungfigur". A native German speaker I'm sure will propose a better term. (2021-10-26)
A.Buchanan: “offensichtlich” it is thanks Mario (2021-10-26)
more ...
comment
Keywords: En passant, Last Moves?, Obvious promotion (L), En passant as key, Promotion (D)
Genre: h#, Retro
FEN: 7n/1p3pp1/4bp1p/pP6/4qPP1/5PrB/4PrPp/3k3K
Input: Gerd Wilts, 1995-06-03
Last update: A.Buchanan, 2021-10-24 more...
27 - P0001937
Werner Kuntsche
6025 Schach , p. 191, 06/1969
2. Preis
P0001937
(4+12) C+
h#2
1. exd3ep Lxg4 2. f3 Le6#
play all play one stop play next play all
Klären: Quelle = Schachmatt? - Felber, Volker: SCHACH ist korrekt, 6/1969, Seite 191 (2010-10-09)

Eliminierung von 4 Steinen zwischen sK und wT im h#2 (mit Hilfe von ep-Schlag)!
more ...
comment
Keywords: En passant as key
Genre: h#, Retro
Computer test: HC+ Popeye 4.61
FEN: 8/8/8/1pn4b/2kPpppR/2q1rB1p/1p4r1/3K4
Input: Gerd Wilts, 1995-06-03
Last update: A.Buchanan, 2023-01-09 more...
28 - P0001970
Thomas R. Dawson
26 Schachkongress Teplitz-Schönau im Oktober 1922 1923
P0001970
(13+13)
h#1
1. exf3ep Lc2#
R: 1. f2-f4 f5xTe4 2. Tg4-e4 & e.g. f6-f5 3. Tg1-g4 f7-f6 4. Le4-h7 h5-h4 5. Lg2-e4 h6-h5 6. Lf1-g2 h7-h6 7. g2xh3
play all play one stop play next play all
1.exf3e.p. Bc2# is the only possible solution, but this necessitates R: 1.f2-f4. Can we prove this?

(13+13) with 1+2 pawn captures. Bf8 captured at home, so to satisfy White appetite, the missing Black pawn (a or b) must have promoted via c2 on c1. Two more White units must be captured to allow this.

The kings cage can only be unlocked by retracting WPc2. But the clock is ticking as there are only 6 black moves which can be retracted.

The promoted piece was captured on e3 or h3. If either capture is undone, then a White bishop square is cut off, so WB must be replaced prior to this.

Now the order of the early moves is: WdP moves, WQB & WQR escape, BP promotes on c1 to X (capturing WR at some point), X captured by WP.

So the first White capture must be dxNe3 and the second White capture releases gxXh3. The second White capture releases WKB & WKR. WKR captured by original BfP.

The clock starts ticking with gxh3. Black has 6 pawn moves. WKB has 3 moves to reach h7. WR has 3 if it goes via d file, or 2 if it starts on g1 (in which case WfP or WQB must also move once). So certainly at least 6 White moves. Last move was therefore White (even if the stipulation didn't tell us), and it can only have been WfP coming from f3 or f4. If it had been coming from f3 it would have blocked WKB in its progress, so the last White move was indeed R: 1.f2-f4.

WKR did therefore move from g1-g4-e4, and R: 1. ... fxRe5 2. Rg4-e4. Prior to that, move order not unique, but counting still exact.

Note that WN loitering on b4, pretending to be part of the cage, is present on the board just to make up the numbers.
Jeliss: "Obstruction of passage square f3 to Bishop of same colour."

"Version 'Pittsburgh Leader' 08.06.1913"
Yoav Ben-Zvi: Appears as the first problem (D445) in the booklet on Dawson's RA problems by G.P. Jellis. The obstruction that occurs in the Try -1.Pf3-f4?, by WP of WB, is described as "obstruction of passage square". It is not considered by Dawson and his disciples to be a Retro opposition. Dawson's conception of RO was quite broad, it included cases where the interference was not by occupation of the target square, so the only valid reason that I can see to exclude this case is that the 2 pieces involved are both of the same color. Fabel's definition explicitly excludes "Monochrome RO". I conclude that it would be preferrable to interpret RO as a bi-chromatic interference. The keyword Retro opposition should be removed. (2018-04-07)
A.Buchanan: To my mind, RO involves some kind of parity-tempo issue between the sides, not just some kind of race-tempo. If it was just about "bi-chromatic interference", one might say that bPe4 blocks wBh7 from an immediate retreat, so it has to be wPf4 that retreats first, legitimizing the ep key. So I agree this is not RO. (2024-01-06)
more ...
comment
Keywords: Last Moves?, En passant as key
Genre: h#, Retro
FEN: nqb5/1rrpp1pB/KRp5/1p4B1/kN2pP1p/2P1P2P/PP2P2P/8
Reprints: D445 Retro-Opposition & Other Retro-Analytical Chess Problems 1989
Input: Gerd Wilts, 1995-06-03
Last update: A.Buchanan, 2024-01-06 more...
29 - P0002476
Werner Kuntsche
1777v Problem 73-78 06/1961
P0002476
(5+7) cooked
h#3 AP
Einheitslösung!
1) 1. cxd3ep Ta5 2. Sf5 0-0 3. Sd4 Txe5#
2) 1. gxf3ep Th5 2. Sd5 0-0-0 3. Sf4 Txe5#
play all play one stop play next play all
Cook: 707 cooks which do not begin with e.p., e.g. 1.Sg3-e2 Ra1-a2 2.Ke4-d3 Rh1-h3 3.e5-e4 Ra2-d2 #
What is the obscure intent behind "Einheitslösung!"? (= "One size fits all solution"?)
Mario Richter: Im Nachdruck in problem 144-147 steht nur auf e5 ein sB (also keiner auf e3). Allerdings gibt es in allen drei Varianten (sBe5+e3, nur sBe5, nur sBe3) jede Menge NL, die ohne Rochade und e.p.-Key auskommen - am wenigsten dann, wenn nur auf e3 ein sB steht. (2010-10-10)
VL: Cf. P0002475. (2012-08-26)
A.Buchanan: Possibly "Einheitlösung!" is a request to unify retro implications derived across separate parts of the PRA solution. But cooked chessically irrespective of these philosophical issues, which still wait to be resolved 60 years on. (2022-02-16)
comment
Keywords: a posteriori (AP) (Type Petrovic - ccee), En passant as key (2), Castling (wb), Symmetrical position, Symmetrical solution, Superseded by (P1401449,P1399112)
Genre: h#, Retro
Computer test: C- Forward play cooked by Popeye v4.87. However the basic idea is good, if a suitable theoretical framework can be found.
FEN: 8/8/8/4p3/2pPkPp1/2n1p1n1/8/R3K2R
Reprints: (66) Problem 144, p. 52, 12/1971
Die Schwalbe 99 06/1986
Input: Gerd Wilts, 1995-06-03
Last update: A.Buchanan, 2022-05-24 more...
30 - P0002850
Michal Dragoun
2974 Sachova skladba 04/1992
P0002850
(6+12) C+
h#2
b) -1w, dann h#2
sDU=Grashüpfer
a) 1. Ka1! gxh7 2. Ga2 h8=G#
b) R: 1. c3-c4, dann Ga6! g7 2. Ga1 g8=G#
play all play one stop play next play all
Witzige Zwillingsbildung mit nettem Umwandlungswechsel.
Henrik Juel: part a) is C+ Popeye 4.61 (2021-05-02)
A.Buchanan: Henrik last move can’t be b2-b4 because of caging bK behind White lines. So must have been b3-b4. Can you try Popeye with halfduplex for b)? (2021-05-03)
Henrik Juel: Thanks to Andrew's observation, the entire problem is now C+ Popeye 4.61 (2021-05-03)
comment
Keywords: Help retractor
Pieces: du = Grasshopper (G)
Genre: Retro, h#, Fairies
Computer test: part a) is C+ Popeye 4.61 part b) is HC+ Popeye 4.61 + trivial retro thought
FEN: 8/7n/6P1/1p3ppp/2P5/1p1*2qp3/kP2*2qPP1/2Kb1n2
Input: Gerd Wilts, 1995-06-03
Last update: A.Buchanan, 2021-05-03 more...
31 - P0003100
John Derek Beasley
Cambridge 1985
P0003100
(3+8) C+
h#2
Zeroposition
a) -sBe4
b) -sSf5
a) 1. Sg3 Kxh4 2. Le4 hxg3#
b) 1. hxg3ep Kh4 2. Lf5 hxg3#
play all play one stop play next play all
Henrik Juel: in a) last move could be Kg2-h3
in b) last move must be g2-g4 (2024-01-14)
more ...
comment
Keywords: En passant as key, Zeroposition
Genre: h#, Retro
Computer test: HC+ Popeye 4.61 & trivial retro-logic
FEN: 8/7b/8/4pn2/4pkPp/4pp1K/7P/8
Input: Gerd Wilts, 1995-06-03
Last update: A.Buchanan, 2024-01-15 more...
32 - P0003122
Albert Zickermann
789 Kieler Neueste Nachrichten 18/03/1934
P0003122
(3+6)
h#2
1. fxg3ep Lxg7 2. Lg4 Lf6#
play all play one stop play next play all
A.Buchanan: Sg7 would be better than T, from both forward & retro perspectives I think? (2021-04-14)
comment
Keywords: En passant as key
Genre: h#, Retro
FEN: 7B/6rp/7K/8/5pPk/5b1r/8/8
Input: Gerd Wilts, 1995-06-03
33 - P0003132
Frank Bentler
10997 Schach-Echo , p. 228, 06/1985
P0003132
(9+13) C+
h#2
1. axb3ep Texf1 2. b2 Kd2#
play all play one stop play next play all
more ...
comment
Keywords: En passant as key, En passant
Genre: h#, Retro
FEN: n7/2ppp1p1/1p6/B7/pP6/5PP1/p4pPP/rk1KRrRn
Input: Gerd Wilts, 1995-06-03
Last update: Alfred Pfeiffer, 2021-02-11 more...
34 - P0003138
Branko Koludrovic
4208 Problem 12/1979
P0003138
(11+9)
h#3 (AP)
0.1...
1. ... axb6ep 2. 0-0-0 0-0-0 3. Td7 a8=D#
play all play one stop play next play all
This problem can be solved presuming that both castlings are executable. It implies that the last halfmoves have been.-1.Bb6xa7 sBb7-b5.
Bb5 could not come from c6 (sBc6xFb5), because all of the missing white pieces(5) have been captured on white squares except wLc1,which must have been captured by Bc7(Bc7xwLb6).Further,it has not played immediately before -1.Bb6-b5 because this would prevent white castling(forcing wT or wK to move)
Here is the proof that 4 remaining white pcs have been captured on white squares:
black d-pawn must have been captured on d4 and h,g f pawns had to promote after Bh3xwFg2,then wBh2moved to h4, enabling Bg4x Fh3 ..h1T, and finnaly sBf7
after fxg2... g1S. sLf1 prevented the check from Th1, and later has been captured by a white officer.
Conclusion:
Retro:-1.wBb6xFa7 Bb7-b5
Forward(AP) 1....axb6 e. p.2.0-0-0 0-0-0 3.Td7 a8Q# (Author)
Branko Koludrovic: P.S.
The black pawn a4(on the diagramm)came from c7 after capturing the white bishop on b6, then moving to b5 and capturing a white officer on a4. (2010-09-28)
A.Buchanan: I think sBf must have captured on g2 & then promoted on f1 to S, otherwise it’s impossible to unlock the southeast cage (2023-07-11)
A.Buchanan: So that means it’s sound. However I don’t see that it needs to have been sL shielding on f1 (2023-07-11)
more ...
comment
Keywords: a posteriori (AP), En passant as key, Castling (sgsgwg), Promotion (D), Valladao Task
Genre: h#, Retro
FEN: r3k3/P3p3/p7/Pp6/p6P/2P3PR/2PPP2b/R3K1nr
Input: Gerd Wilts, 1995-06-03
Last update: A.Buchanan, 2023-07-11 more...
35 - P0003142
Marco Bonavoglia
3652 Sinfonie Scacchistiche 01-03/1982
P0003142
(6+5) C+
h#2
b) +sTd4
a) 1. Sxb4 Sxc3 2. Ka5 Sxc4#
b) 1. cxb3ep Sc4 2. Sb4 Sxc3#
play all play one stop play next play all
Henrik Juel: In a) last move could be Kb2-a1 (2022-12-05)
more ...
comment
Keywords: En passant as key, Cant ep, Zilahi
Genre: h#, Retro
Computer test: HC+ Popeye 4.61
FEN: 8/8/8/1P6/kPp5/N1p5/n1pP4/KN6
Input: Gerd Wilts, 1995-06-03
Last update: A.Buchanan, 2022-12-06 more...
36 - P0003143
Zvi Roth
947 Shahmat 08/1972
P0003143
(6+7) C+
h#2*
*) 1. ... Kxb3 2. Txd5 Txd5#
1) 1. bxc3ep Kxb3 2. Txd5 Lxc3#
play all play one stop play next play all
A.Buchanan: Attractive. It reminded me of P0003347, which I have just discovered is sound as set play. (Just omit the first move!) Here part of the interest, although the intermediate moves are the same, the mating move is different, but in both cases has to cover d4 & e5. (2020-12-09)
more ...
comment
Keywords: En passant as key
Genre: h#, Retro
Computer test: C+ Popeye 4.61
FEN: 8/8/1P6/BRrP4/KpPkp3/1r1nq3/8/8
Input: Gerd Wilts, 1995-06-03
Last update: A.Buchanan, 2020-12-09 more...
37 - P0003151
Laszlo Talaber
Tivadar Kardos

3649 Arbejder-Skak 01/1954
1. Preis
P0003151
(9+16) C+
h#2.5*
* 1. ... ... 2. fxe3ep Sge4 3. Txe4 Sf3#
1. ... e5 2. dxe5 Sde4 3. fxe4 Sf5#
play all play one stop play next play all
Sally: Der letzte Zug war e2-e4! (2010-04-06)
A.Buchanan: Great harmony between the phases (2021-11-23)
comment
Keywords: En passant as key
Genre: h#, Retro
Computer test: HC+ Popeye C-Version 3.52 & trivial retro-logic
FEN: 8/7p/3p3p/2nq1ppP/p1PkPp1r/P1p3N1/n2N2Pb/1b1r1B1K
Reprints: (XX) Problem 37-40 09/1956
Input: Gerd Wilts, 1995-06-03
Last update: A.Buchanan, 2021-11-23 more...
38 - P0003180
Jozsef Korponai
6875 Arbejder-Skak 02/1968
P0003180
(12+12)
h#2
1. hxg3ep Sc1 2. gxf2 Se3#
play all play one stop play next play all
Welches ist die Originalquelle? Oder wurde es zweimal als Urdruck gebracht?

vgl. P0003291
Henrik Juel: White pawns captured all four missing black men
Last black move was not b7xPa6 ([Ta8]) nor e7-e6 ([Lf8]), so it was Lb7-c8, which is possible only if last white move was g2-g4
C+ Popeye 4.61 and above analysis (2021-04-23)
comment
Keywords: En passant as key
Genre: h#, Retro
FEN: 2b5/p2p2pp/p3p2b/6Pr/5BPp/2P5/NRP1PP1P/3Nrk1K
Reprints: F84 The Problemist 05/1970
Input: Gerd Wilts, 1995-06-03
Last update: Mario Richter, 2021-04-23 more...
39 - P0003187
Thomas R. Dawson
6577 British Chess Magazine 12/1944
P0003187
(12+14)
h#1.5
1. ... bxc6ep 2. b5 Txa3#
R: 1. ... c7-c5 2. f4xTg5 Tg6-g5 3. f3-f4 Ta6-g6 4. f2-f3 Sc6-a7 5. h4-h5 Ta8-a6 6. h3-h4 a6-a5 7. h2-h3 Ka5-a4 8. b3-b4 Ta4-a3
or R: 1. f4xTg5 Tg6-g5 2. f3-f4 Ta6-g6 3. f2-f3 Sc6-a7 4. h4-h5 Ta8-a6 5. h3-h4 a6-a5 6. h2-h3 Ka5-a4 7. b3-b4 Ta4-a3
but we are told that White has the move.
play all play one stop play next play all
Henrik Juel: Stipulation should probably be interpreted to mean h#1.5 . -1... c7 -2.f4xTg5 Tg6 -3.f3 Ta6 -4.f2 Sc6 -5.h4 Ta8 -6.h3 a6 -7.h2 Ka5 -8.b3 Ta4 etc. (2004-03-18)
A.Buchanan: Idea I suppose is that Black just moved, and moreover it was pawn double hop. I don't see clearly far enough back to confirm that. (2021-10-21)
Mario Richter: If Black moved last, the only legal retraction is R: 1. ... c7-c5! But the position can also be resolved if White moved last, starting with R: 1. f4xTg5 Tg6-g5 2. f3-f4 Ta6-g6 3. f2-f3 Sc6-a7 4. h4-h5 Ta8-a6 5. h3-h4 a6-a5 6. h2-h3 Ka5-a4 7. b3-b4 Ta4-a3.
Therefore the keyword "No legal last move for White" doesn't apply here; and to make the solution work, the Stipulation should be something like "h#2 0.1.1." or "h#1.5". (2021-10-21)
Henrik Juel: HC+ Popeye 4.61 (2021-10-22)
A.Buchanan: To my mind, the retro thinking is serious enough that we can’t claim HC+. If a tool like Replicator 2.0 was to validate the retro status I’d be ok. But I think it would be misleading to put HC+ to indicate the trivial h#1.5 is sound. We can certainly mark that Popeye has checked the h#, but not click the C+ flag. What do you think? (2021-10-22)
Henrik Juel: HC+ is not recognized (yet) by PDB, and I suppose it just meant that some sort of human analysis is needed in addition to a standard forward solver like Popeye
You are suggesting that we distinguish between easy and difficult human analysis and say HC+ for easy analysis only
I suggest saying HC+ for all problems where Popeye is not sufficient; you can write easy or difficult in the C+ comment
If and when a human analyst proves the analysis wrong, we delete the HC+ label, of course (2021-10-22)
A.Buchanan: What does Mario think? He invented the term, I seem to remember (2021-10-22)
comment
Keywords: En passant as key
Genre: h#, Retro
FEN: 2b5/np1p2pp/7q/pPp3PP/kP6/r1PPP3/RKpp2P1/BB2n3
Input: Gerd Wilts, 1995-06-03
Last update: A.Buchanan, 2021-10-21 more...
40 - P0003188
Fritz Hoffmann
11614 Schach 04/1988
P0003188
(4+9)
h#2
b) Schlüsselfigur nach f3
a) 1. Sxg4 f3+ 2. Kh5 fxg4#
b) 1. fxg3ep fxg3 2. Kh5 g4#
play all play one stop play next play all
Henrik Juel: C+ Popeye 4.61 (2021-05-04)
comment
Keywords: En passant as key
Genre: h#, Retro
FEN: 8/4pp2/7p/5K2/5pPk/4r2r/3n1P1n/4B3
Input: Gerd Wilts, 1995-06-03
Last update: Rainer Staudte, 2021-05-04 more...
41 - P0003189
Tivadar Kardos
2144 Diagramme und Figuren 19/09/1967
P0003189
(7+15) cooked
h#2
1. bxc3ep Lxe2 2. Sa3 0-0-0#
play all play one stop play next play all
Cook: 2. ... Td1#
Sally: Der letzte Zug war: Bc2 - c4!
Nr.138 200 Ausgewählte S .Probleme T. Kardos(W. Fentze1983) (2010-09-30)
A.Buchanan: All 9+1 captures are visible by pawns. Note R: 1. Kd1-e1 d3xe2 is illegal because it costs 2 more captures for Black. So e.p. is unconditionally legal, requiring no AP justification. I see no reason why White has lost castling rights. Indeed someone might make a demo game ending with c2-c4. So this one seems cooked. Neither WinChloe nor yacpdb contains this problem. (2022-01-08)
A.Buchanan: If this is AP, it will be rendered sound by removing wPa2. But is this the intention? Does anyone have access to the collection of selected Kardos problems mentioned? (2024-01-25)
more ...
comment
Keywords: En passant as key, Castling (wg)
Genre: h#, Retro
FEN: 8/8/3pp3/2ppp3/1pPkrq2/4pb1P/P3p1rP/Rn2KBb1
Input: Gerd Wilts, 1995-06-03
Last update: A.Buchanan, 2022-01-08 more...
42 - P0003194
Tivadar Kardos
1597 Schachmaty 05/1963
P0003194
(6+14) C+
h#2
1. cxd3ep? bxc3 2. Lxf4 Lxg2# Last move not d2-d4
1. gxf3ep! bxc3 2. Lf4 Lf5# Last move f2-f4
play all play one stop play next play all
White's last move was certainly one of the two double pawn hops. The last move couldn't be d2-d4 because Black's pawns had already captured 10 of White's pieces, and playing d2-d4 would have resulted in wBc1 being White's 17th piece. So, the only solution is 1.gxf3 e.p.
Viktoras Paliulionis: The last move couldn't be d2-d4 because Black's pawns had already captured 10 of White's pieces, and playing d2-d4 would have resulted in wBc1 being White's 17th piece. So, the only solution is 1.gxf3 e.p. (2023-12-30)
A.Buchanan: Yes that's right! (2023-12-31)
comment
Keywords: En passant as key
Genre: h#, Retro
Computer test: HC+ Popeye C-Version 4.87 & trivial retro-logic
FEN: 2n1q3/5prp/5K1b/3b4/2pPkPp1/2p1p1pB/1P4pP/8
Reprints: 546 FIDE Album 1962-1964 1968
Input: Gerd Wilts, 1995-06-03
Last update: A.Buchanan, 2023-12-31 more...
43 - P0003195
Thomas R. Dawson
2130 Die Schwalbe 07/1932
P0003195
(14+14) C+
h#2
WTM
1. ... bxc6ep 2. Da7 c7#
retro tries:
1) 1. Lxg3 b6 2. Lxd6 Lxd6#
2) 1. Dc6 b6 2. Dc7 dxc7#
3) 1. Ta7 f8=D,f8=T 2. Ka8 Dxc8#,Txd8# ("untolerated" dual)
R: 1. ... c7-c5 2. b4-b5 Dc4-a6 (move order variable beyond this point) 3. b2-b3 Ta6-a8 4. d5-d6 Te6-a6 5. b3-b4 Te8-e6 6. a2xSb3 Th8-e8 7. d4-d5 Lb4-e1 8. d3-d4 Lf8-b4 9. Kg5-h5 e7xSf6+
play all play one stop play next play all
Missing: Wh: SS Bl: SS
Pawn captures: Wh: axb, hxg Bl: exf, fxg
bPf6 comes from e7, else collision with wPf7. So can't uncapture fxg2 until wP retracted. White has 6 pawn retractions before retro-pat, just giving Bl time to put Le1 & Ta8 (impostor) back home, and uncapture exf6. So Black must retract first, and it's WTM. sD must clear out of the way, and c4 is the only possible square, wPaxSb can't happen on b5, or it would block sD. And axSb4 would block sLe1. It was axSb3. So wPb5 must retreat to b4, and thus Black cannot play R: 1. Lb4-e1. So neatly R: 1. c7-c5 is forced.
Ladislav Packa: Is wRh1 needed? (2021-10-22)
Mario Richter: Yes, wRh1 is needed - without it the ep-key would not be justified (last moves could have been R: 1. ... Th1-g1 2. Lg1-h2 or even R: 1. ... Ld2xTe1). (2021-10-23)
A.Buchanan: + three mostly clean retro tries. (2021-10-24)
more ...
comment
Keywords: En passant as key, No legal last move for White, Impostor (t)
Genre: h#, Retro
Computer test: HC+ Popeye v4.87 & simple retro-logic
FEN: rkb5/1p1p1Ppp/q2P1p2/pPp4K/6PR/1P4PQ/2P1P1pB/4bBrR
Input: Gerd Wilts, 1995-06-03
Last update: A.Buchanan, 2023-12-08 more...
44 - P0003200
William H. Reilly
British Chess Federation 05/1944
2. Preis
P0003200
(4+9) C+
h#4
1. bxc3ep Sxg2 2. cxd2 Sxh4 3. Ke3 Kg3 4. Te4 Sf5#
play all play one stop play next play all
A.Buchanan: So no twinning in the original? What's the point of sDb3? It can be replaced by sB without issue. This problem doesn't appear in WinChloe, so can't check (2021-02-12)
VL: The author could, e.g., prefer the position with black Queen as richer and more hidden for solution. (2021-02-12)
A.Buchanan: Hi Valery well dressing the board is a hypothesis. There is no try to compensate for loss of economy. Indeed here sD forms a distracting battery which “pins” sLd3. It’s all too easy to add this kind of “richness” (noise) to a problem but more is less in my humble opinion. Unless it was a thematic tourney or we are missing a twin, my top hypothesis is that the composer did very well without a computer to make a sound h#4, and it’s easy these days to see a minor improvement. (2021-02-13)
more ...
comment
Keywords: En passant as key, En passant
Genre: h#, Retro
Computer test: C+ Gustav 4.1d
FEN: 8/8/8/8/rpPk3p/1q1b3K/3Pp1rn/4N3
Reprints: 49 PCFT 1944-1945
Input: Gerd Wilts, 1995-06-03
Last update: James Malcom, 2021-02-13 more...
45 - P0003206
Hans Joachim Schlüter
4443v Schach , p. 349, 11/1963
P0003206
(5+8)
h#2
b) wTf1 tauschen mit wLg1
a) 1. Kxb4 Lb6 2. a4 Tb1#
b) 1. cxb3ep gxf3 2. Sc1 Tg4#
play all play one stop play next play all
Korrektur 1964, Seite 155: sBh3->f3
Henrik Juel: C+ Popeye 4.61 (2022-11-24)
comment
Keywords: En passant as key
Genre: h#, Retro
FEN: 8/8/8/p7/kPp5/p1p2p2/4n1Pp/5RBK
Input: Gerd Wilts, 1995-06-03
Last update: Felber, Volker, 2022-11-24 more...
46 - P0003207
Klaus Gerber
4497 Schach , p. 30, 01/1964
P0003207
(4+9) C+
h#3
1. exf3ep c3 2. Df7 Ke4 3. Dh5 exf3#
play all play one stop play next play all
A.Buchanan: Can shift bQd5 to c4, while replacing bRg5 with bP to reach Meredith status. bBh4 can also be downgraded to bP. But probably even more economy is possible by shifting pieces one file to the right, keeping the ep, White tempo & model mate, which seem to be the main features (2022-11-30)
more ...
comment
Keywords: En passant as key
Genre: h#, Retro
Computer test: HC+ Popeye 4.61
FEN: 8/8/8/3q2r1/2p1pPkb/4K1pp/2PrP3/8
Input: Gerd Wilts, 1995-06-03
Last update: A.Buchanan, 2022-11-30 more...
47 - P0003211
Tivadar Kardos
6476 Skakbladet 10/1957
P0003211
(7+14) cooked
h#3
1. dxe3ep+ Kxc4 2. Lc3 Sxc2 3. Da5 axb3#
play all play one stop play next play all
See the correction P1396159
Cook: 1. dxe3ep+ Kxc4 2. Td5 Sxc2 3. Ta5 axb3#
A.Buchanan: Easy enough to eliminate what I suppose is the intended solution (sL unblocking sD and blocking sT) by e.g. just removing sDf5. However I don't see a way to say goodbye to the cook instead. (2021-11-23)
James Malcom: I believe I may have a fix Andrew. If we substitute the Black queen with a Black rook, the same theme is kept even is if considered watered down.

h#3 8/3p4/3p1p2/3Kbr2/k1ppPp2/1n1r3P/P1p4P/N3nq1b (2021-11-24)
A.Buchanan: Hi James, hurray cool I thought I couldn't get through that way: I was confused in my mind about wPe2 blocking that line, but with the freshness of a new day, I can look at it again. The precise diagram that you propose doesn't quite work because R: 1. e3-e4 Sf3*e1+,Sg2*e1+,g2xf1=Q+,g2xh1=Q+, but this is all fixable with +bPd2,-bSe1. Now Popeye 4.87 says the forward solution is works uniquely, and Retractor 2.0 says R: 1. e2-e4 Rf3-d3+ 2. Ke4-d5 are unique. So for full C+ we only need some demo game that results in the position prior to Ke4-d5.
I don't see any "watering down" here: I think we have faithfully represented Tivadar's idea. Let's post it here as TK, corr JM&AB? (2021-11-24)
A.Buchanan: I have made an unedifying demo game with lichess, so our correction is fully C+. (2021-11-24)
James Malcom: "TK, corr JM&AB" sounds good to me. (2021-11-24)
more ...
comment
Keywords: En passant as key, Superseded by (P1396159)
Genre: h#, Retro
Computer test: Popeye C-Version 4.87 & trivial retro-logic
FEN: 8/3p4/3p1p2/3KbqP1/k1ppPpp1/1n1r3P/P1p4P/N6b
Input: Gerd Wilts, 1995-06-03
Last update: A.Buchanan, 2021-11-24 more...
48 - P0003221
Thomas R. Dawson
1511 The Chess Amateur 10/1929
P0003221
(9+16) C+
h#2
1. ... bxc6ep 2. Le7 cxd7#
1. Lc6? bxc6 2. Le7 cxd7# WTM
R: 1. c6-c5? Kd3-e3 2. Dc5-c4+ Ke3-d3 3. Dc4-c5+ retro-loop.
R: 1. c7-c5! Kd3-e3 2. Dc6-c4+ Ke3-d3! clear
play all play one stop play next play all
Easy to see no last move for White. Moreover, Black just double-hopped to allow sD/wK to avoid retro-loop.
Single clean retro try.
more ...
comment
Keywords: En passant as key, No legal last move for White
Genre: h#, Retro
Computer test: HC+ Popeye v4.87 & moderate retro-logic
FEN: 3nkr2/3p1r2/1p1npb2/PPpb4/Bpq2PP1/1p2Kpp1/2PPP3/8
Input: Gerd Wilts, 1995-06-03
Last update: A.Buchanan, 2023-12-08 more...
49 - P0003236
Thomas R. Dawson
10381 The Fairy Chess Review 12/1955
P0003236
(9+16) C+
h#2
1. ... bxc6ep 2. Dxa5 Db7#
1. Sc6? bxc6 2. Dxa5 Db7# but WTM
R: 1. ... c7-c5 2. g5-g6 Sc6-e7 3. De7-b4,D~ Sb4-c6+ unpromote wD on h8, etc.
play all play one stop play next play all
Clean retro try
Henrik Juel: Typo: 2.Dxa5 Db7#. wTa1 is missing, I think (10+15). -1... c7 -2.g5 Sc6 -3.De7 Sb4, unpromote wD on h8, etc. (2004-03-18)
A.Buchanan: I agree Henrik and have made changes. A number of these problems have been marked "Weisz zieht an", which is inappropriate as part of the point of the problem is to deduce this. I think it's 2. c2-c3! not 2. g5-g6? or the cross-captured pawns become retro-locked. (2021-10-21)
Henrik Juel: 2.g5-g6 also works, because the only cross-capture is axb,bxa
White captured [Pd7] with an officer (2021-10-22)
Henrik Juel: HC+ Popeye 4.61 (2021-10-22)
A.Buchanan: If g5-g6 also works, then can’t we have both? I.e. White moved last. I would be surprised if this was TRD’s intention, given the others of this general style (2021-10-22)
Henrik Juel: TRD died in 1951, so someone else may have messed up the problem
FRC is not generally available after 1953, so I cannot check the source (2021-10-22)
A.Buchanan: WinChloe has 29(!) posthumous problems by Dawson - the latest in 1958, FCR's final year. The composition in question is not one of them. (2021-10-23)
A.Buchanan: PROOF OF COOK + SUGGESTED FIX
I agree with much of what you say Henrik.
Capturing history. Basically 3 possibilities for the Black pawns.
Notation: ~,o,| denote "cross-capture", "original" & "waylaid"(=captured on home file by officer).
1) a~b co d~e f~g. This implies either (wPhxg & bPh=X) or (wPh=X & bPh|)
2) a~b co d~e hxgxf f|
3) a~b co hxgxfxexd d|
Suppose WTM. Then R: 1. c6-c5 c2-c3 2. c7-c6 g5-g6 3. Sc6-e7 De7-b4 4. Sb4-c6+ Kc5-c4 5. Lb2-d4+ etc. Now we're free to uncapture wDh8 (or g8 for scenario 1), and any of the three pawn scenarios can work. No e.p. rights means a cook.
If we fix it by e.g. shifting wPg6-g5 then I think that fixes this cook, and also it means that there is no solution if BTM.
Is there any possibility for premature capturing hxg? Not by Bl because scenario 1 says bPfxg, and the other two scenarios rely on wPh=X. If by Wh, then there's no waylaying on d or f, so we are back to scenario 1, but bPh is blocked from promotion, so that won't work.
But: what about R: 1. c6-c5 c2-c3 2. Sc3-b1 Th1-a1. Doesn't this completely cook it anyway?? How about k7/4n3/6P1/PPpPPP2/qQKbrrp1/ppRppp2/b1P5/1n6? h#2 (with Art 15 so it's really h#1.5)
Do you agree? (2021-10-23)
Mario Richter: In the diagram, both for WTM and BTM the retraction sequence c2-c3 Sc6-e7 De1-b4 Sb4-c6+ works and furthermore Andrew's R: 1. c6-c5 c2-c3 2. Sc3-b1 Th1-a1 shows, that Henrik's guess of the omitted wTa1 might be wrong. Instead, I think that Andrew's modification might be in fact the original Dawson diagram. (2021-10-23)
A.Buchanan: Yes that makes sense. I can well believe that there was a lot of constructions in flight when TRD passed away. It's not surprising that something appeared a few years later. Can we get access to FCR from that year to validate? (2021-10-23)
A.Buchanan: Brian Stephenson kindly checked the original magazine. He wrote:
"Eventually got round to looking out this TRD problem. It was published in the issue you quote as problem 10381. The diagram in PDB is wrong. There should be a bPc2. The solution was published in FCR April 1956 as: 1.Pb5xPc5ep Qxa5 2.Qb7#. Last moves must have been 1.Pc7-c5 Pg5-g6 2.Sc6-e7 Qe7-b4 Sb4-c6 ch etc. No flaws were noted, but I have looked later than that issue. Hope this helps." (2021-11-13)
A.Buchanan: The appearance of the 16th Black unit thanks to the offices of Brian Stephenson simplifies the retro logic substantially, and the solution seems sound. Amazing how much disruption a simple typo can cause, but at least this one I believe is laid to rest (2021-11-13)
more ...
comment
Keywords: En passant as key, No legal last move for White
Genre: h#, Retro
Computer test: HC+ Popeye 4.6 + non-trivial retro
FEN: k7/4n3/6P1/PPpPPP2/qQKbrrp1/ppPppp2/b1p5/1n6
Input: Gerd Wilts, 1995-06-03
Last update: A.Buchanan, 2023-12-08 more...
50 - P0003240
Gabor Tar
Magyar Sakkelet 06/1972
P0003240
(4+10) C+
h#2
b) sLe3 nach f3
a) 1. f3 Lxa7 2. Lg5 Lxf2#
b) 1. fxg3ep Lxc7 2. Lh5 Ld8#
play all play one stop play next play all
Henrik Juel: a) C+ Popeye 4.61
b) the ep capture is unjustified, last move could be Ke2-f1 or f3xg4 (2023-08-06)
A.Buchanan: There was an error in the stipulation: "b) sLe3 nach h2 (-wSh2)" is completely incorrect. However "b) sLe3 nach f3" leads to a very nice problem which matches the solution already present. This problem is not in WinChloe, but I am sure this is the resolution of the mystery (2023-08-07)
Henrik Juel: You are probably right, Andrew; maybe there was a misprint in the source
Now part b) is C+ Popeye 4.61 after very simple analysis (2023-08-07)
comment
Keywords: En passant as key
Genre: h#, Retro
Computer test: HC+ Popeye v4.61 & simple retro-logic for b)
FEN: 1B6/npp5/8/8/5pPk/4b2p/3p1p1N/5Kn1
Input: Gerd Wilts, 1995-06-03
Last update: A.Buchanan, 2023-08-07 more...
51 - P0003269
Tivadar Kardos
149 British Chess Magazine 10/1956
P0003269
(5+17) cooked
h#3
1. exf3ep e3 2. 0-0-0 Tb4 3. Tg4 Tb8#
I think that this cooked problem is an early A osteriori. I can see ways that it might be made sound, but I would like to see the published version. San anyone help please? Thanks.
play all play one stop play next play all
Cook: 17 Black pawns, retro cook and numerous forward cooks
klären wK im Schach, vielleicht wBb2?
Alfred Pfeiffer: 9 schwarze Bauern! (2012-02-07)
Ladislav Packa: Auch mit wBb2 NL, z.B. 1.exf3 e.p. Txg5 2.0-0-0 Tb5 3.Tg~ Tb8# (2012-02-07)
A.Buchanan: And adding to earlier comments, if we do swap sBb2 for wB, the retraction is still not unique with R: 1. f3-f4 Lh6xg5+ (2021-11-26)
A.Buchanan: I think that this cooked problem is an early A Posteriori. I can see ways that it might be made sound, but I would like to see the published version. Can anyone help please? Thanks. (2022-04-25)
Gerald Ettl: Verbesserungsvorschlag: -sBa2, -sBb2, -sSg7, +sBg7
r3k3/3b2p1/5p2/6b1/4pPRp/2pq2rp/2p1P1pB/2K3n1/ (2022-04-25)
Gerald Ettl: und +sSg8 (wegen exf4 Vermeidung) (2022-04-25)
Gerald Ettl: +sSa4 nicht g8 (2022-04-25)
A.Buchanan: Hi Gerald. Thanks for this. Your final proposed version is r3k3/3b2p1/5p2/6b1/n3pPRp/2pq2rp/2p1P1pB/2K3n1, yes? There are 16 candidate solutions, all using castling. But wPa cannot leave the a-file, and is required for capture balance, so Black cannot castle. Five of the candidates begin with e.p.: play might have just been R: 1. Kb1-c1 b3xPc2+, so e.p. is not permitted in any case. What am I missing? (2022-04-25)
Gerald Ettl: Hi Andrew,
Those were 2 mistakes I made. I improved it again.
r3k3/3b4/4pp2/6b1/4pPRp/n1pq2rp/PPp1P1pB/2K3n1/
(den wBa2 habe ich auf das Brett gestellt, da sonst wieder Lh6xXg5 geht.) (2022-04-26)
A.Buchanan: Hi Gerald: Deine Retro-Logik ist gut. Aber es gibt 18 Lösungsvorschläge für die Zukunft. Ich habe heute im Discord meinen eigenen Vorschlag zur AP-Korrektur von Kardos veröffentlicht. Ich werde ihn hier im PDB hinzufügen. (2022-04-26)
comment
Keywords: En passant as key, Castling
Genre: h#, Retro
FEN: r3k3/3b2n1/5p2/6b1/4pPRp/2pq2rp/ppp1P1pB/2K3n1
Input: Gerd Wilts, 1995-06-03
Last update: A.Buchanan, 2021-11-26 more...
52 - P0003281
Fritz Hoffmann
Problem 08/1974
P0003281
(4+11) C+
h#2
3.1...
1) 1. Tb1 Lxf3 2. Tb3 Lc6#
2) 1. Lc6 Le4 2. Lb5 Lc2#
3) 1. cxb3ep Lxc3 2. Lc6 Lxc6#
play all play one stop play next play all
more ...
comment
Keywords: En passant as key
Genre: h#, Retro
Computer test: C+ Popeye 4.61
FEN: 8/8/1b6/B7/kPp3p1/p1p2b1p/6Bp/3r1n1K
Input: Gerd Wilts, 1995-06-03
Last update: A.Buchanan, 2021-05-05 more...
53 - P0003287
Tomislav Petrovic
(VIII) Problem 37-40 09/1956
P0003287
(5+9) C+
h#2
1. fxg3ep h8=S 2. Sg4 Sxg6#
play all play one stop play next play all
A.Buchanan: The point I think is that the en passant is a tempo move. Otherwise wBf2 serves no purpose. How about sBh5 instead of sT? (2021-10-19)
more ...
comment
Keywords: En passant as key, Promotion (S), Tempo Move
Genre: h#, Retro
Computer test: HC+ Popeye v4.61 & trivial retro-logic
FEN: 8/7P/6pP/6pr/5pPk/4np1n/5P1p/7K
Input: Gerd Wilts, 1995-06-03
Last update: A.Buchanan, 2021-10-19 more...
54 - P0003288
Vladimir Korolkov
211 Biuletyn 03/1960
P0003288
(9+10)
h#2
1. fxg3ep Lxe3 2. Ld5 cxd5#
play all play one stop play next play all
Henrik Juel: C+ Popeye 4.61 and very simple analysis (2022-08-09)
comment
Keywords: En passant as key
Genre: h#, Retro
FEN: 8/8/8/1PP1P2r/PRPBbpPk/ppKppp1r/8/8
Input: Gerd Wilts, 1995-06-03
Last update: James Malcom, 2022-08-09 more...
55 - P0003291
Jozsef Korponai
5412 Arbejder-Skak 02/1961
P0003291
(11+13) C+
h#2
1. hxg3ep Sd4 2. gxf2 Se3#
play all play one stop play next play all
vgl. P0003180
Henrik Juel: Last move was g2-g4 giving Black a preceeding move like Da8-b8 or Lb7-c8
C+ Popeye 4.61 (2021-04-24)
comment
Keywords: En passant as key
Genre: h#, Retro
Computer test: HC+ Popeye 4.61 + simple retro thought
FEN: 1qb5/b1pp1ppp/pp6/6P1/5BPp/2P5/1PN1PP1P/3Nrk1K
Input: Gerd Wilts, 1995-06-03
Last update: A.Buchanan, 2021-04-25 more...
56 - P0003297
Wilhelm Hagemann
Deutsche Nachrichten (Sao Paolo) 1960
P0003297
(5+9)
h#2
1. cxd3ep f4 2. Lf3 Kxc5#
play all play one stop play next play all
Henrik Juel: C+ Popeye 4.61 (2022-04-18)
Henrik Juel: Obviously, last move was d2-d4 (2022-04-18)
comment
Keywords: En passant as key
Genre: h#, Retro
FEN: 8/8/p7/P1r2p2/RKpPk3/p1p1p3/5P2/3b4
Input: Gerd Wilts, 1995-06-03
Last update: Rainer Staudte, 2022-04-18 more...
57 - P0003312
Gerhard Paul Latzel
2680 Deutsche Schachzeitung 02/1970
P0003312
(7+14) C+
h#3
1. cxb3ep e5 2. Lh6 g5 3. Txh4 Txh4#
play all play one stop play next play all
Henrik Juel: The only possible last move is b2-b4, so the ep key is permissible
C+ Popeye 4.61 (2021-07-26)
A.Buchanan: The main idea is to unblock the cluttered 4th rank, and this is complemented thematically by *blocking* four other lines: g8-g4, h8-h4, h8-d4 & "blocking the blocker" h6-f4.
The last move was b2-b4, so by the ep rule, capture is definitely legal. The convention only serves to resolve ambiguous situations where multiple possible histories exist. In retro problems, such ambiguity is rare if the retro logic does its job. However the pessimistic nature of the e.p. convention motivates *why* the retro play must exist. (2021-07-26)
comment
Keywords: En passant as key
Genre: h#, Retro
Computer test: Last move was b2-b4, so by ep rule, capture ok as key. HC+ Popeye 4.61
FEN: 6rq/pn6/K7/8/kPprPbPP/p1pppppR/7P/8
Input: Gerd Wilts, 1995-06-03
Last update: Rainer Staudte, 2021-07-26 more...
58 - P0003316
Eduard Schlatter
590v Schach-Echo 05/07/1955
P0003316
(4+15)
h#4
1. bxc3ep Kxa3 2. Ld5 Sg2 3. Tce1 Kb4 4. T1e4 bxc3#
play all play one stop play next play all
Henrik Juel: C+ Popeye 4.61
The only possible last move is c2-c4 (2021-08-04)
comment
Keywords: En passant as key
Genre: h#, Retro
FEN: 8/2p5/2p3pq/n3r3/KpPk1pp1/bp1p4/1P4b1/2r1N3
Input: Gerd Wilts, 1995-06-03
Last update: Dieter Berlin, 2021-08-04 more...
59 - P0003323
Peter Takacs
Imre Telkes

285 The Chess Amateur 12/1922
P0003323
(3+13) C+
h#2
1. gxf3ep Dxh2 2. Ld3 Dxh4#
play all play one stop play next play all
more ...
comment
Keywords: En passant as key, Selfblock, Line opening
Genre: h#, Retro
Computer test: C+ Popeye 4.61 and very simple analysis
FEN: 6n1/2r4q/3pK3/7p/4kPpp/4r1pQ/7n/b4b2
Input: Gerd Wilts, 1995-06-03
Last update: A.Buchanan, 2022-06-06 more...
60 - P0003325
Tivadar Kardos
4272 Arbejder-Skak 08/1956
P0003325
(6+13) C+
h#5
a) 1. gxf3ep g4 2. Kd5 gxf5 3. Kc6 fxe6 4. Kb7 exd7 5. Ka8 dxc8=D#
play all play one stop play next play all
A.Buchanan: Popeye v4.87 (via Olive v1.4 ) for PDB problem P0003325 h#5 delivers the correct solution with en passant set to f2f3f4. However no solution when the "intelligent" flag is also set. Reported to Popeye Github. (2021-11-23)
Henrik Juel: Popeye 4.61 with 'opt var int enp f3' works fine
Maybe you just have to change f2f3f4 to f3, Andrew (2021-11-23)
A.Buchanan: Thanks for the suggestion, Henrik, but even without the "intelligent", opti vari enpa f3 fails with v4.87. I think that the syntax has been rendered more complicated since v4.61 in order to support fancy fairy en passants. (2021-11-23)
comment
Keywords: En passant as key, Promotion (D), Excelsior, Pac-Man
Genre: h#, Retro
Computer test: HC+ Popeye C-Version 3.52 & trivial retro-logic
FEN: 2n5/p2p4/4b3/p4p2/P3kPp1/Ppp5/Pp4P1/bK2n3
Input: Gerd Wilts, 1995-06-03
Last update: A.Buchanan, 2021-11-23 more...
61 - P0003326
Tivadar Kardos
4252 Arbejder-Skak 07/1956
P0003326
(3+13) C+
h#5
1. dxc3ep Kxg7 2. f6 Kxf6 3. e5 Kxe5 4. Lc4 Kd4 5. Lb5 dxc3#
play all play one stop play next play all
A.Buchanan: Can be done with lighter material e.g. sLa5, sBa4, sSa3g7 (2021-11-23)
more ...
comment
Keywords: En passant as key
Genre: h#, Retro
Computer test: Popeye C-Version 4.87 & trivial retro-logic
FEN: 6bK/5pr1/2p1p3/p4p2/rkPp4/qp1p4/3P4/8
Input: Gerd Wilts, 1995-06-03
Last update: A.Buchanan, 2021-11-23 more...
62 - P0003328
Tivadar Kardos
4311 Arbejder-Skak 09/1956
P0003328
(7+16) C+
h#3
1. cxb3ep Sb4 2. Sc4 Sc2 3. Sa5 Sxc3#
play all play one stop play next play all
Henrik Juel: The only possible last move is b2-b4
C+ Popeye 4.61 (2021-04-24)
comment
Keywords: En passant as key
Genre: h#, Retro
Computer test: HC+ Popeye 4.61 & simple retro thought
FEN: 8/2p5/2P5/2p2p2/kPp2Pn1/n1pp1Kp1/N2P2pb/rNr1q2b
Input: Gerd Wilts, 1995-06-03
Last update: A.Buchanan, 2021-04-25 more...
63 - P0003335
Tomislav Petrovic
(XXI) Problem 37-40 09/1956
P0003335
(5+9) cooked
h#9
1. dxc3ep dxc3+ 2. Ka5 c4 3. d2 c5 4. d1=L c6 5. Lg4 hxg4 6. Kb6 g5 7. Ka7 g6 8. Kb8 g7 9. Kc8 g8=D/T#
play all play one stop play next play all
Cook: 1. b2 c5 2. b1=S c6 3. Sc3 dxc3+ 4. Ka5 cxd4 5. Kb6 d5 6. Ka7 d6 7. Kb8 dxc7+ 8. Ka7 c8=D 9. a5 Db7#
paul: See P0003212 as version. (2011-08-06)
Anton Baumann: C+?? Die Aufgabe ist eindeutig inkorrekt! (2021-01-27)
James Malcom: Fixed. (2021-01-27)
comment
Keywords: En passant as key, Promotion (D), konsekutive Umwandlungen 2 (L, D/T), En passant, Kindergarten Problem
Genre: h#, Retro
FEN: 8/2p5/p7/8/1kPp3p/1p1p2pP/3P2Pp/7K
Input: Gerd Wilts, 1995-06-03
Last update: James Malcom, 2021-01-27 more...
64 - P0003337
Rudolf Buljan
Problem 41-44 03/1957
2. Preis
15. Thematurnier
P0003337
(11+15) C+
h#2
1. ... exf6ep 2. Sf5 f7#
play all play one stop play next play all
White captured fxg and promoted [Pd2] on d8
Black captured axbxa2-a1=Y, b7xc6, d5xe4, and gxf
White cannot have moved last, and in fact Black's last move can only have been f7-f5 in order to give White a prior move. R: 1... d5-e4 would have retro-blocked wPd marching to its promotion.
So the retroplay was R: 1... f7-f5 2.f5xYg6
Numerous retro tries but none distinct.
"Autor Ing. Rudolf Buljan, Zagreb"
AB: Have therefore marked him in the author table as Croatian. Yugoslavia has been removed from the PDB country list, I see.
A.Buchanan: Any ideas how to distinguish usage of “whose move?” and “no legal last move for white” (or black) keywords (2020-10-01)
Henrik Juel: This problem is clearly the latter
I would use 'Whose move?' in retro problems where a deeper analysis is required to determine the move, often with stipulations like '#1 (who?)' (2020-10-01)
A.Buchanan: I think "Whose move?" is where there is no forward stip, or where we have to determine if the forward stip should be interpreted as orthodox or as half-duplex. "no legal last move for..." is where the mater/stalemater/etc is the usual one, but we add or remove a single move at the beginning of the forward stip. I guess a third possibility is "last move?" where in a Type A position we also have to figure out who moved last.
With respect, I don't think the level of analysis can be relevant. (2023-12-06)
more ...
comment
Keywords: En passant as key, No legal last move for White
Genre: h#, Retro
Computer test: Popeye 4.61
FEN: 5nkb/1qp1p1n1/2p3PQ/4PpKb/4prRp/P4prB/1PP3PP/8
Input: Gerd Wilts, 1995-06-03
Last update: A.Buchanan, 2023-12-06 more...
65 - P0003339
Tivadar Kardos
2519 Revista Romana de Sah 01/1948
2. Preis
P0003339
(10+15) cooked
h# in wieviel?
1. ... cxd6ep 2. Tc8 dxe7 3. Tf8 exf8=S#
play all play one stop play next play all
AL ('Revista Romana de Sah', 06/1948, S.225-226):
Bekanntlich beginnt im Hilfsmatt Schwarz, also hat in der Diagrammstellung Weiß den letzten Zug gemacht. Man findet aber, wenn man die Position analysiert, daß Weiß den letzten Zug nicht gemacht haben kann.
Der weiße Doppelbauer auf der d-Linie läßt sich nur durch Schlagen des einzig fehlenden schwarzen Steines, des sBa7, erklären. Dieser konnte aber erst nach Umwandlung geschlagen werden. Dazu mußte er zunächst bis a3 vorrücken, dann nach b2 schlagen, und dann umwandeln.
Also kann Weiß nicht zuletzt b2-b3 gezogen haben.
Die anderen schwarzen Bauern haben mindestens 5 Schläge gemacht, um die Position im Diagramm zu erreichen (man beachte, daß der sBh7 zweimal schlagen mußte, um den wBh2 durchzulassen)
also kann auch g2-g3 nicht der letzte Zug gewesen sein, denn sonst hätte der wLf1 nicht von einem sB geschlagen worden sein können.
Für alle anderen weißen Steine ist leicht zu sehen, daß sie den letzten Zug nicht gemacht haben können.
Es stellt sich heraus, dass der letzte Zug von Schwarz gemacht wurde, und eine einfache Analyse zeigt, daß dies nur möglich ist durch d7-d5 (sonst wäre es unmöglich, daß die weißen Steine ihre Position im Diagramm erreichen könnten).
In diesem Fall Weiß kann 1. ... c5:d6(e.p.) spielen, gefolgt von 2. Tc8 d:e7 3. Tf8 e:f8C matt!
Das kürzeste Hilfsmatt wird also in 2,5 Zügen erreicht!
Cook: 1. ... Txg5+ 2. Kxg5 f3 3. Lg6 gxf4#
1. ... Kxd5 2. Kf6 e4 3. Tg6,Dg6 e5#
A.Buchanan: Popeye v.487. WinChloe has the same diagram, and although it just claims 1 solution, it's C? Yet WinChloe engine gives the cooks right away (2021-10-27)
Mario Richter: Shouldn't this problem get the Label "No legal last move for White"?
(White pawn d2 has captured the missing black piece (Pdxe), so black pawn a7 had to promote to get to the capture square. This requires one capture (Pa3xb2), together with bxc,cxd,hxg,gxh,gxh this accounts for all missing white pieces.)

Btw. "for me, "C?" stands for "Not yet computer-tested or computer test didn't come to a conlusion" ... (2021-10-28)
A.Buchanan: Hi Mario, thank you.
(1) The stipulation is non-standard. I think this is why the WinChloe database doesn't include a solution, and just marks it as "C?" However if I run the WinChloe solving engine manually it (like Popeye) gives 121 h#3 solutions and 4 h#2.5 solutions. I think you, me and Christian would all agree on the meaning of "C?"
(2) The problem does not rely on Article 15 to eliminate the first single move, but implicitly requests the shortest. Currently the PDB usage of "No legal last move..." is to indicate that Codex Article 15 applies, so I did not add it. But I'm ok if you want me to add the keyword. In my correction, I shift the stipulation to h#3 so it's clear that Article 15 and not any desire for brevity which push the problem to h#2.5.
(3) However the PDB usage is confused. Ideallly think the two "No legal..." keywords should be replaced by "Retro-stalemate", while two new keywords "Article 15" & "Whose mate?" indicate the choice of implications of the retropat for forward play. "Board rotation" and other jokes are also implications of retropat (and of illegal diagram).
(4) But this is complicated to implement, requiring either database-level access or screen-scraping. The latter is quite practical for some future time when I have more leisure.
(5) https://www.stere.ro/biblioteca-digitala/reviste/revista-romana-de-sah/ is an excellent archive of Revista Romana de Sah. My sincere compliments and thanks to whoever prepared it. And the magazines themselves contain thorough annual indices. p225-6 of 6/1948, contains the solution to Kardos. I haven't translated it, nor have I found time to locate the prize announcement.
Comments welcome! (2021-10-29)
Mario Richter: I still believe, that the label "No legal last move for White" is justified here, because this is exactly what the author wanted to show.
(And this is too, what the reprint in 'Problem (Zagreb)' 37-40 09/1956 is all about - it appears there inside asn article

Btw., I wouldn't call the presentation of the solution in 'Revista Romana de Sah', p.225, 06/1948, a "reprint" - it lacks the presentation of the position ... (2021-10-30)
A.Buchanan: Added the keyword.
Yes I know that's the usage of "reprint", but then there should be a field for the location of the solution :-) It's idiotic that if I want to say where the solution is contained, I have to embed it in text like this: 2519 Revista Romana de Sah , p. 225-6, 06/1948 (2021-10-30)
more ...
comment
Keywords: En passant as key, Promotion (S), Superseded by (P1394978), No legal last move for White
Genre: h#, Retro
Computer test: Popeye v4.87 & WinChloe both say cooked
FEN: 7b/4p1nP/2r1P1kq/1nPpKbrp/2pp1pRp/1P4P1/P3PP2/8
Reprints: (III) Problem 37-40 09/1956
Input: Gerd Wilts, 1995-06-03
Last update: A.Buchanan, 2021-10-30 more...
66 - P0003343
Yaakov Mintz
344 Canadian Chess Chat 11-12/1980
P0003343
(4+15) cooked
h#4
1. hxg3ep Lxb1 2. Kf5 fxe4 3. Kg6 e5+ 4. f5 exf6ep#
play all play one stop play next play all
Cook: NL:
1. hxg3ep fxe4 2. f6 e5 3. Kf5 exf6 4. Kg6 Lxb1# uvm
YM: The correct mechanism is in P1109109 (2010-07-13)
YM: Correction option: P1229434 (2021-06-28)
comment
Keywords: En passant as key, Superseded by (P1109109)
Genre: h#, Retro
FEN: q2r4/4Kp2/3r3b/3p2pp/4nkPp/1n3P2/B4p1p/1b6
Reprints: 805 FIDE Album 1980-1982 1988
Input: Gerd Wilts, 1995-06-03
Last update: A.Buchanan, 2021-06-28 more...
67 - P0003347
Marko Klasinc
3. Makuc-Moder-Gedenkturnier 1971-1973
1. Lob
P0003347
(7+6) C+
h#2
1. axb3ep bxc6+ 2. b5 cxb6ep#
play all play one stop play next play all
Marko Klasinc: The most economical h#2 with two e.p. captures. (2002-01-28)
A.Buchanan: Very nice - and i think this is unbeatable at least in the case where ep is first and last move (2020-08-02)
A.Buchanan: Also works as set play! Just truncate Black's first move! :) (2020-12-09)
more ...
comment
Keywords: En passant as key, Economy record, En passant as mating move
Genre: h#, Retro
Computer test: C+ Popeye v4.61 + a little thinking
FEN: 8/1p6/B1p5/RPP5/pPkp4/K1p5/P7/8
Input: Gerd Wilts, 1995-06-03
Last update: A.Buchanan, 2020-12-09 more...
68 - P0003359
André Hazebrouck
3256 Themes-64 07-09/1977
P0003359
(14+10) C+
h#2
2.1...
1) 1. 0-0 Le6+ 2. Kh8 Sg6#
2) 1. bxc3ep 0-0 2. 0-0-0 Tc4#
play all play one stop play next play all
If all three castlings are legal, then the ep is on. So evidence must be accumulated from the two solutions. White 0-0 serves only in this retro role, eliminating 26 other candidate solutions.
A.Buchanan: There was a diagram error here. It should be bPd7 not bPe7. The existing diagram is vastly cooked with e.g. 1. Ke8 Kd1 2. Te8 Lb6#. WinChloe has the correct diagram here. (2022-05-24)
comment
Keywords: a posteriori (AP) (Type Petrovic), En passant as key, Castling (wksksg)
Genre: h#, Retro
Computer test: HC+ Popeye v4.87 & retro logic
FEN: r3k2r/B2p2pp/8/NP6/RpP4N/pP5B/1p1PPPP1/n3K2R
Reprints: feenschach 54 04/1981
Input: Gerd Wilts, 1995-06-03
Last update: A.Buchanan, 2022-06-06 more...
69 - P0003365
Gyula Bebesi
41 Problemas 04-06/1962
P0003365
(8+14) cooked
h#2
1) 1. axb3ep bxc6+ 2. b5 cxb6ep#
2) 1. exd3ep g8=D,L+ 2. d5 cxd6ep#
play all play one stop play next play all
PRA: 1 solution with 2 parts
Henrik Juel: White captured [sLc8] on c8 and axb, so last move was either b2-b4 or d2-d4
C+ Popeye 4.61, except for the promotion dual (2020-08-03)
A.Buchanan: There's no keyword for an untolerated dual promotion! (2020-08-03)
Henrik Juel: Yes and no, Andrew
What about the cooked label? (2020-08-03)
A.Buchanan: Hi Henrik: this is obviously intentional. I tried to replace the pawn promotion with wDh8-h7+ but I ran out of pieces. I guess that Bebesi had the same difficulty. The word "cooked" seems too strong to me, but it's obviously a significant defect though (2020-08-04)
A.Buchanan: The bar is higher for helpmates than other problems, and this is unsound. However it’s not cooked but dualled. Yet all we have is “cooked” so I suppose it is in PDB-speak. I’ve added that and removed the 2.1… (2023-08-07)
Joost de Heer: Is there an 'unsound' keyword? (2023-08-07)
A.Buchanan: Hi Joost: no there isn't. It can be added, the challenge for any new common keyword would be populating it for more than a few values. (2023-08-07)
A.Buchanan: This time round, I've managed to find a fix. Bebesi's correction in WinChloe 66992 has non-standard wL. I have a feeling I've seen a similar one by Andrey Frolkin & someone else recently with a different matrix, and also non-standard wL. I've sent my effort to Joaquim Crusats at Problemas for his assessment (2023-08-08)
more ...
comment
Keywords: En passant as key (2), Partial Retro Analysis (PRA), En passant as mating move (2), Superseded by (P1411659, P1413906)
Genre: h#, Retro
FEN: 3qn3/1p1p2P1/B1r3p1/1PP1Kp2/pPkPp3/p1p4n/N7/2b5
Reprints: 233 Ungarische Schachproblem Anthologie , p. 188, 1983
1 Problemas 44, p. 1456, 10/2023
Input: Gerd Wilts, 1995-06-03
Last update: A.Buchanan, 2023-12-04 more...
70 - P0003366
Laszlo Lindner
3040v Europe Echecs 07-08/1982
P0003366
(8+16)
h#2
b) sBc7 nach f6
a) 1. Dc4 Lxb6+ 2. c5 dxc6ep#
b) 1. fxe3ep Lxb4 2. Sc4 bxc3#
play all play one stop play next play all
Lindner in 'Mattbilder eines Lebens':
In a) ist die Lösung der stellung b) nicht möglich, weil das e.p.-Schlagen durch Schwarz nicht legal ist. Der letzte Zug von Weiß muß nicht unbedingt e2-e4 geweseb sein. Es kommt als letzter zug auch Kh3-g2 in Betracht, mit den vorherigen Zügen h4:g3 e.p.+ und g2-g4.
In b) demgegenüber sind Kh3-g2 und vorher f4:g3 e.p. illegal, weil die Rücknahme von g2-g4 unmöglich ist: der sB würde 7 Schlagfälle benötigen, und es fehlen nur 6 weiße Steine. Der letzte weiße Zug muß also e2-e4 gewesen sein.
In 'Mattbilder eines Lebens' abgedruckt mit sTh7 statt h8 und der Quellenangabe: Europe Echecs, 1964
AB: (1) Where is wK?
(2) Why is 1.fxe3ep legal in b) but not a)? (2002-01-31)
Henrik Juel: wK is probably on g2. In part a) last move could have been Kh3-g2, I think (2002-02-01)
A.Buchanan: Very convincing, Henrik. I've repaired the diagram accordingly. (2023-05-28)
comment
Keywords: En passant as key, En passant in the retro play
Genre: h#, Retro
FEN: 7r/2pn4/1nqRb3/B2Pp3/pb1kPp2/2p2Pp1/1PP2pKp/7r
Reprints: 501 Mattbilder eines Lebens , p. 379, 1996
Input: Gerd Wilts, 1995-06-03
Last update: Mario Richter, 2023-05-28 more...
71 - P0003369
Valerian Onitiu
6643 The Fairy Chess Review 02/1946
P0003369
(14+11) C+
h#1*
1. ... fxg6ep#
1. dxe3ep fxe3#
play all play one stop play next play all
Gerald Ettl: Der rumänische Autor zeigt, dass der letzte Zug jeweils ein Doppelschritt g7g5 oder e2e4 gewesen sein muss. Die wBc6,d7 schlugen ueber Kreuz und der wUW-L hat bxaB und a7xXb8 geschlagen. Die Forderung lautet wohl h#1 (2010-07-12)
Alfred Pfeiffer: Ja, die Forderung ist h#1*.
Online: "http://problem64.beda.cz/silo/fcr_4_1946.pdf" (2010-12-15)
more ...
comment
Keywords: En passant as key (2), Promotion (L), Non-standard material (L)
Genre: h#, Retro
Computer test: C+ Popeye 4.61 and analysis
FEN: 4b3/1p1P1p2/2PRBR1P/2prpPpK/3pPkr1/5pN1/P4P1B/6B1
Input: Gerd Wilts, 1995-06-03
Last update: A.Buchanan, 2021-06-30 more...
72 - P0003384
Tivadar Kardos
4299 Stella Polaris 12/1971
P0003384
(9+16) C+
h#2
b) alles eine Reihe nach links
a)
1. gxf3ep Sxg7 2. d5 Lf5#
1. cxd3ep? Sxc7 2. f5 Ld5#
b) 1. bxc3ep Sxb7 2. e5 Lc5#
1. fxe3ep Sxf7 2. c5 Le5#
play all play one stop play next play all
Sally: a)Der letzte Zug war: Bf2 - f4!
b)Der lrtzte Zug war: Bc2 - c4!
Nr. 139 200 Ausgwwählte S. Pr. T. Kardos (W. Frentze 1983) (2010-09-30)
Henrik Juel: Black pawns captured all 7 missing white men, incl. [Pa1], which promoted on a8
Only possible last moves are d2-d4 and f2-f4
In each twin the closing off for an original white bishop determines the last move (2021-11-22)
Henrik Juel: HC+ Popeye 4.61 (2021-11-22)
comment
Keywords: En passant as key, Twin by board shift
Genre: h#, Retro
Computer test: HC+ Popeye v4.87 & Retractor 2.1.1 Human contribution is "positions prior to retractions look legal enough."
FEN: 2bKNb2/2rn1pqn/3pBp2/4B1P1/2pPkPpr/2p1p3/1P2p2P/8
Input: Gerd Wilts, 1995-06-03
Last update: A.Buchanan, 2021-11-27 more...
73 - P0003387
Tivadar Kardos
(4) Land og Folk 22/12/1973
P0003387
(6+15) C+
h#1
1. bxc3ep Lxc5#
1. fxe3ep Le5#
R: 1. c2-c4,e2-e4
play all play one stop play next play all
Originalforderung? "2 Lösungen"
A.Buchanan: The original stipulation here asks for 2 solutions. I don't think this would work as 2 solutions under Retro Strategy protocol which was the default in those days. Dubious e.p. captures are just not permitted. This is why AP was invented as a funny work-around.
However this problem works under the PRA protocol, which decomposes the history as "one solution, two parts". Note that this problem does not make use of the e.p. convention, as it is certain that the two e.p. cannot be simultaneously legal. (2021-12-22)
Henrik Juel: I can strengthen the human contribution
The positions before c2-c4 and e2-e4 are surely legal (2021-12-22)
more ...
comment
Keywords: En passant as key, Partial Retro Analysis (PRA)
Genre: h#, Retro
Computer test: HC+ Popeye v4.87 + Retractor 2.1.1 Human contribution is "position prior to retractions looks legal enough."
FEN: 3K4/1rp1n3/3Bbnq1/N1p2Qr1/1pPkPpp1/1p1p1p2/8/8
Input: Gerd Wilts, 1995-06-03
Last update: A.Buchanan, 2021-12-22 more...
74 - P0003411
Norman Alasdair Macleod
3970 Themes-64 04-06/1982
P0003411
(4+6) C+
h#2 (AP)
2.1...
1. Kc3 0-0-0 2. Txc4 Txd3#
1. bxc3ep e4 2. Kc4 Ta4#
play all play one stop play next play all
The idea is that the ep in one solution is validated by the castling in the other solution. Since no other solutions exist, there are no parasites which might "piggyback" off the proof given by the castling solution. This is not PRA: both solutions have the same history with both castling & hence ep legal.
Mario Richter: How is the (AP) to be interpreted here? Is the intention as follows: Since in solution 1) White castles, Black is allowed to capture e.p. in solution 2)? (2011-05-28)
Henrik Juel: It seems to be an unusual situation, not covered by keywords like AP or PRA.
If last move was c2-c4, both solutions work; if not, there is no solution, because White may not castle.
(In the second solution, Kxc4 should be Kc4). (2011-05-28)
VL: This a generalized type of AP, which I call "consolidate AP": both solutions are considered as parts of one complete solution. However the order of both parts is significant. I know two similar problems: P0003437 and P0003186. (2011-06-01)
A.Buchanan: @Valery: I've started to classify all the AP problems in PDB. This and its kin are "Type Petrovic - cons" (standing for consolidated, but there's only a limited number of characters for a PDB parameter). I truly hope that the theory in the end will not have this as a special case, but rather is part of a more general pattern.
It's not just that PRA & RS each scales up within its own paradigm to handle seamlessly very complex collections of conditional move dependencies. They also scale *down* to handle situation when there is just one conditional move, or even zero conditional moves. And when n=0 or n=1, PRA & RS are in complete agreement. This is the kind of robust seaworthy behaviour we need if are going to set sail in the stormy seas of fairydom.
But small steps. The first thing is to classify all the problems in PDB into buckets, to find out what we have. I doubt there are 37 categories, so will probably need the results of your scholarship too. (2022-02-17)
more ...
comment
Keywords: a posteriori (AP) (Type Petrovic consol), En passant as key, Castling (wg)
Genre: h#, Retro
Computer test: HC+ Forward play proved by Popeye v4.87 AP logic requires some framework but I think we know we want this cool idea to end up being sound.
FEN: 8/8/8/1nr5/1pPk4/1p1p4/4P3/R3K3
Input: Gerd Wilts, 1995-06-03
Last update: A.Buchanan, 2023-09-12 more...
75 - P0003417
John Frederick Keeble
2206 The Problemist Fairy Chess Supplement 16, p. 173, 02/1936
P0003417
(4+5) C+
h#2.5 (AP)
1. ... exd6ep 2. 0-0-0 dxe7+ 3. Tf8 exf8=T,D#
play all play one stop play next play all
H.Juel: If Black may castle, last move was not made by Ke8 or Ta8, but by Pd5, so it was d7-d5, and hence White may capture en passant. 1... exd6ep 2.0-0-0 dxe7 3.Tf8 exf8=D,T# where the castling (belatedly) justifies the e.p. capture.
Not 1... exd6ep 2.Kd8? dxe7 3.Kc8 e8=D,T
AB: Another 120 retro tries (=60x2 due to tolerated dual promotion) of general form 1. exd6ep 2. T~ dxc7 3. T~ c8=D,T.
No set play
Nachdruck in "Die Schwalbe" 33 mit Diagrammfehler (wBe6 fehlt).
VL: The first AP-problem! Thematic illegal try:
1. ... exd6ep 2.Kd8? dxe7+ 3.Kc8 e8=T(D)#?? (2007-02-11)
Mario Richter: 46 years between this Problem and P1366663 - perhaps this was not Keeble's first AP-Problem?! (2019-08-29)
Ladislav Packa: Of course I know the idea a posteriori. But let my heresy be forgiven - I do not consider it logical. In this particular example, I consider e.p. for proof that black can castling - not vice versa. (2019-08-29)
A.Buchanan: Hi Ladislav. If ep is ok, then Black just moved a pawn with double hop. That doesn't stop Black having e.g. moved bK right at the beginning of the game. So you can't prove that Black definitely can castle this way. What am I missing here? (2019-08-29)
Ladislav Packa: Your statement is perfectly fine. However, the AP condition is usually interpreted as castling is evidence of e.p. However, the second party does not have to do castling - the possibility has already been demonstrated by the existence of e.p. (2019-08-29)
A.Buchanan: Hi again Ladislav. Please try again, I am obviously being very stupid. I don't understand how 1. exd6ep is any more indicative that Black can castle than 1. K~. (2019-08-30)
Ladislav Packa: Hi, Andre!
e.p. is just a result of the d7-d5 move. Black can castle only if it was the last move of black (even without wPe5). Therefore, it does not have the logic of joining the e.p. with castling. For example, ask yourself if black is allowed to castle in the position without wPe5. (2019-08-30)
A.Buchanan: @Ladislav: ep is not "proof that black can castle", rather it is a *necessary* condition for Black to be able to castle. The logical implication is the other way round. We can castle because of the optimistic convention, and because sPd might just have moved. Without wPe, there is no solution, but whatever White plays, Black can castle (2020-12-11)
Hans-Jürgen Manthey: was heißt hier Thematisch Illegal ??
Mir völlig wurscht ob ep möglich oder nicht, nach dem weißen Zug ist natürlich Kd8 möglich !!
also ist auch 1. ... exd6ep 2. Kd8 dxe7+ 3. Kc8 e8DT# eine Lösung, sonst müßte auf die erzwungende Rochade in der Aufgabenstellung hingewiesen werden ! (2020-12-11)
A.Buchanan: Can one tolerate “tolerated dual promotion” in Valladao? Well Valladao himself did, so it must be ok (2022-03-26)
more ...
comment
Keywords: a posteriori (AP) (Type Petrovic), En passant as key, Castling (sg), Promotion (D/T), Tolerated dual promotion (D/T), Valladao Task, Retro Strategy (RS)
Genre: h#, Retro
Computer test: Popeye 4.61
FEN: r3k3/2p1p3/2P1P3/2KpP3/8/8/8/8
Reprints: V. Die Schwalbe 33, p. 323, 06/1975
2001 CHM Themes 12/2000
3 Die Schwalbe 215, p. 240, 10/2005
A1 Problemkiste 169 02/2007
402 Eigenartige Schachprobleme , p. 130, 2010
14 Die Schwalbe 241, p. 374, 02/2010
Input: Gerd Wilts, 1995-06-03
Last update: A.Buchanan, 2022-06-10 more...
76 - P0003423
Matti Arvo Myllyniemi
3975 Stella Polaris 01/1971
P0003423
(7+11)
h#3 (AP)
0.2.1...
1. ... g6 2. 0-0 gxf7 3. Kh8 Le5#
1. ... cxb6ep 2. 0-0-0 bxa7 3. Td7 a8=D#
play all play one stop play next play all
Henrik Juel: Analysis
If Black can castle both ways, last move must have been b7-b5
C+ Popeye 4.61 (2020-10-25)
comment
Keywords: a posteriori (AP), En passant as key, Castling (sksg), Promotion (D), Valladao Task
Genre: h#, Retro
FEN: r3k2r/p4p1p/8/KpP2PP1/4PBpP/5pp1/6p1/8
Reprints: Nordisches Turnier 1970-1971
Input: Gerd Wilts, 1995-06-03
Last update: James Malcom, 2020-10-25 more...
77 - P0003428
Imre Sztankovszky
130 Die Schwalbe 09-10/1953
P0003428
(10+11) C+
h#1.5 AP
1. ... fxg6ep 2. 0-0 gxh7#
play all play one stop play next play all
Henrik Juel: Black captured c7xd6x..x.h2 and once more with an officer
White captured a2xb3, b2xa3, f2xg3, and e.g. exf and once more
If Black may castle, last move was g7-g5 (not b7-b6, because of Lh3) (2022-04-26)
Henrik Juel: HC+ Popeye 4.61
The castling serves two purposes:
enabling the mate and legitimizing the ep capture
so the a posteriori legitimizing is impure (2022-04-26)
A.Buchanan: Yes Henrik. The absence of other candidate solutions (beginning e.p. but excluding 0-0) declines an opportunity to embed additional content, and is arguably an artistic defect. However this very early AP problem is sound, and purity of motive is just one school. IMHO, there is space for such problems, if other content compensates: e.g. 4k2r/p2ppp1p/p7/5PpK/8/1PBB2Pb/2PP2Pp/8 h#2* AP. (2022-04-27)
more ...
comment
Keywords: En passant as key, Castling (sk), a posteriori (AP) (Type Petrovic), Volet Pawn
Genre: h#, Retro
Computer test: HC+ Popeye 4.61
FEN: 4k2r/3ppp1p/1p6/4BPpK/P7/pPPB2Pb/3P2Pp/8
Input: Gerd Wilts, 1995-06-03
Last update: A.Buchanan, 2022-04-27 more...
78 - P0003430
Tomislav Petrovic
(XIII) Problem 37-40 09/1956
P0003430
(4+15) cooked
h#3
1. dxe3ep Td1? 2. Le4 Txd6 3. Sg4 Tf6# usual AP try
1. dxe3ep 0-0-0! 2. Le4 Txd6 3. Sg4 Tf6#
play all play one stop play next play all
Cook: 1. dxe3ep 0-0-0 2. Lf5 Txd6 3. Sf3 Td4#
not
1. dxe3ep Td1? 2. Lf5 Txd6 3. Sf3+ ?? usual AP try and also check Bl 3rd move
Henrik Juel: The missing black man is dark-squared, so last move was not fxe4. In the solution Td6 should read Txd6. There seems to be a 'variation': 0... dxe3ep 1.0-0-0 Le4 2.Txd6 Sg4 3.Tf6#, so maybe the stipulation should have 1.1;2.1;1.1 added. (2004-09-23)
A.Buchanan: Well spotted Henrik. Your kindness is legendary, but in the harsh world of helpmates, this counts as a cooked I believe. In fact, it's the 2.Lf5/3.Sf3 version which is the cook. Yours is the actual solution! This is because 3.Sf3 would check if White hadn't castled, and we want the castling to be because of the AP condition only, not for 2 reasons, which one might term a "logical dual". The problem can be trivially fixed by sticking sSe5 on f6. I have no idea whether it's a typo or a cook: WinChloe has the same diagram, and the same confusion between cook & intended solution. The fact that AP does not appear in the stipulation might be a factor. WinChloe as far as I can see has no concept of AP as a keyword, and just picks it up based on a substring of the stipulation. In other compositions, T.Petrovic wasn't shy of flashing the AP acronym in the stipulation, but as usual the detective work peters out in the absence of primary records. (2020-12-08)
more ...
comment
Keywords: En passant as key, Castling (wg), a posteriori (AP) (Type Petrovic), Superseded by (P1382802)
Genre: h#, Retro
Computer test: Popeye v4.85 + retro thinking
FEN: 1n1r4/p1B4p/1p1p4/p3n1p1/3pPk2/3b4/pr5q/R3K3
Reprints: (60) Problem 144-147 12/1971
Input: Gerd Wilts, 1995-06-03
Last update: A.Buchanan, 2022-02-16 more...
79 - P0003431
Tomislav Petrovic
(XIV) Problem 37-40 09/1952
P0003431
(8+14) cooked
h#3
1. cxb3ep+ c3 2. Ld3 Tg4+ 3. Kh1 0-0-0#
play all play one stop play next play all
paul: White captures was axb and g2xh3, so the retro move c3xb4 is not possible. If b3-b4, the retro check is not justified. So last move was b2-b4 (preceded by Rc3-a3). (2011-08-06)
A.Buchanan: There's a lot to unpack here. Firstly, the intended solution has a definite transposition dual in the order of Black's 2nd & 3rd moves. Second, the pawn capture count is wrong. What happened to the original wBh, given sBh2 never left the h-file? Thirdly, and not a showstopper, White's only solution involves castling: so that's a logical dual. (2020-12-08)
more ...
comment
Keywords: En passant as key, Castling
Genre: h#, Retro
Computer test: Popeye v4.85 + retro thinking
FEN: 3r3n/p7/pp6/bP1q4/RPp5/r3p2P/p1P1P1kp/Rb2K3
Reprints: (61) Problem 144-147 12/1971
Input: Gerd Wilts, 1995-06-03
Last update: A.Buchanan, 2021-05-02 more...
80 - P0003434
Jozsef Bajtay
2432 Problem 101-102 09/1966
P0003434
(10+11)
h#2
1. fxg3ep 0-0 2. Lg4 hxg3#
play all play one stop play next play all
A.Buchanan: Surely diagram typo. Change to sBh3, then everything works (2022-03-29)
Mario Richter: Yes, Pawn h3 is black (2022-03-29)
comment
Keywords: a posteriori (AP) (Type Petrovic), En passant as key, Castling (wk)
Genre: h#, Retro
FEN: 8/b3p3/4p3/6pp/2P2pPk/1pPP3p/2PP1P1P/r2bK2R
Input: Gerd Wilts, 1995-06-03
Last update: Mario Richter, 2022-03-29 more...
81 - P0003442
Janko Furman
3. Makuc-Moder-Gedenkturnier 1971-1973
1. Preis
P0003442
(12+13) C+
h#2
b) wBd4 nach d5
a) 1. cxd3ep Sd5 2. 0-0 Se7#
NOT 1. ... Lf6? 2. Kf8 Tx8# because no AP justification
b) 1. Kd7 Lf6 2. Te8 Sxb6#
NOT 1. 0-0? Tf6 2. Kh8 Txf8# because rights lost
play all play one stop play next play all
Assume that bPb6 is really on b7:
Bl captures: dxe, exf, fxg, a|
Wh captures: bxc, g|, cxd=
So all pawn captures are accounted for.
a) If Bl 00 rights remain, then only way to give Black a prior move is by d2-d4.
b) No way to give Black a prior move, so Bl 00 rights must be lost.
Cook: NL
b) 1. 0-0 Tf6 2. Kh8 Txf8#
A.Buchanan: There is definitely something wrong here, with both (a) & (b). I think it's a simple diagram error: bPb7 has been misplaced on b6. Then the AP logic for (a) works great, and the castling "NL" for (b) is seen to be a thematic retro try. Can anyone confirm? (2022-03-21)
comment
Keywords: a posteriori (AP) (Type Petrovic), En passant as key, Castling (sk)
Genre: h#, Retro
Computer test: Popeye v4.87 & simple retro-logic
FEN: 4k2r/7p/1pR5/2P5/NNpP4/KB2PPPP/p3pppq/B4bnr
Input: Gerd Wilts, 1995-06-03
Last update: A.Buchanan, 2022-03-21 more...
82 - P0003444
Janko Furman
Miroslav Stosic

7273 Schach-Echo 11/1972
P0003444
(8+14) cooked
h#2
b) Gespiegelt (a1<->h1) & wKd1->e1
a) 1. dxe3ep 0-0 2. Te4 Txf3#
b) 1. exd3ep 0-0-0 2. dxe2 Ld5#
play all play one stop play next play all
Cook: a) 1. Kg3 Kf1 2. Df4 Th3#
1. Kg3 0-0 2. Dh4 Txf3#
A.Buchanan: This is a very heavy position to prevent the possibility of R: 1. c2xb3. I'm not sure why this was done. A much lighter position 8/8/8/5np1/1r1pPkr1/2Bp1p2/1p1P2P1/4K2R with 7 less units achieves the mates soundly. Am I missing something? (2022-02-16)
comment
Keywords: a posteriori (AP) (Type Petrovic), En passant as key, Castling (wk,wg), Superseded by (P1399967), Twinning by board reflection
Genre: h#, Retro
FEN: 8/1p1pp1p1/8/5nq1/1r1pPkr1/1PBp1p2/Pp1P2P1/2n1K2R
Input: Gerd Wilts, 1995-06-03
Last update: A.Buchanan, 2022-03-23 more...
83 - P0003451
Knud H. Hannemann
3823 Stella Polaris 03/1971
P0003451
(6+8) cooked
h#2 AP
1) 1. gxf3ep 0-0-0 2. Te7 Th4#
2) 1. cxd3ep 0-0 2. Da4 Txa4#
play all play one stop play next play all
AB: I think this is cooked. I don't see how AP castling can justify earlier ep here.
1. Kf3 Bf5 2. Kg2 Be4#
1. Kf3 Bxd7 2. Kg2 Bc6#
1. cxd3 Rf1 2. Qa4 Rxa4#
1. gxf3ep Ra5 2. Kf4 Rh4#
1. gxf3ep Rd1 2. Re7 Rh4#
Also, second solution given has typos 1. *c*xd3ep & 2. Q*a*4. (2002-03-21)
V.Liskovets: Indeed this problem is cooked, and I
failed to correct it preserving symmetry.
Here is a possible correction:
W: Ke1 Ra1 Rh1 Be6 Pd4 Pf4
B: Ke4 Qd7 Rf7 Pc3 Pc4 Pe3 Pg3 Pg4 Bh7 Rb2 Pb3

Another story, justifying its contents (e.p.).
In my opinion, all published treatments are
insatisfactory (the same concerns P0004295
(corrected) by Werner Kuntsche as well).
There is a way to make it sound under the
sophisticated HYBRID stipulation 'AP, pRA':
2 partial solutions legalized JOINTLY by
both castlings (details to be published). (2002-04-02)
VL: Sorry, Andrew, I see no issues with my version (excepting the mentioned special retro-convention/genre for justifying its soundness, of course!). h#2 (pRA&AP). The full solution consists of 2 partial AP-based ones: I 1.gxf3 e.p.(!?) 0-0-0! 2.Te7 Th4#; II 1.cxd3 e.p.(!?) 0-0! 2.Da4 Txa4#.
C+(popeye): h#2 & two h#1.5 after the keys (added Tb2 & Lh7 are cookstoppers).
The main features of the corresponding suitable AP-genre ("consolidated")are rather clear. However, not all subtle aspects have been analyzed thoroughly yet. I have got only a draft manuscript with few examples, still. (2022-02-15)
A.Buchanan: Yes Valery I think your correction is technically sound. Theological debate is something else, heh. Have posted it as separate entry. Please tell me if it was original here, or appeared somewhere else (2022-05-29)
Henrik Juel: I looked in Stella Polaris 1971 p.202-203, where the problem is given without AP under the diagram
It turns out that both the author and the editor (Jan Knöppel) consider the problem correct
After explaining the AP convention the latter writes [my translation]:
The author has elegantly applied the AP conventionn in combination with partial analysis. If we assume that last move was not f2-f4 or T-h1, the possibilities d2-d4, K-e1, and T-a1 remain, but the two latter disappear by 1... 0-0. The other assumption (last move was not d2-d4 or T-a1) is analogous.
So something like Valery's HYBRID stipulation was the intention, and the problem was considered correct in 1971; later developments do not change this fact (2022-05-29)
Henrik Juel: Of course, the obvious cooks 1.Kf3 etc. remain (2022-05-29)
A.Buchanan: I agree. Basic AP Petrovic does not allow multiple ep, nor does RS, so some alignment of AP with PRA is required. I would like to see this codified though. (2022-05-29)
more ...
comment
Keywords: Castling (wb), En passant as key (2), a posteriori (AP) (Type Petrovic), Superseded by (P1401508, P1399178), Quasi-symmetrical position, Partial Retro Analysis (PRA)
Genre: h#, Retro
Computer test: C- Forward play cooked by Popeye v4.87. However the basic idea is good, if a suitable theoretical framework can be found.
FEN: 8/3q1r2/4B3/8/2pPkPp1/2p1p1p1/8/R3K2R
Reprints: (65) Problem 144-147 12/1971
(8) StrateGems SG19, p. 156, 07/2002
Input: Gerd Wilts, 1995-06-03
Last update: A.Buchanan, 2022-05-29 more...
84 - P0003460
Johannes Jacob Burbach
1089 Probleemblad 06-08/1950
P0003460
(5+4) C+
h#3
1. 0-0-0 Ke2 2. Kb8 Td1 3. Ka8 Txd8#
play all play one stop play next play all
1. 0-0-0 0-0? usw. ist nicht erlaubt, da Weiß im letzten Zug einen schwarzen Stein geschlagen haben muss, was nur mit K oder T ging.
A.Buchanan: Tried some experiments to see if other formats of position entry are possible, but having reverted those, the solution animation no longer works. The usual PDB problem of hidden & undocumented variables between stipulation, diagrams & solution animation, sigh. (2021-04-01)
A.Buchanan: I found a "fix" - I added a bogus empty diagram b! How random! (2021-04-02)
A.Buchanan: For h#, I run Popeye without opti noca, so can see all the retro tries explicitly. For d#, I would apply opti noca if relevant, else lines of play might be blocked (2021-04-04)
more ...
comment
Keywords: Cant Castler, Castling (wksg), Homebase (s)
Genre: h#, Retro
Computer test: Gustav 4.1d C+ Popeye 4.61 also (with 'opt noc h1')
FEN: r3k3/pp6/8/8/8/6P1/5P1P/4K2R
Reprints: (27) Problem 144-147 12/1971
Input: Gerd Wilts, 1995-06-03
Last update: A.Buchanan, 2021-04-04 more...
85 - P0003463
Tivadar Kardos
1501 Problemista 68 03/1967
P0003463
(9+16) C+
h#4
b) sSe6->d5
a) 1. Df7! h4 2. 0-0 h5 3. Lh8 hxg6 4. Sg7 gxh7#
b) 1. De6! h4 2. 0-0 h5 3. Kh8 hxg6 4. Tg8 Txh7#
play all play one stop play next play all
1. 0-0-0?? 0-0 2. Kb8 Txc1 3. Ka8 Txb1 4. Tb8 Ta1# (0-0-0 ist illegal)
1. 0-0-0? ist nicht gestattet, da wegen der schwarzen Bauernstellung der weiße Bauer a2 auf a8 umgewandelt wurde!
Marcin Banaszek: Die Urdruck-Stellung ist mit sSe6 (nicht sSd5) und über die Zwilling-Stellung ist dort keine Erwähnung. (2022-06-09)
A.Buchanan: Yes I don't know where the twin concept was introduced to this record (the visible "Last Update" history does not mention it). The twin adds very little: the retro try is identical, and the solution shares most of the White moves, ending in a duller mate. Did the 1983 reprint include it? (2022-06-10)
more ...
comment
Keywords: Cant Castler, Castling (wksg)
Genre: h#, Retro
Computer test: HC+ Popeye v4.87 & trivial retro-logic
FEN: r3k1qr/1p5p/4n1p1/4bp2/6p1/1pPpPpP1/1P1P1P1P/1bn1K2R
Reprints: 144 200 Ausgewählte Schachaufgaben 1983
Input: Gerd Wilts, 1995-06-03
Last update: A.Buchanan, 2022-06-10 more...
86 - P0003501
Jozsef Korponai
1224 Sahovski glasnik 01/1970
P0003501
(6+3)
h#2*
2 solutions
set play
1. ... a3,a4? 2. cxb2 0-0?# castling illegal
1. ... Kf1! 2. cxb2 Kg2#
1. ... Kd1! 2. cxb2 Kc2#
actual play
1. c3xb2? 0-0#? castling illegal
1. Ka1! Kd1 2. c2+ Kxc2#
1. Kxb2! Kd1 2. Ka1 Kc2#
play all play one stop play next play all
A.Buchanan: Not sure what is intended here: should one include set play? It's a bit unclear. (2021-10-22)
A.Buchanan: I have marked stip from "h#2" to "h#2* 2 solutions" (2021-10-22)
comment
Keywords: Cant Castler, Castling (wk)
Genre: h#, Retro
Computer test: HC+ Popeye v4.87 & trivial retro-logic
FEN: 8/8/8/8/1N6/2p1p3/PP2P3/1k2K2R
Input: Gerd Wilts, 1995-06-03
Last update: A.Buchanan, 2021-10-22 more...
87 - P0003505
Herbert Hultberg
7153 The Fairy Chess Review 02/1947
P0003505
(2+3) C+
h#3
1. Kb3 Kd2 2. Ka4 Kc3 3. a2 Txa2#
Nicht 1. Kb3 0-0-0?? 2. Ka2 Td3 3. Ka1 Txa3#, da Weiß zuletzt mit K oder T gezogen hat
play all play one stop play next play all
more ...
comment
Keywords: Cant Castler, Castling (wg), Minimal, Miniature, Homebase
Genre: h#, Retro
Computer test: HC+ Popeye v4.87 + trivial retro logic
FEN: 8/8/8/1p6/8/p7/2k5/R3K3
Input: Gerd Wilts, 1995-06-03
Last update: A.Buchanan, 2023-06-29 more...
88 - P0003506
Leon Loewenton
4495 Arbejder-Skak 06/1957
nach T.R. Dawson
P0003506
(7+5) C+
h#2*
* 1. ... Tg1 2. e4 g3#
1. Kg3 0-0? 2. h4 Tf3#
1. Kg3 Tg1! 2. Kh4 g3#
play all play one stop play next play all
A.Buchanan: I suppose set play is intended, because it's the only function of sBe5. But I think it's fine without. (2021-10-22)
Mario Richter: The set play is not mentioned in the "official solution" ('Arbejder-Skak' 08/1957 p.231) (2021-10-23)
A.Buchanan: Thanks. Odd! (2021-10-23)
comment
Keywords: Cant Castler, Castling (wk)
Genre: h#, Retro
Computer test: HC+ Popeye v4.87 & trivial retro-logic I suppose set play is intended, because it's the only function of sBe5. But I think it's fine without.
FEN: 8/8/8/4p1pp/6Pk/4p2P/4P1PR/4K2R
Input: Gerd Wilts, 1995-06-03
Last update: A.Buchanan, 2021-10-24 more...
89 - P0003529
Tivadar Kardos
2290 Magyar Sakkelet 05/1957
P0003529
(10+11) C+
h#2
1. Kxd8 Ta7 2. Te8 Td7#
1. 0-0 0-0-0? 2. Th8 Txg1# try
play all play one stop play next play all
White castling not possible as sBa= on a1. (Also sBgxh= or sBhxg=)
A.Buchanan: Retractor 2.1.1 does not notice that promotion happened on a1 (2021-11-25)
comment
Keywords: Cant Castler, Castling (wgsk), Sacrifice of white pieces
Genre: h#, Retro
Computer test: HC+ Popeye 4.87 & trivial retro-logic
FEN: b1nNk2r/8/3p4/2Pp1B2/3p4/1PpPp3/1PPP4/R3K1nr
Input: Gerd Wilts, 1995-06-03
Last update: A.Buchanan, 2021-11-25 more...
90 - P0003531
Tivadar Kardos
3563 Schach 24, p. 379, 12/1960
P0003531
(10+12)
h#2
1. Kd8 Txh8 2. Kc8 Txg8#
play all play one stop play next play all
Henrik Juel: The diagram pawns captured all missing men, so the missing pawns [Pa7,Ph2] promoted on a1,h8, and Ta1,Th8 have moved, ruling out the apparent solutions 1.Ke7 0-0-0 2.Kxe6 The1#, 1.Dg1+ Ke2 2.0-0 Taxg1#
C+ by Popeye 4.61 and analysis (2021-08-31)
comment
Keywords: Cant Castler, Castling (wbsg)
Genre: h#, Retro
FEN: r3k1qr/1bp2p2/2p1Pp2/5p2/1P3p2/1Pp2P2/1PP2P2/R3K2R
Reprints: 664 Die Schwalbe 05/1961
Input: Gerd Wilts, 1995-06-03
Last update: A.Buchanan, 2022-02-10 more...
91 - P0003539
Hilding Fröberg
332 Land og Folk 01/03/1970
P0003539
(4+2) C+
h#1.5
1. ... Sf5? 2. 0-0-0?? Sd6# - aber die s0-0-0 ist illegal, den zuletzt muß Schwarz mit K oder T gezogen haben.

1. ... Sg8! 2. Td8 Sc7#
play all play one stop play next play all
Mario Richter: Luboš Kekely (Slovakia) correctly points out that 1. ... Sf5! 2. 0-0-0 Sd6# is only a try and not a solution (last black move must have been by King or Rook, so the queenside castling is illegal).
I have changed the solution accordingly. (2023-03-10)
comment
Keywords: Cant Castler, Castling (sg), Aristocrat, Minimal, Miniature, Homebase
Genre: h#, Retro
Computer test: HC+ Popeye v4.87 & trivial retro-logic | C+ rawbats
FEN: r3k3/6K1/N6N/8/8/7B/8/8
Input: Gerd Wilts, 1995-06-03
Last update: Mario Richter, 2023-03-10 more...
92 - P0003561
Tivadar Kardos
5396 Arbejder-Skak 01/1961
P0003561
(7+15) C+
h#2
not: 1. Sd7 0-0-0 2. 0-0 Tdg1#
but: 1. Lxe6 Th7 2. 0-0-0 Ta8#
and: 1. Ke7 0-0 2. Kxe6 Tae1#
play all play one stop play next play all
A.Buchanan: Hard to see this as anything but a version of P0563989. Popeye reports again three apparent mates, with all 4 castlings. However pawn capture analysis indicates that Black qside castling is illegal, while the two kingside castlings are mutually exclusive. (2022-01-08)
comment
Keywords: Cant Castler (sg), Castling, Partial Retro Analysis (PRA)
Genre: h#, Retro
Computer test: Popeye v4.87
FEN: r1b1kn1r/1n6/1pPpP3/1p1p4/1bp5/1qpp4/1PP5/R3K2R
Input: Gerd Wilts, 1995-06-03
Last update: A.Buchanan, 2022-01-08 more...
93 - P0003574
Juan Ruiz Luque
Jean-Claude Gandy

RA37 diagrammes 21 05-06/1976
P0003574
(8+8)
h#2
Duplex
s) 1. Kf8 g6 2. Se7 Th8#
w) 1. ... Kg1 Td8 2. Lh2 Td1#
play all play one stop play next play all
Rochaderecht
Henrik Juel: C+ by Popeye 4.61 and analysis
If Black could castle, both parts would be cooked:
1.0-0-0?? Txg7 2.Sb8 Tc7#
1... Kg1 0-0-0?? 2.Lh2 Td1#
But he cannot, because explaining wTh7 by h7-h8=T would require nine white pawn captures, and only eight black men are missing; hxg8=T is even worse, requiring 10 captures (Th7 being original or promoted further towards the west obviously would require Ke8 to have moved) (2021-07-04)
comment
Keywords: Cant Castler, Castling (sg)
Genre: h#, Retro
FEN: r3k3/6pR/2n4p/5PP1/pp5b/5PBP/6PK/8
Input: Gerd Wilts, 1995-06-03
Last update: A.Buchanan, 2021-07-03 more...
94 - P0003607
Ljubomir Ugren
anonymous

881 Sahovski glasnik 07-08/1964
P0003607
(6+8)
h#3*
*) 1. ... 0-0 2. Tf8 Tc1 3. Tf7 Tc8#
1) 1. 0-0 Tf1 2. Txf2 Txf2 3. Kh8 Tf8#
play all play one stop play next play all
Henrik Juel: Ljubomir had an anonymous co-author?? (2021-10-31)
A.Buchanan: I've no idea. There were three users called "anonymous" of different kinds, so I merged them into one, that's all. (2021-10-31)
A.Buchanan: I think anticipated in 1948 by Peter Kniest P0003606 (2021-11-01)
comment
Keywords: Cant Castler, Castling (wksk), Homebase (w), anticipated (P0003606)
Genre: h#, Retro
FEN: 4k2r/3pp1pp/pp6/8/8/8/3PPP1P/4K2R
Input: Gerd Wilts, 1995-06-03
Last update: A.Buchanan, 2021-11-01 more...
95 - P0003622
Janez Moder
Problem 07/1962
P0003622
(4+3)
h#2*
*) 1. ... 0-0 2. Dh4 Txf3#
1) 1. Kh4 Kf2 2. fxg2 Sf3#
play all play one stop play next play all
SCHRECKE: NL: 1. Dg5,Kh4 gxf3 2. Kh4,Dg5 Sf1# (2023-09-13)
Ladislav Packa: Retro content is not needed, the solution is preserved even without it.
Pg4 Pg2 Sh2 Ke1 Rh1 (5)- Ph5 Kg3 (2) h#2* C+

1...0-0 2.h5-h4 Rf1-f3 #
1.Kg3-h4 Ke1-f2 2.h5*g4 Sh2-f3 # (2023-09-14)
comment
Keywords: Cant Castler, Castling (wk)
Genre: h#, Retro
FEN: 8/8/8/8/6q1/5pk1/6PN/4K2R
Input: Gerd Wilts, 1995-06-03
Last update: hpr, 1997-05-07 more...
96 - P0003630
Jean Oudot
652 Themes-64 10-12/1962
P0003630
(9+6) C+
h#2*
*) 1. ... 0-0#
1) 1. Dc5 Tf1+ 2. Df2 Txf2#
play all play one stop play next play all
A.Buchanan: White must cover g2. If not by check, then by rook. However Rf1 is check, so both Black's moves are forced to shield bK (2022-11-27)
more ...
comment
Keywords: Cant Castler, Castling (wk)
Genre: h#, Retro
Computer test: HC+ Popeye 4.61
FEN: b7/7P/6p1/6qP/6Pb/3p1kNP/3P3P/4K2R
Input: Gerd Wilts, 1995-06-03
Last update: A.Buchanan, 2022-11-27 more...
97 - P0003657
Vladislav Bunka
4204 Problem 12/1979
P0003657
(4+13)
h#3
b) -wBh3
a) 1. Txh3 0-0 2. Sd3 Txf4 3. Sd6 Ta4#
b) 1. Sd1 Txh4 2. Se3 Th8 3. Sd6 Ta8#
play all play one stop play next play all
Adrian Storisteanu: Fix for P0534003. (2015-10-17)
Yuri Bilokin: version: -wPh3 8/3pq3/5r2/5p2/4nppp/5r2/kpP2n2/1b2K2R (3+13) h#3 2 Sol
1.Rh3 0-0 2.Sd3 Rxf4 3.Sd6 Ra4# (MM)
1.Sd1 Rxh4 2.Se3 Rh8 3.Sd6 Ra8# (MM) (2021-06-24)
Ladislav Packa: Yuri: Without wPh3, the castling is impossible (theme Cant Castler). But after wPh3-d5, two solutions are OK. (2021-06-24)
Yuri Bilokin: Ladislav, sorry, was wrong. Thank you for your vigilance. A slightly different edition, too, with two solutions: 8/2p1q3/4r3/5pPp/4npp1/5r2/kpP2n2/1b2K2R (4+13) h#3 2 Sol. (2021-06-25)
comment
Keywords: Cant Castler, Castling (wk)
Genre: h#, Retro
FEN: 8/3pq3/5r2/5p2/4nppp/5r1P/kpP2n2/1b2K2R
Input: Gerd Wilts, 1995-06-03
Last update: Gerd Wilts, 1999-04-24 more...
98 - P0003659
Maurice Jago
8202v Die Schwalbe , p. 148, 10/1980
version of cooked 8202 Die Schwalbe 08/1977
P0003659
(12+14) cooked
h#2
b) sBb4->b5
a) 1. Sf2 Dxf2+ 2. Kh1 0-0-0#
b) 1. Sb1 Ta4 2. Sxg3 Th4#
play all play one stop play next play all
Cook: 1. Txg3 0-0-0 2. La2/Lb3/Lc4 Dxh1#
1. La2/Lb3/Lc4 0-0-0 2. Txg3 Dxh1#
See P0000642
Mario Richter: Hier stimmt was nicht: In der Schwalbe 08/1977 erschien als Nr. 2202 die folgende Aufgabe: 8/Pp2p1p1/Bq1p3p/4p3/1p6/n4pP1/1PPPPr1k/R3KQ1n/; h#2, b) sBb4 nach b5 und der AL a) 1.Tf2-g2 0-0-0 2.Tg2xg3 Df1xh1#; b) 1.Sa3-b1 Ta1-a4 2.Sh1xg3 Ta4-h4#, die aber in a) und b) durch 1.f3xe2 La6xb7 2.Tf2-g2 Df1xg2# nebenlösig war.
Ferner ist in nebenstehendem Diagramm sowohl in a) als auch in b) sehr wohl die w0-0-0 möglich
(Weiß hat als letzten Zug Sc7-a8, Schwarz entwandelt auf b1 und h1, Weiß auf f8); und selbst unter der Annahme, daß die w0-0-0 nur mit sBb4 möglich sei, wäre die Aufgabe mehrfach nebenlösig, z.B. durch 1.Ld5-c4 0-0-0 2.Bd6-d5 Df1xh1# (2010-05-29)
A.Buchanan: Here, the source was incorrectly given as "8202v Die Schwalbe 08/1977". No, that was where the original problem appeared, with a "v" added. So when, as here, we suspect the diagram has a typo, we would have to look through many back issues of Die Schwalbe! It turns out that this correction (itself cooked) was published (as text) in p.148 Die Schwalbe 65 10/1980. I have corrected the source. The reference to the earlier buggy problem can then cleanly be put in the "after" field. The actual typo was missing bPc7. (2024-03-06)
A.Buchanan: With the typo fixed, I think the retro-logic works ok, but there is a single cook family, fortunately repairable (2024-03-06)
more ...
comment
Keywords: Cant Castler, Castling (wg), Superseded by (P1415606)
Genre: h#, Retro
Computer test: both twins cooked by Popeye v4.87
FEN: NBr5/N1p3p1/1qPpp3/2bb4/1p6/n1pP1rP1/1PP1P2k/R3KQ1n
Input: Gerd Wilts, 1995-06-03
Last update: A.Buchanan, 2024-03-06 more...
99 - P0003666
Johannes Jacob Burbach
Alain C. White

392 Probleemblad 03/1947
P0003666
(5+8) C+
h#2
b) +sBh6
a) 1. Ta6! Lxc3 2. Ta2 0-0#
b) 1. Ta5! Kd1 2. Ta2 Kc2#
play all play one stop play next play all
Henrik Juel: In b) White may not castle, because last move was done by Ke1 or Th1 (2023-12-02)
A.Buchanan: Cute. Do wPe2 & bPe3 serve any purpose, however? (2023-12-03)
more ...
comment
Keywords: Cant Castler, Castling (wk), Superseded by (P1413924)
Genre: h#, Retro
Computer test: HC+ Popeye 4.61 & simple retro logic
FEN: 5n1q/6B1/5r2/7r/5p2/2p1p3/1P2P3/k3K2R
Input: Gerd Wilts, 1995-06-03
Last update: A.Buchanan, 2023-12-05 more...
100 - P0003720
Yuri Markovic Gordian
4496 Schach , p. 30, 01/1964
P0003720
(7+9) C+
h#2
b) sBh7 nach a7
a) 1. 0-0 Tcg3 2. Sh8 Txg7#
b) 1. 0-0-0 Tgc4 2. Sb8 Txc7#
play all play one stop play next play all
more ...
comment
Keywords: Castling, Cant Castler, Obvious promotion (T), Twin
Genre: h#, Retro
Computer test: HC+ Popeye 4.61
FEN: r3k2r/1pp3pp/2n3n1/8/6R1/2R4P/4P1P1/5B1K
Input: Gerd Wilts, 1995-06-03
Last update: Gunter Jordan, 2022-12-01 more...
Show statistic for complete result. Show search result faster by using ids.

https://pdb.dieschwalbe.de/search.jsp?expression=COMMENTDATE%3E%3D20200919+AND+NOT+S%3D%27Buffalo+Sunday+Times%27+AND+G%3D%27h%23%27

The problems of this query have been registered by the following contributors:

Gerd Wilts (100)