No.751 (KM)

Karol Mlynka


Original Problems, Julia’s Fairies – 2015 (I): January – June

   →Previous ; →Next ; →List 2015(I)

Please send your original fairy problems to:

No.751 by Karol Mlynka – Nice thematic play of the two Pawns! (JV)


Circe: Captured units (not Ks) reappear on their game-array squares, of the same colour in the case of pieces, on the file of capture in the case of pawns, and on the promotion square of the file of capture in the case of fairy pieces. If the rebirth square is occupied the capture is normal.

Face to Face: When a white piece is just one rank below a black piece on the same file, they exchange their walk.

Phantom Chess: Any unit except a king may move either normally (from its current square) or as though from its Circe rebirth square if the latter is vacant.

No.751 Karol Mlynka

original – 14.03.2015

Solutions: (click to show/hide)

Black Ph2 White Ph7

h=2                                         (1+1)
b) Ph7->e6, c) Ph7->b6, d) Ph7->a4
Phantom Chess
(No Kings)

44 Responses to No.751 (KM)

  1. Luce Sebastien says:

    no solutions Winchloe 🙁

    • Kjell Widlert says:

      Must be a bug? I see nothing wrong with the composer’s solutions.

      I suppose Circe is just against cooks such as c) 1.h1Q b7 2.Qa8,Qc8 PxQ=.

      • Mike Neumeier says:

        Not stalemate there. Since Circe applies, bQ reborn on d8.

        • Nikola Predrag says:

          Yes, that’s what Kjell has said, without Circe that would be a stalemate and therefore a cook.
          A merely cook-stopping condition is “in principle” worse than a merely cook-stopping white officer.
          However, with only 2 Pawns on the board, the Circe “appears” as a very efficient and economical “invisible” tool.

  2. Nicolas Dupontdupont says:

    Interesting is to see why WinChloe doesn’t find any solution. In the diagram position the sequence is sound according to this program, but doesn’t lead to a stalemate, because of the (unique) legal move 3.Sb7. Idem for solution b), the (unique) legal move being now 3.Bc7.

    • Georgy EvseevGeorgy Evseev says:

      Then, this means that WinChloe considers that Phantom chess move of a pawn from source knight square is possible…

    • Kjell Widlert says:

      That’s a strange interpretation of the combination FTF + Ghost Chess. The idea must be that in the final position in a), the bSh6 moves like a bP (by FTF), and therefore it can move from its Circe rebirth square (by Ghost rules), which is considered to be b8 because the piece is still a bS, but it moves like a bP from that square?!

      I suspect that Py considers the rebirth square to be h7 for the bSh6 moving like a bP, and so Py would try to play Sh6-h7-h6 … which, however, is illegal in Py because the program doesn’t allow any null moves! That could be the reason why Py finds solutions where WinChloe doesn’t.

      • Nicolas Dupontdupont says:

        “I suspect that Py considers the rebirth square to be h7 for the bSh6 moving like a bP”.

        Indeed it is probably the case, but is it correct? I would say no, because nothing in the rules tells that bS changed its “deep nature” – which is to be a Knight – but only that, temporarily, it changes the way it normally moves.

        So IMO the rebirth square or the bSh6 should always be b8, even when its moving possibilities are modified according to the Face-to-Face condition (a rebirth is not a move).

        Now it is generally accepted – although not for every fairy condition – that a “delocalized Pawn”, e.g. bPb8, may move only one step – straight line without capture and diagonally with capture. And indeed if one adds some white guy on square a7 or c7 to the position bSh6 and wPh5, WinChloe allows Sxa7/c7.

        According to this short analysis, I see nothing wrong in the way WinChloe is dealing with such a mix of conditions.

        • Kostas Prentos says:

          Nicolas, I agree that it is more logical to have the rebirth square of bSh6 on b8, but WinChloe’s treatment of the Face-to-Face condition, when it comes to a pawn on the first/last rank is outright wrong. See the official definition of FTF (from the Batumi Bulletin), which mentions how a pawn on the last rank moves. For convenience, I quote it here:
          “Face-to-Face: When a white piece is just one rank below a black piece on the same file, they exchange their roles. P on the first rank cannot move unless it is given a power to move by FTF. Any piece can make an en passant capture when it is given a power to move as P by FTF.”

          However, if this is the case, and the rebirth square is that of a Knight, then the example I gave for twin d is a cook, whether null moves are allowed or not: 1.h1=Q a3 (a4-a2-a3) 2.Qa5 (Qh1-d8-a5) a4= (resulting in an immobile bQd8, moving like a pawn on the last rank).

          The only way for this problem to be sound, is if the rebirth square of the piece (e.g., bSh6) is that of a pawn (h7) and null moves are forbidden. But both of these requirements seem rather illogical to me.

          • Nicolas Dupontdupont says:

            “P on the first rank cannot move […]”.

            Ok this is a FTF rule, but what does it mean really? To my mind the bSh6 transferred to b8 is NOT a bP on its first rank – it is still a Knight, although moving like a bP…

            So I think the above rule is rather imprecise – I would have preferred (but is it really the author’s intention?) something like “Any piece on its first rank moving like a Pawn is immobile unless […]”.

  3. Karol Mlynka says:

    Excuse me please, but I think that by Face to Face white and black pieces exchange their walk – now bSh6 = bPh6 and wPh5 = wSh5, which is stalemate or not?

  4. Kostas Prentos says:

    This unusual combination of fairy conditions results in 4 possible variations of rebirth and movement:

    i) bSh6 is reborn like a pawn (on h7) and moves like a pawn (3.h7-h6)
    ii) bSh6 is reborn like a pawn (on h7) and moves like a Knight (3.Sh7-~)
    iii) bSh6 is reborn like a Knight (on b8) and moves like a pawn (3.b8-b7)
    iv) bSh6 is reborn like a Knight (on b8) and moves like a Knight (3.Sb8-~)

    If the movement of the reborn piece is according to the Knight (variations ii or iv), then it is obvious that the Knight can move and there is no stalemate.

    If the rebirth square is h7, and the Knight moves like a pawn (option i), then 3.h7-h6 seems perfectly legal to me, despite being a null move.

    The only case that looks like stalemate is the option iii: The Knight h6 is reborn on b8 and then must move like a pawn. According to the FTF definition, a pawn on the first (or eighth) rank cannot move, unless it is given the power to move by a FTF effect (e.g., by a white piece standing on b7). Therefore, this position is stalemate, because 3.b8-b7 is illegal per FTF rules.

    Now, if you ask me which of these 4 possibilities is the “right” one, I will have difficulty giving my opinion. Without much thought, I think that the rebirth square of bSh6 must be that of the Knight (b8, per Phantom Circe rules) and then, it must be forced to move like a pawn 3.b8-b7, which is illegal per Face-to-Face rules. If this is the correct interpretation, then Popeye is right (and WinChloe is wrong) and the problem is sound. Any other thoughts?

    • Kostas Prentos says:

      Looking at the other twins, it seems that d) is cooked:

      1.h1=B a4 (null move, a4-a2-a4) 2.Ba6 (Bh1-c8-a6) a5=

      Unless, a null move is not allowed at all (Kjell’s interpretation) and the option (i) is the reason Popeye finds the problem sound. If someone has a more detailed definition of Phantom Circe that disallows null moves, please provide it to us. In my confused mind, this interpretation does not seem reasonable, but if null moves are strictly forbidden in Phantom Circe, I will accept it.

  5. Kostas Prentos says:

    Another interesting fact about twin (d) that reinforces Kjell’s no-null-move explanation is the third promotion 1.h1=Q a3 (a4-a2-a3) 2.Qa5 (Qh1-d8-a5) a4 is not stalemate, according to option (i) because now 3.Qa6 is possible (3.Qa5-a7-a6). Whether null moves are allowed or not is a different story.

    My apologies for the consecutive comments, but I kept discovering new things after posting my previous comments. I hope this is the last one for now.

  6. Mike Neumeier says:

    My opinion. I would hope Popeye has it right. Allowing them just seems too weird. Sounds like WinChloe’s promotion routine is not doing all the maintenance when the Phantom condition is present. It incorrectly retains the original movement powers of the diagram pawns. It goes so far as to reestablish the new birth square correctly, then exits.

  7. Kjell Widlert says:

    “Kjell’s interpretation” of null moves?

    That Py does not allow null moves is not my interpratation, it is something I know for a fact (unless I am greatly mistaken). So I Believe Py always compares the “before the move” and “after the move” positions, and if they are identical, then the move is illegal – in all fairy forms.

    Personally, I don’t agree with this definition. I think any move that is in conformance with the current rules of play should normally be allowed, whether the resulting position looks the same as the previous position or not – unless the rules of a particular fairy form expressly forbids such moves. Mathematically speaking: an identity transformation is also a legal transformation!

    • Mike Neumeier says:

      It is natural to expect that a “move” entails actual movement. How does Black know it’s his turn to move if White just played a null phantom move?

      • Mike Neumeier says:

        I wonder who the inventor of Phantom is, and what his/her opinion is on null moves.

      • Kostas Prentos says:

        Mike, a null move does not necessarily imply that the side to move passes its turn to move. For example, the Rose travels across the full board to finish its move on the square from which it started (I am curious how Popeye treats a Rose’s null move). In Phantom chess, the null move can be analyzed in two steps: First the moving piece is reborn on its Circe rebirth square and then it moves back to the square it was standing before the move. It means an actual movement and a rebirth, not just doing nothing.

        Forbidding null moves (in Phantom chess, at least) seems highly illogical to me. Suppose we have a bQ standing on e8. She can move anywhere on Queen lines from e8, and anywhere on Queen lines from the rebirth square d8, except back to e8! It just does not seem right.

        • Nicolas Dupontdupont says:

          “I am curious how Popeye treats a Rose’s null move”.

          It doesn’t allow it – seems to be a general principle, as pointed out by Kjell. For an illustration, you might have a look to JF-problem 640 of mine.

          • Kostas Prentos says:

            Thank you, Nicolas. I guess I was too lazy to do my own research.

            In my mind, null moves by a Rose are generally accepted; that’s why I used this fairy piece as an example. I wonder now if this is really the case, or if some problemists think that null moves by the Rose are not allowed. If so, is it because Popeye does not allow null moves? In other words, how much power do chess solving programs have in forming a set of rules (right or wrong) that people follow without second-guessing them?

            • Nicolas Dupontdupont says:

              WinChloe contains a themes list, among them standing the null move. A quick look at the related problems list shows an amount of them – with a Rose but also with a Ubi-Ubi, a Spider etc. Conditions such as cylinder board for example also show this trick.

              So it seems, at least, that the lack of null moves from Popeye didn’t prevent composers to use this opportunity…

              • Kjell Widlert says:

                Yes, null moves with roses are not unusual, several composers (including myself!) have used them. But you have to use some other program than Py to test them.

                I have also seen several examples of null moves on a cylinder board (again, including myself).

                Kostas states very well the main reason why null moves shouldn’t generally be forbidden. It doesn’t make sense to me (either) that one of all the normal moves, according to the power of movement of a particular piece, should be illegal just because the resulting position happens to look the same as the previous position.

  8. Mike Neumeier says:

    In all these discussions on rules, I forgot to say that I like this problem. Surely hope it survives as sound!

  9. Nicolas Dupontdupont says:

    In a), why the sequence 1.h1=Q/R h4 2.Q/R×h4(h2) h3=
    doesn’t lead to another solution? I mean with the hypothesis that Sh6-b7 is illegal as in Karol’s solution.

    • Kostas Prentos says:


      Yes, this is also a cook, provided that the rebirth square is that of the Queen or Rook, and not h7.

      Pawns on the first/last rank: It is very unpleasant that there are such differencies from one fairy condition to another. A pawn on the last rank can be immobile, move one, two or even three squares forward. It is almost impossible to keep up with each particular case.

      You wrote: “To my mind the bSh6 transferred to b8 is NOT a bP on its first rank – it is still a Knight, although moving like a bP…”

      Yes, I agree and disagree, at the same time. The bSh6 transferred to b8 is still a Knight. But it is a Knight moving like a pawn, therefore, not being able to move at all (per FTF rules). If it could move, what moves would it be allowed to play? One, two, or three squares forward? This is not specified by the rules. What is specified, is that the pawn (or a piece that has the moving properties of a pawn, is the logical consequence) cannot move at all.

      Finally, we should all keep in mind that any fairy condition may be fully defined and perfectly clear, but when two or more fairy conditions are applied together, the result may be comical. It might be a good idea for a composer to specify in advance which special cases apply each time, and how conflicts of fairy conditions are resolved. But then, adding more text to an already heavy stipulation does not seem very economical, does it?

      • Nicolas Dupontdupont says:

        There are so many comments I would like to share Kostas, that it needs several messages. For the moment I still concentrate on Karol’s problem.

        If we add to it Winchloe’s condition “Un Pion est impuissant sur sa première rangée” (which is part of the FTF definition, as you pointed out), then plenty of “cooks” appear, seemingly as valid as author’s own solutions. For example 1.h1=B é7 2.Bé4 é3= in the b) case.

        So it now looks complicated (at least to my eyes) to declare sound the problem, whatever the rules interpretations we are adopting…

        • Kostas Prentos says:

          I quote from one of my previous comments:
          “The only way for this problem to be sound, is if the rebirth square of the piece (e.g., bSh6) is that of a pawn (h7) and null moves are forbidden.”

          I could not find a cook, in any of the twins, that would work with these restrictions, but I may be missing something.

          • Nicolas Dupontdupont says:

            Oh yes this is my mistake, thanks!

            To summary, in the b) solution the final position is bBe6 wPe5, which is a stalemate – black on move – provided that Be6=P->e7-e6 is forbidden. In Winchloe’s “solution” quoted above the final position is bBe4 wPe3, which is not a stalemate – black on move – provided that Be4=P->e7-e6 is allowed.

    • Mike Neumeier says:

      Not a cook due to the available phantom moves by black, for example Qh6 (as a pawn, disembodies the h4Q and plays h7-h6). Qh5 possible, too (h7-h5). Qh4 not possible (no change on the board occurs as a result of the perceived phantom move (null move). The Py Rose handling and the way Py handles this problem are at least consistent, no nulls.

  10. Mike Neumeier says:

    LOL. Not Qh4, because the phantom pawn cannot move three squares to h4 from h7.

  11. Kjell Widlert says:

    We had a somewhat similar case last year in Bala’s no 617 with BTB and Madrasi: in the constellation bRh4 wBh5 bBf5, the wB moves like a rook because of BTB, and therefore observes bBf5, but still really is a bishop, and therefore paralyses bBf5 which is also a bishop. A reasonable way to look at a complicated situation, I think.

    By the same reasoning, the bSh6 moves like a P (because of wPh5 and BTB) but still is a bS. So it seems right to have b8 as the rebirth square for this piece. But once it moves from b8 (by Ghost Chess), it seems natural to me to have it move like a bS: it doesn’t sit back-to-back with a wP on b8!

    So Kostas’ variation iv) seems to me most logical.
    But neither Py nor WinChloe agrees with me 🙁

    • Nicolas Dupontdupont says:

      Concerning the end of your second paragraph (I already mentioned that I fully agree with the first one and the beginning of the second one), this is indeed an interesting (and new in the discussion!) viewpoint, but I’m not sure this could be the right way…

      Consider the position wBa2 bPb3 bSb4 in T&M and FTF. According to both Popeye and WinChloe, the capturing (full) move is Ba2xb3(Bb2). It means that the wBb3 is not affected by the face-to-face with bSb4 – it moves like a Pawn and not like a Knight. The reason (on which I agree) is probably that Bxb3 is not a move in T&M, and hence the wB was in fact never face-to-face with the bS.

      Now reverse the process in the problem situation after the a) solution. The bSh6 is face-to-face with wPh5 and should (to my mind) be considered to still be face-to-face after the rebirth Sh6->b8, as this latter is not a move in Ghost and hence doesn’t have the power to “disconnect” bS and wP (for the same reason that Bxb3 doesn’t have the power to “connect” wB and bS).

  12. Mike Neumeier says:

    Allowing null moves opens up other questions. One is the 50-move draw rule (50 moves without a pawn move or capture). Does a null pawn move in a phantom study mean the count should return to zero?

  13. Kostas Prentos says:

    A null move by a pawn is still a pawn move and restarts the count. A move (null or not) by a piece that moves like a pawn is probably not a pawn move, but a move by a pawn that moves like a piece should probably restart the count. This is just my impression. However, it is very unlikely that we will ever see this in a problem.

    • Dmitri TurevskiDmitri Turevski says:

      Interesting. I thought “pawn move or capture” in the 50-move rule was just an explicit way of saying “irreversible move” – i.e. move after making which you can never return to the position before the move.
      In this sense, pawn null-move should not necessarily restart the counter.

      • Kostas Prentos says:

        Dmitri, maybe you are right, I don’t know. But then, any pawn move in Phantom chess has the potential of resulting in the same position before the move, even if not right away with a null move. For example, Ph4-h7-h5, followed by h5-h4 on the next move repeats the position of the pawn. Even if the pawn moves further forward to h3 or h2, it can always return to h4 later on. However, as I said before, I do not see this dilemma to have practical interest, unless an extreme adventurer retro composer decides to explore the possibilities of Phantom chess and the 50-move rule.

        • Nikola Predrag says:

          Dmitri has given the genuine motivation for the existence of the rule.
          Unfortunately, the formulation of the rule is not adequate and should have be re-formulated for the fairies.

          It should be re-defined even for OTB games by including the restart of counting also when the castling-right is lost, as it is for the 3-fold repetition.

          • Kostas Prentos says:

            About castling and the 50-move rule: There was a time when castling would restart the count. Later, the FIDE rules changed, excluding the castling part, probably because castling was very unlikely to have any connection to the 50 moves anyway, as most of the times, this rule was used for the end of the game when castling was already done or lost forever.
            However, this small change had an influence to retro compositions. There are some classical retro problems, mainly by Nikita Plaksin, that included castling as one of the requirements for a 50-move draw. It was not the first time that a change in FIDE rules would make a problem unsound.

            If we add a fairy condition in the equation of the 50-move rule, things can get very complicated, like in the case of Phantom chess. What makes sense to me is to use the letter of the rule, otherwise we open a can of worms.

  14. Nikola Predrag says:

    Popeye finds the solution for the following:
    White Kc6 Pc4 Qa2 Pc2
    Black Ka4 Bb4 Qa3 Bb3
    Stipulation H#2
    Condition AntiCirce Circe

    1.Qa3*a2[bQa2->d8][+wQd1] Qd1*d8[wQd8->d1][+bQd8] 2.Qd8-a5 c2*b3[wPb3->b2][+bBc8] #
    So, it’s not just about the “identical positions” before and after the move in all cases.

    • Mike Neumeier says:

      Interesting, white does take a pass, in effect a null move, on his first. I suppose one could debate the black rebirth on d8 since the capture occurred there. It’s as though Anti-Circe is resolved first (wQ back to d1), negating Circe permanent capture (occupied d8).

      • Mike Neumeier says:

        On that, it would be interesting to see if switching the two conditions’ order on the Condition row finds the same solution.

    • Kostas Prentos says:

      Maybe Popeye accepts the null move here, because it involves a capture. To what level does that make a difference? If a capture is considered a change, then what about a rebirth?

Leave a Reply

Your email address will not be published. Required fields are marked *

You can add images to your comment by clicking here.