Fairy chess composition

# No.1731 (AG&DP)

Original Fairy problems
JF-2023:
01.01.2023 - 31.12.2023

☝ Definitions

No. 1731 Armin Geister &
Daniel Papack

Germany
original - 11.02.2023

white Be1 Kf2 Pd2a2f3a3h4f4 Sh3 Rg8 Qb8 black Sc1c8 Rh2e2 Pg3h6a6h7f7a7 Kh5 Qf8 Bd8
h#4            2 solutions            11+13
Mars Circe
Anti-Mars Circe

Solution: (click to show/hide)

Subscribe
Notify of

Inline Feedbacks
Siegfried Hornecker
February 11, 2023 17:00

Mars Circe
Anti Mars Circe

The “Anti Mars Circe” is not on this page, but seems intended.

Also typo in the second solution:

Spoiler
Final move should be f3xg3, not fxg3 (otherwise f4xg3 is possible also).

Kjell Widlert
February 20, 2023 23:38

This combination is really simpler than the two single conditions: every move must start with a Circe rebirth, whether it is a capture or not. It is also more limited, as there are so few possible moves – in this position, for example, Black has only 8 legal moves! But complex logic can be shown anyway, of course, as this problem proves.
Black’s Kh5 can only be attacked and mated by a wQ from d1 or a wR from h1, so we must clear either h1-h5 or d1-h5. In addition, Black must play Sb1-b8-d7 and Sc8-g8-e7 in order to block the two flight-squares for the bK. So the wQ and the wR must get away from b8/g8, either by white moves from d1/h1 or by black captures from a8/h8 – in which case the capturing bR must itself get away by a move from h8/a8. It turns out this last procedure is the only one that works.
There is a (non unique) set-play with mate from h1: 1…Rg8-h1-f1 2.Sc8-g8-e7 Qb8-d1-a4 3.Rh2-h8-g8 Sh3-b1-c3 4.Sc1-b8-d7 h4-h2xg3#, but Black has no tempo move (f7-f7-f6 gives a new flight on f7). There is no analogous set-play with mate from d1, as White must play three moves in sequence to clear the diagonal (Be1-c1-b2, Kf2-e1-f1, Pf3-f2xg3) so he has no time to move both the wQ and the wR.
For a mate from h1, Qb8 isn’t needed, so Black can start with Re2-a8xb8. As he must later clear b8 by R-h8-g8, he has no time to move Rh2, so White must play R-h1xh2 and must follow this with R-a1-b1 as h2 is of the wrong colour – but this blocks b1, so S-b1-c3 must be played before R-a1-b1. Black’s play echoes this: S-g8-e7 must be played before R-h8-g8.
For a mate from d1, Rg8 isn’t needed, so Black can start with Rh2-h8xg8. As he must later clear g8 by R-a8-b8, he has no time to move Re2, so White must play Q-d1xe2. White must then continue B-c1-b2, K-e1-f1, and Pf3-f2xg3 in that order, and Black must play S-b8-d7, R-a8-b8, and S-g8-e7 in that order (Black must start with Sc1 also in order to clear c1 for Be1 – but Sc1 cannot be anywhere else because of a dual R-a1-d1 in the first solution).

Walter Lindenthal
March 7, 2023 14:35

“…EVERY move must start with a Circe rebirth, whether it is a capture or not” … this can NEVER EVER result in the given diagram positions! And as both conditions include the K, so the K could only be checkmated on very few squares!

Thus, even if you accept the given diagram as a valid starting position, it is immediately BLATANT the bK can only be checkmated on d8 or d7 by wQ via d1 (after possibly having moved himself) or on his current(?) position h5 by wR via h1 or by wQ via d1. There is NO other possibility! Such, the solution is no surprise, you just have to “remove” the in-between stones one after another.

And as long as you block d8 with bB vis-a-vis f8 with bQ there is no way to unblock any of them as long as a wP is blocking the d-file (only the wQ via d1 could resolve the blocking by capturing bBd8, but who could before remove the wP from the d-file as the bQ is immobile?…) – so, that’s a permanent deadlock situation… Whatever happens, the wK(f2) could NEVER EVER be in any danger, and equally the bK here could NEVER EVER be threatened on d7 (or d8)… So the final mate on h5 by wR or wQ is ABSOLUTELY clear from scratch… I don’t think this requires any complex logic, just simply “count” the moves…

Joost de Heer
March 7, 2023 15:25

Codex footnote 18:
18. Retroanalysis does not apply to illegal positions, except for the purpose of determining that they are illegal. It also does not apply to fairy compositions unless it is essential to the content of the composition.
Since retro isn’t essential for the content of the composition, the question ‘is this position reachable from the begin condition using the fairy conditions’ is irrelevant.

Last edited 1 year ago by Joost de Heer
Walter Lindenthal
March 7, 2023 17:22

Joost: I know Codex Art.14(1), and so they say “irrelevant”… like the “impossible” bKh5 here is for the solution? It’s a matter of perspective! And please notice that I conceded this in my post. Still, the combination of these conditions seems rather fruitless, this is what I wanted to express. May be a combination with mirror- or alternative-options of the Mars-conditions could raise the complexity…
And possibly there’s a truly “relevant” illegality in the diagram when considering the rule “…unless it is essential to the content of the composition”: 4 bPawns on files f-h (but no extra bPawn can have crossed to the g-file via f7)! As the bPh6 in the diagram can only be established under the given conditions from the g-file by his capturing on h6 a w stone (via g7) this would make bPg3 supernumerous (=”illegal”?), but the existence of bPg3 seems rather essential to the content of the composition in both solutions!
Of course you could alternatively say it’s bPh6 which is supernumerous, a Pawn with absolutely no relevance for the solution… But then, how did the relevant bP arrive at g3 under the given conditions?… It’s all a matter of perspective!

Last edited 1 year ago by Walter Lindenthal
Joost de Heer
March 9, 2023 12:47

Walter,
We’ve had this discussion before. Unless explicitly stipulated (e.g. ‘last N moves’ or ‘proofgame’), for fairy compositions the de facto standard is ‘there is no previous play’.

Last edited 1 year ago by Joost de Heer
Walter Lindenthal
March 10, 2023 11:08

Joost: …we had, you’re fully right, and I do NOT want to go against any prevailing interpretation! All I still want to find out is the “borderline” between essential(=relevant) and irrelevant as given by the “fairy section” of FN18 in the Codex’ rule 14(1), quote: “Retroanalysis… it also does not apply to fairy compositions UNLESS it is essential to the content of the composition”. What would be “essential to the content” then if not even the solution, nor the diagram, nor the positions of relevant stones, nor the moves, etc are? There must be something that is “essential” otherwise the rule would be completely meaningless… This is what I want to find out. The wording seems clear, the second sentence of Fn18 refers to “Retroanalysis… it ALSO does not apply to fairy compositions UNLESS…”, it does explicitely NOT refer primarily to “illegal positions…” (like in the first sentence)! Please read it word by word! Or is an interpretation prevailing that is deviant from the Codex? THANKS!

Joost de Heer
March 11, 2023 14:38

Easy test: Does the past of the diagram play a role in the solution?

Yes – Retro is relevantNo – Retro is not relevantFootnote 18 says that you can’t deduce any retro-aspects in illegal positions, so e.g. in an illegal position, ep capture will never be allowed as you can’t prove that the previous move was the double-step, since there is no history. The only application of retro analysis possible in an illegal position is to prove its illegality.

Last edited 1 year ago by Joost de Heer
Walter Lindenthal
March 11, 2023 16:51

Joost: Thanks, your example with the ep capture is catchy, this seems a good general example valid for each and every type of chess composition covered by the Codex. Provided (e.g.) an orthodox #2 also would be understood as a “fairy” composition insofar in the meaning of sentence2 of Fn18, everything would be clarified now, unfortunately I got misdirected by the term “fairy” then…, sorry!

Joost de Heer
March 11, 2023 21:48

Another way to look at this:
In an orthodox composition, retro content can be applied, e.g. to determine the legality of an ep-capture as key move, or to prove the illegality of castling.
If the retro analysis shows that the position is illegal (i.e. not reachable from the begin position by any legal sequance of moves), no retro logic can be applied to the position, which means that e.g. ep capture is not possible as a key, or that any castling is possible, and mutually exclusive castling doesn’t exist. This is what footnote 18 means.
In fairy land, the world is reversed: We assume that any position has no retro history, unless stated (either by the stipulation or by explicitly mentioning that retro content should be considered). Fairy pieces are considered to be promoted pawns in this case.
To quote Sam Loyd when someone asked him how a piece got to its diagram position in a clearly illegal position (https://pdb.dieschwalbe.de/P1054680): “Because I placed it there”.

Walter Lindenthal
March 12, 2023 10:19

Joost: Thanks for your efforts, I appreciate the focused discussion of this issue! The wording of the second sentence in Fn18 seems to require(!) an “essentiality”-check specifically in fairy(!) compositions. Insofar your statement (quote) “…In fairy land, …we assume that any position has no retro history, unless stated…” (= act of volition), seems to clash with Fn18 “Retroanalysis… does not apply to fairy compositions unless it is essential to the content of the composition…” (= a factual condition, independent of volition) – I have not yet succeeded to “decode” this seeming discrepancy… (Sam Loyd’s “free posit” might have referred to a NON-essential illegal positioning in his fairy composition). May be there is no good example for a fairy-specific essentiality, anyhow the general ep capture and castling issues can also occur in a fairy composition and could be essential to its content… and possibly this might be meant by the second sentence in Fn18. By the way, based on former discussions, I am not convinced everybody will agree to (quote) “Fairy pieces are considered to be promoted pawns” in fairy land…?

Joost de Heer
March 12, 2023 23:53

The statement about fairy pieces being promoted pawns is only applicable to compositions with retro content and fairy pieces in the diagram.

Kostas Prentos
March 14, 2023 01:46

Fairy pieces are considered to be promoted pawns for retro problems, but also to determine the rebirth square in many forms of circe, anticirce, etc.
While it is not a requirement for non retro fairy problems to have a legal (by orthodox means) diagram, it is always desirable to have a diagram that is economical in terms of legality. I know that some composers opt for legal diagrams in their fairy problems, even if it’s not a requirement, and only if that’s not possible, then accept a position that is illegal. But nobody would sacrifice the thematic content of a fairy problem in order to achieve legality.

shankar ram
March 16, 2023 18:05

Some fairy pieces could never have come about by promotion. For example, White Equihoppers or Alfils on odd-numbered ranks.
Also, the combination of some fairy pieces with Circe results in some funny results. For example, a White Fers or Vao captured on b3 is reborn on b8, though they could never have promoted there!
So, any retro problems would have to avoid these situations!

Kostas Prentos
March 17, 2023 08:03