Julia's Fairies

No.636 (GF)

Geoff Foster (Australia)


Original Problems, Julia’s Fairies – 2014 (III): September – December

→Previous→Next ; →List 2014(III)

Please send your original fairy problems to: julia@juliasfairies.com

No.636 by Geoff Foster – Valladao theme with Anti-Take&Make condition! (JV)


Anti Take & Make: Every capture (“take”) must be complemented by a further step (“make” which is not a capture) by the captured piece (Kings excluded), which must move from its square of vanish (according to the wishes of the capturing side in case of options). The capture is forbidden if the captured unit has no possible moves. Promotions at the end of the “make” element are normal.

No.636 Geoff Foster

original – 31.10.2014
Dedicated to Sébastien Luce

Solution: (click to show/hide)

white Kh6 pb2 black Ke8 Ra1a8 Pa4

h#4                                       (2+4)

Notify of

Newest Most Voted
Inline Feedbacks
View all comments
S. K. Balasubramanian
S. K. Balasubramanian
November 1, 2014 12:44

Yes, definitely there is a controversy that whether the wP is captured on b3 or b4. I also would have assumed that the wP is captured on b4 as it has already moved to b4.

Nikola Predrag
Nikola Predrag
November 1, 2014 15:02

Which rules allow capture a4xb4?
Neither T&M nor standard rules mention it but standard rules allow a4xb3 and a4xb3ep.
bPa4 can capture wP on b3 after both b2-b3 and b2-b4.
That’s what the rules say.
If someone applies different rules, it should be mentioned in a definition of a fairy condition.

November 1, 2014 15:58

Geoff is giving his own notation for the solution, which is not the one used by WinChloe. Rules are still the same: b2 reaches b4 and is captured on b3.

If we follow your own interpretation, Nikola, between b2-b4 and a4xb3, bPa4 claimed for a pull-back b4-b3. Thus in the ep-procedure wPb2 is perfectly orthodox while bPa4 is a fairy piece, able to pull-back on b3 an opponent Pawn which just crossed this square. Convincing…

Dominique Forlot
Dominique Forlot
November 1, 2014 20:02

the notation above is a presentation of the solution for the use of “Py2Web” ( magical illustration used on this site )

[the Animated diagrams use Py2Web by Dmitri Turevski and WinChloe’s graphics.]

The solution need to be formated as follows on this Py2Web’s page:

1.0-0-0 b2-b4 2.a4*b3 ep.[+wPb5] b5-b6 3.Ra1-a7 b6*a7[+bRc7] 4.Rd8-d7 a7-a8=Q #

winchloé accepts this problem and gives the same solution :

1.0-0-0 b4 2.a×b3 e.p.(b5) b6 3.Ra7 b×a7(Rç7) 4.Rdd7 a8=Q‡

but Popeye don’t find no solution!

that makes two approches of this fairy, strangely.! Just like we have a circe of “type cheylan” or of “type calvet” it seems that we have one “pep type I” and one “pep Type II” in our mind

Nikola Predrag
Nikola Predrag
November 1, 2014 21:20

Literally, bPa4 is an completely orthodox piece because it is defined by orthodox rules.
The same is about wPb2 and the ep. capture, that all belongs to the orthodox rules.
However, we notice that the ep. capture requires a special paragraph to define it, because the general definition of the pawn-movement is specifically altered.

A logic of that alteration is not fully and clearly described in the rules, so we may speculate about it but of course, within the limits of general rules.
bP captures wP on b3 and the only space for speculations is about how wP has come to b3 after b2-b4.

The very first alteration of the general logic is the exclusive ability of a pawn in its first move: it may “ALTERNATIVELY ADVANCE two squares along the same file”.

That formulation pretty strongly suggests that “one-step move” is the default principle and that “double-step advance” is the alternative to one-step move.

So, b2-b4 is a “fairy” move, if you wish. wPb2 may choose the alternative but passing near bPa4 to avoid its attack on b3 would be an unfair advantage for wP.

Therefore, bPa4 is also allowed to choose between the default play and the alternatively “fairy play”, of wP.
But that all begins with the “fairy” double-step advance of wP.

white Pa2, black ep-Ra8, 1.a4 ep-Rxa3ep. might be possible. Although the Pawn’s” alter-ego” on a4 apparently shields its own default-ego on a3, the ep-Rook has the ultimate choice 🙂

November 1, 2014 22:00

Dominique, the relevant question is whether b2 reached b4 after the ep-sequence. And both Popeye and WinChloe answer “yes” (see e.g. the Parrain Circe framework). Maybe Popeye doesn’t find Geoff’s solution because of a bug?

The fact that wPb2 is captured on b3 or on b4 doesn’t matter nor in this problem, neither in Sébastien’s. In fact I don’ know of any practical situation where it could be relevant (I mean a problem which is correct or not according to the choice of the square where the ep-capture occurs).

Finally, if we follow the convention that b2 is captured on b4 (as Geoff and S.K. suggest), the difficulty doesn’t come from some potential inconstancy (at least I can’t find some), but from the fact that there is no more “en-passant” capture in this setting!

Nikola, yes indeed Dominique and I already remarked that a2-a4 epRa8xa3 must be considered legal in “ep-chess”. But it would also be the case for each white piece standing on a2 and able to move on a4 via 2 leaps – e.g. Rook or Queen. That’s why I don’t like to consider Pa2-a4 as fairy, because it would imply that Qa2-a4 is also a fairy move! (in the more general setting of ep-chess I mean).

Dominique Forlot
Dominique Forlot
November 1, 2014 22:46

Nicolas, I agree with this point of view, for me if I chose 1.a2-a4 rather than 1.a2-a3, the pawn reach a4 of course!( it’s factual!)

I am persuaded that our vision of the ep-power is an logic extension of the pep’s rules of pep and the castle’s constraints.

I wanted just to underline the difference of processing of a same problem by two programs taken for reference.

All pieces seem to possess an ep-power, they have to let us see how they use this talent.

Geoff Foster
Geoff Foster
November 1, 2014 23:07

It seems there is a bug in Popeye whereby en passant is not allowed with anti-Take&Make. The best way to show this is by using the “h~” (help-anymove) stipulation. Consider a position with just a wPg5 and bPh7, with Black to play. The Popeye code is:

stipulation h~1
pieces white pg5 black ph7

Popeye gives the following 4 possibilities, as expected:

1.h7-h5 g5-g6
1.h7-h5 g5*h6 ep.
1.h7-h6 g5-g6
1.h7-h6 g5*h6

However now add “condition CirceTake&Make” (this is what Popeye calls anti-Take&Make), and there is no en passant possibility:

1.h7-h5 g5-g6
1.h7-h6 g5-g6
1.h7-h6 g5*h6[+bPh5]

Nikola Predrag
Nikola Predrag
November 2, 2014 00:13

Yes Nicolas, I tried to express it shortly and it was not clear enough.
Literally, a2-a4 is not a fairy move, it is allowed by orthodox rules as an orthodox alternative to the default one-step move.
The choice of “alternative perception” is allowed to bP.
But this is not explicitly stated by the rules.
We can only SPECULATE that the formulation “as though…” might be interpreted as an “alternative perception” by bP.

So, this SPECULATION is a “fairy-stuff” because it is not defined by orthodox rules.
Qa2-a4 is not a fairy move but an EP-piece has a fairy power of “alternative perception”, i.e. to perceive each single picture of the whole series of moving-pictures.
Other pieces see that series as a motion in a movie.

However, some piece completes its move with all consequences like capturing on g7 in your example 1.Ba1xg7 axBb2ep..
If bPa3 can choose the moment when wB was crossing b2, bQg7 was not captured yet, on that “single picture”.

A simpler fairy definition of EP-pieces might allow this play:
1.Bxg7 axb2ep.(bQg7)#
This would be a rather logical extension of the orthodox pawn’s ep.-ability, to the EP-pieces.
The main fairy element here would be the particular speculation about how to interprete “as though…”.

Disappearance of bQ, after 1.Bxg7 axb2ep. requires additional fairy logic, something like:
bPa3 can’t fully react before wB completes its move and can’t undo the consequences.
But bP can catch the wB’s “tail” (or “trace”) on b2.
wB can stretch to g7 as “tied” by a “rubber band/elastic force”.
Then it is pulled back to b2, but only if EP-Pawn will indeed capture it.
Thus, wB would be captured on b2.

The idea that bEP-Pawn captures on b2 but that wB is captured on g7, would require quite a tricky fairy logic.
For instance, bEP-Pa3 catches wB while crossing b2, leaves its own elastic pinned to b2 and is pulled by wB to g7. Then wB captures bQ, bEP-P captures wB and is pulled back to b2.

What about wBc1,bEP-Pa3; 1.Bxa3 bEP-Pxb2ep.?
bEP-Pa3 catches wB while crossing b2 and is pulled by wB to a3. How that “clash” on a3 ends?
If wB arrives first to a3 bEP-Pa3 was never captured and it should be written as 1.Ba3 bEP-Pxb2ep.

Dominique Forlot
Dominique Forlot
November 2, 2014 01:00

it’s a joke! 🙂 that is not the mechanism of ep-chess, of course;
Otherwise, by this way, after 1.g2-g4 if the pawn blanc h4 choose to bring back the pawn on g3, you can consider now that the g pawn don’t have finished its move, and to finish it by g3xh4 capturing the black during its own move!

You would obtain the magic move 1.g2xh4 !!

i love this move but its another fairy 🙂

November 2, 2014 01:06

Roughly speaking, I see an EP-piece as a policeman along a road. If you drive too fast while crossing the point where he is standing, he may whistles you and you have to stop, then go-back to him to be “captured”. But now if you crash into the policeman and “kill” him, he can no more whistle you…

The chess-translation of this metaphoric little story is clear: it is legal to capture an EP-piece – even with a cross of its observation range (and even if it is not legal to kill a policeman!), and end of the story, the EP-piece lost all his power. Note that it is coherent with the orthodox framework – wPa5 is a kind of semi EP-piece and may be captured, even by a piece crossing square b6.

Btw, another interesting position is bPb3 and wGd1 where the Pawn has (full) ep-power. Then after Gd1-a4 the ep-capture Pb3xGc2 should be legal… Indeed a grasshopper is moving leap after leap before it meets the hurdle (hence it crossed square c2 – I checked this rule).

Nikola Predrag
Nikola Predrag
November 2, 2014 01:26

It seems that Popeye at least tries to obey the standard ep.-rules and admits that bP was captured on h6.

But Popeye also thinks that bP is not yet “removed” from h5, so the Make part, bPh6-h5 is impossible because the same bP still occupies h5.

it is not factual that wP has reached a4 after a2-a4, before Black admits that.
If you never read the rules and try to understand them only by watching the players, you will see wP on a4.
But the rules don’t say that wP may MOVE from a2 to a4.
If wP MOVES to a4, with the full meaning of MOVING onto some square, as it is defined for other pieces, bP can’t capture it.
ADVANCE to a4 still can be treated “as though wP MOVED to a3”. Standard rules don’t allow such treatment of the MOVES by other pieces. Only a2-a3 is treated as MOVING completely to a particular square.

Moving the material pieces on the material board is only a projection of the true chess-play into the material world.
Literally applied rules would perhaps require that after a2-a4, the material wP is placed partially on both a3 and a4 on the material board.

Anyway, ep. capture requires that the whole Pawn occupies a3 before it could be captured, as though it played a2-a3!
The projection into the material world should actually show the “clash” of two opposing pawns. Only that “clash” may result with a capture.
That is the most profound nature of standard chess, and claiming that the projection in material world “proves” that wP was removed from 4th rank only reveals the distortions in the projection.

Georgy Evseev
Georgy Evseev
November 2, 2014 10:11

“Capture for dummies”

Generally a move consists of moving a single piece M from one square to another. Sometimes another piece is affected, as a rule of an opposite color. In this case this move is called a capture and an affected piece – a captured piece C.

In traditional chess the effect of capture is a removal of piece C from the board. In this case it does not matter if the act of capture is tied to specific square. But in many kinds of fairy chess (Circe, Anticirce, Take&Make families) there is some effect which depends on the “square of capture”.

In general, two squares may be considered as “square of capture”: the square to which a move of piece M is made (“m”) and the square on which piece C stands before the capture (“c”). In simple cases squares “m” and “c” are really the same square which may be called “the square of capture” unambiguously. In more difficult cases (en passant, Locust, marine, Dynamo, etc.) the squares “m” and “c” are different.

As there are no “correct” or “wrong” fairy conditions – only different ones if they are not contradictory inside, we, in fact, may define different conditions: Circe M and Circe C, Anticirce M and Anticirce C, etc. The added letter shows if we consider square “m” or “c” as “square of capture”.

Traditionally though, the conditions affecting captured piece (Circe, Anti Take&Make) use “c” square; while conditions affecting moving piece (Anticirce, Take&Make) use “m” square.

Nikola Predrag
Nikola Predrag
November 2, 2014 13:56
Reply to  Georgy Evseev

bPh4 can move to g3 only if g3 is occupied by some opponent’s piece. So wP must already occupy g3 before the capture, otherwise bP can’t move to g3.
wP is captured “as though it was occupying g3” before Black started to move.
All consequences of Black’s ep-move should be “as though wP was occupying g3” before it was captured.

Nikola Predrag
Nikola Predrag
November 2, 2014 23:42

the logic of your police-story seems convincing enough. Hopefully it can be defined with sufficient exactness, and that should be enough for a fairy condition.

Interaction of the capturing ep-piece and the captured piece, is based on an “administrative force” to “pull back” a piece. It is more abstract than “elastic force” but the point is the same and the abstraction is actually more suitable for definitions.

What about the priority of effects?
ep-piece attacks square X, opponent’s piece A crosses X and captures piece B which should be reborn on X.
Is B reborn first and ep. capture is not possible or ep. capture happens first and the rebirth is impossible?

November 3, 2014 18:24

Yes Nikola, we have to precisely define what’s going on while adding the EP-condition to another condition X. The EP-X-procedure goes as follows:

1) An X-legal move is played
2) Pull-back (this is not a move but a kind of rebirth, hence no legality is needed)
3) An X-legal capture is played

It might happen that the pull-back forbids the capture to come – in that case it is not allowed to ask for this pull-back, as 2) and 3) are parts of the same move. Indeed we are obliged to link 2) and 3) because this is the same side which may previously ask for a pull-back, and then capture the pulled-back unit. The following example will make fully clear this stuff, I hope:

W: Ka1 Pa2a4 where Pa2 has ep-power
B: Bc2 Ga5

The ep-sequence 1.Bxa4 1…(Ba4-b3) Pa2xBb3 is legal, although the pull-back move (Ba4-b3) is a self-check…

Now concerning your Circe example, point 1) shows that the full Circe move has to be played, including the rebirth. If this rebirth occurs on a square which forbids either point 2) or then point 3), then no ep-capture can occur. Example:

W: Ba1 Pb4 where Bb1 has ep-power
B: Bd2

After the Circe move Bd2xb4(b2), there is no possible ep-capture, although Bd2 crossed square c3 which is attacked by Ba1.

The “reverse” case (an observed square appears only after the rebirth) addresses a potentially controversial setting: Replace the white Bishop by a white Grasshopper in the above position. Is Bd2xb4(b2) epGxBc3(Bf8) legal ?!

It is not true that Bd2 crossed a square observed by Ga1, but it is true that Bd2 played a Circe move such that, as a consequence, the crossed square c3 becomes observed by Ga1! Hence the above ep-sequence should be considered legal, at least to my mind. Do you agree on this funny setup?

Georgy Evseev
Georgy Evseev
November 3, 2014 19:36


I do not like a pullback idea. Somehow, Nikola interprets it as the only possible one.

So, here is another possible interpretation)).

During a long move the body of a piece moves faster than its soul and while the body arrives to destination, the soul still lingers along the the path. This soul, while immaterial and transparent, may be captured as a normal piece by an opponent piece with ep-power (you may also name some pieces ep-susceptible, defining that only such pieces may be ep-captured). Of course, the soulless piece immediately dies and goes where all dead pieces go))).

This interpretation positively solves most difficulties, including Begley/Caillaud double capture or your Grasshopper example.

November 3, 2014 21:16

Dear Geoff,
no controversy on my side !
Thanks a lot for the dedication !!

Nikola Predrag
Nikola Predrag
November 3, 2014 22:42

Nicolas, it looks consistent, so I have no reason to disagree.
Georgy, I have no reason to disagree with your logic for it also looks consistent.

However, it would be good to predict at least some possible complications with additional fairy elements.
Georgy’s explanation abandons the very clear concept of capture in standard chess, a clash of the opposing pieces on the same square.
Without that concept, it might be hard to predict the troubles.

Sebastian, the problem is nice and Georgy’s change of the standard concept of a capture would make it correct. Of course, such a change should be mentioned in the definition of T&M.

However, even with that change it’s not a true Valladao!
Valladao is interesting exactly because normally 1.b4 axb3ep. gives the same position as 1.b3 axb3.
The author must invent something to make 1.b3 illegal and in good problems also to invent something to make bP necessary on b3.

Solution of Geoff’s problem wouldn’t change with bPa5 instead of bPa4, that makes Valladao essentially unconvincing.

November 3, 2014 23:46

Yes, I think it is worth to keep a standard capture at step 3) of the generalized ep-procedure. This way the only fairy element is step 2) – the “pull-back” or “death and rebirth” or whatever it is called, which appears just before the ep-capture.

Btw, Dominique and I got the idea of looking at the anti-form of EP-chess – A move that would lead to a possible generalized ep-capture is illegal. In orthodox chess this already happens with e.g. the white King crossing square f1 during short castling. That’s why we have in mind to call OO-chess this new possibility.

It is rather interesting to see that the 2 special kind of pieces in orthodox chess are linked by a fairy condition and its anti-form: Pa5 is a “semi EP-piece” while Ke1 is a “semi OO-piece”.

Kostas Prentos
Kostas Prentos
November 4, 2014 06:29

Although my opinion about the soundness of the recent entries featuring the Valladao task remains the same (I consider them sound), I find Nikola’s conclusion of his last comment (from November 3, 22:42) quite convincing. The en passant capture loses much of its effectiveness when a “convenient” fairy condition is used. For example, the value of the e.p. capture in Madrasi is minimal. In Circe variants, it is a little better, but the fact that instead of a black pawn on h4/a4, a pawn on h5/a5 would do the same job with a normal capture on g4, makes the whole task more trivial. In a similar way, Excelsiors are less interesting and rather ordinary in long helpmates. So, regardless of the achievement of a formal theme (like Valladao), other criteria need to be applied, in order to evaluate the artistic merit of a problem.

Nikola Predrag
Nikola Predrag
November 4, 2014 10:48
Reply to  Kostas Prentos

Correction to my previous post, Geoff’s problem is nice and correct but the fairy definition isn’t.
At least for solving. Every solver, including me, should be able to understand the condition on the basis of default rules, before solving, except perhaps in case of a joke problem.

“Advance two squares along the same file” has two possible results, “as though” moving one-step and “as though” moving double-step.
Next, immediately following opponent’s move, determines the “game-reality”.
“Material-reality” is only a projection of “game-reality” into material world. Seeing the double-step on a material board, before the opponent’s move, is a distorted projection in a material-reality.

“Move to” and “advance towards” some square, is perhaps a terminology to be considered by Nicolas and Dominique?

Dominique Forlot
Dominique Forlot
November 4, 2014 17:03

Here is my metaphor of the “prise en passant”
(just my subjective point of view of course)

Let us consider this position
B : Fa1
N :UUa8 ( Ubi-Ubi)

And, now, Let us consider the move 1. Fa1-h8!

The bishop crossed six squares before occupying h8, all under the supervision of the UUa8.
This last piece( UbiUbi) is so powerful that it covers here all the chessboard at the same time.
Nicolas defined it to me as ” a quantum piece* ”

* i.e. during its move, it can be everywhere at the same time, the raison why we call it “Ubiquity”

In the same way, when the bishop arrives on h8 we can consider “the set of squares visiting by the bishop” as a subset of the chessboard ( b2,c3,d4,e5,f6,g7,h8) where the bishop exist on an “Ubiquity mode” as long as black has not finished its next move.

Now, if the UbiUbi choose to play on any square of this sub-set, it meet the bishop and naturally capture it. ( orthodox on h8 but “en passant” every whre in this sub-set… and the bishop disappear from the chessboard )

November 4, 2014 18:27

Here is a possible definition of EP-piece which seems to follow, in a clear-cut fashion, the needed criteria:

“As a preliminary part of its move, an EP-piece X may take the opponent piece Y that just moved, and then put it on any square just crossed by Y, providing X may capture Y on this square, thereby completing its move.”

As a consequence, the orthodox ep-procedure (with wPa2 and bPb4) could be detailed as follows:

Step one: 1.a2-a4 (normal move, in particular a4 is really reached)
Step two: Pa4->Pa3 (preliminary part of the next black move – the white Pawn is taken from a4 and is putted on a3, just crossed by Pa2)
Step three: 1… b4xPa3 (completion of the black move, via a normal capture occurring on a3)

Nikola Predrag
Nikola Predrag
November 4, 2014 21:49

Capture is THE most essential concept in chess. It’s highly desirable to define the capture exactly and clearly in one paragraph. And the basis on which everything else stands,
should not be altered in other and “special” paragraphs.

The unstable foundations is not foundation, it’s neither convincing nor worth serious attention.

The Begley/Caillaud speculation requires a new definition of capture, based on new definition of square-occupation by own/opponent’s piece, a new definition of the pieces’ movement and yet to make it fully convincing, some natural logic of the whole system would be desirable.

Nicolas, so far you seem to keep the standard basic concepts so, it’s easy to follow the specificity of the innovation.
Just to check it, what would the innovation say about Begley/Caillaud case?

November 4, 2014 23:43

It is based on the following scheme:

W: Ra8 Pg5
B: Rh8 Ph7
Parrain Circe

1.Rxh8 h5(Rh6) everything normal for now.

Authors are considering that the ep-capture is allowed even if square h6 is occupied by the bR. Despite the deep respect I feel for Kevin & Michel, I think it is not acceptable, even near from a joke move…

I looked at WinChloe’s analysis of the same position. After the normal move 1, the next one is 2.gxRh6 where h5 is still alive – thus the program denies the ep-capture. It seems to me this is the correct way to handle this particular situation, and it perfectly fits my above theoretical description of the ep-capture (equivalently how an EP-piece is working).

Indeed, as h6 is occupied, the step 2 – transferring Ph5 to h6, is impossible (if we keep as a definite axiom that a square may contain at most one unit). As a consequence the Begley/Caillaud ep-capture is illegal while WinChloe’s move is correct.

Juraj Lörinc
Juraj Lörinc
November 5, 2014 09:52
Reply to  dupont

Authors of Begley/Caillaud have found “hole” in existing rules (13 years ago), successfully exploited it, the judge of primary tourney was impressed (I know that, because it was Juraj Lorinc) and FIDE Album judges were impressed as well (only fraction of problems successfully falls through their sieve as a rule).

Therefore I find it wrong to mark such case as unacceptable, illegal or joke just because many years later the theoretical discussion might result in different consensus and newly defined terms.

I support the effort to define rules consistently (including fairy rules and their combinations) as I prefer to work with well defined terms. But I dare to reserve right to work with any interpretation that works. If the author of problem can show some particular effect successfully within his own interpretation, I am prepared to give him credit and praise even if anybody claims his work is unsound based on other interpretation of rules.

November 5, 2014 20:10
Reply to  Juraj Lörinc

Considering the Begley/Caillaud double capture as “near from a joke move” is just my own feeling, it tells nothing about the cleverness of the idea behind it, and the possibility of exploiting the holes of the rules with great inventiveness…

You are speaking of a “different consensus”, hence suggesting that there were a consensus in 2001 to consider legal this double capture. Frankly speaking it is difficult for me to imagine that each composer agreed at that time, and then were all disclaimed both by Popeye and WinChloe… Indeed, if those programs are now rejecting the double capture trick, there are necessarely some good reasons that can’t have been completely eluded in 2001.

Juraj Lörinc
Juraj Lörinc
November 5, 2014 20:42
Reply to  dupont

I am not suggesting there was any consensus in 2001. Over time just about 6 people (two authors maybe earlier, one StrateGems in 2002 and three FIDE Album judges even later) thought this interpretation of e.p. in Circe Parrain is okay. Nobody was looking for consensus. However, if there is any consensus as a result of long discussions now, this does not invalidate the previous interpretation(s) of the rule.

Nikola Predrag
Nikola Predrag
November 6, 2014 00:32
Reply to  Juraj Lörinc

Well, if there was indeed a hole in the standard rules 13 years ago, then the problem is correct, showing great imagination and creativity.
The most impressive imaginative creativity is creating within the limits of the rules.

However, in case that there was not a hole in the mentioned rules, then the authors have created the “new” rules. The logic of that “new” rules is not convincing, it’s actually repelling.

So, what was the hole in the rules 13 years ago?

Nikola Predrag
Nikola Predrag
November 4, 2014 23:54

Well, if the change of standard rules is not defined in Parrain Circe, the ep. capture on h6 is legal.
wPg5 may capture bP as though it has moved to h6, which includes the rebirth of bR on h7.
It could hardly be clearer.

November 5, 2014 00:54

If I understand correctly you would like, after the move h7-h5(Rh6), to transfer the whole couple (h5,Rh6) to (h6,Rh7) and then to capture g5xPh6. I’m afraid this would lead to intricate situations and anyway it doesn’t fit the Begley/Caillaud double capture.

I prefer to consider h7-h5(Rh6) as if it was a series of 2 distinct moves in a ser-h2 problem and, after such a series, it is clearly illegal for white to end with an ep-capture.

Nikola Predrag
Nikola Predrag
November 5, 2014 02:23
Reply to  dupont

Nicolas, I didn’t speak about my preferences, just about what the standard rules say.
wPg5 may capture bP as though it was moved to h6, in which case bR was reborn on h7.

The rule doesn’t say “as though bP occupies h6 for some mysterious reason, although it has really moved to h5”, which might cause ambiguity when Parrain Circe is added, because bR would really occupy h6 and bP hypothetically.

I can’t see that (without Parrain Circe) bP has really moved to h5 as long as the ep. capture possible, because as long as the reality may have a different outcome, “as though…”, bP can’t really occupy h5.
The rules do not suggest any possibility that wPg5 may move to h5, empty or occupied by an opponent’s piece.
Instead of personal preferences, we just have to apply the rules, wPg5 may REALLY move to h6 only if it is occupied by an opponent’s piece.
And that is the only reality allowed by the rules, so bP can’t be really on h5 if wP “may move” to h6. Or bP would have to be really on both h6 and h5 which is not allowed by the rules.
In the reality, bP is neither on h6 nor on h5. When White decides to play the particular move, one option becomes real.

I will not explain anymore what I see as the only possible result of the standard rules.
Instead, after so much time wasted, I expect to see any consistent interpretation of standard rules from all those who claim that bPh5 occupies h5 while it’s been captured on h6.
So far, it was “proved” by observing the otb players as though they play by some other rules.

November 5, 2014 04:16

Rules are sometimes incomplete and hence, while waiting for a general ep-theory that will apply without any ambiguity to (almost) each fairy condition, we are more or less in the dark, as for Parrain Circe.

This is a typical case where ep-rules should be sharpened to be able to logically answer to critical situations such as the Begley/Caillaud double capture – I don’t think you can deduce here something definite from the classical ep-rules. That’s why is added in the EP-definition that the square where the piece is transferred must be empty. It now reads:

“As a preliminary part of its move, an EP-piece X may transfer the opponent piece Y that just moved, from its arrival square to any VACANT square necessarily crossed by Y, providing X may capture Y on this square, thereby completing its move.”

Note that the word “necessarily” has also been added to handle the case of a quantic piece such as ubi-ubi.

Nikola Predrag
Nikola Predrag
November 5, 2014 09:15

Yes, your definition says that bP has completed its move on h5.
But what in standard rules says that wP has definitively moved to h5 before ep. capture? Such interpretation would indeed be incomplete.

A different question, I thought that Rose can move e.g. from d1 back to d1 if the path is empty. Popeye doesn’t allow it.
Probably Winchloe does but I don’t have it.

Juraj Lörinc
Juraj Lörinc
November 5, 2014 09:36
Reply to  Nikola Predrag

WinChloe by definition allows null moves of any kind (complete round of rose, rook move around vertical board…), but it can be switched off by option Coups nuls interdits.

Georgy Evseev
Georgy Evseev
November 5, 2014 10:15

“But what in standard rules says that wP has definitively moved to h5 before ep. capture?”

The statement 3.7b in FIDE rules: “on its first move the pawn may move as in 3.7a or alternatively it may advance two squares along the same file provided both squares are unoccupied.”

Does when in OTB game, when white play 1.e2-e4, white pawn arrives to e4, or there is something else to it?

Nikola Predrag
Nikola Predrag
November 6, 2014 00:09

Pe2 may move to e3 or alternatively ADVANCE two squares…
Only the so called “double-step” is defined as “advancing”.
All pieces (including the other movements of pawns) move to a square, except the “transfer” in castling.

So, the rules do not allow MOVING a wP from e2 to e4.
They allow moving TO e3 or alternatively advancing two squares…
Moving to e3 is a general (default) P-movement, already described in 3.7a, so why should that be repeated?
Because the formulation “may alternatively advance” needs a referential point to which it will be alternative.

Pe2 may MOVE TO e3 by general Pawns-rule but since it is the first move by that Pawn, it may “progress further” but may not REALLY MOVE FURTHER (where “to move” is specifically used for DEFINING the movements of chess pieces).

Strictly respecting the terminology of the rules, to ADVANCE 2 squares… is not the same as to MOVE to e4.
Whatever we might speculate, Pe2 may not MOVE TO e4 and thus not arrive definitevely on e4 in one move.

It is not defined for all cases WHEN wP definitively (unconditionally) arrives to e4, so we must speculate about it.
And there is the point, if we suppose that the rules are consistent, the EP (ep. capture) will confirm their consistency.
After e2-e4, the supposed legality of the EP would prove that at least a partial result of “may advance” is “may occupy” e3.

But the legality of the EP, as it is defined, absolutely excludes the definitive unconditional arrival to e4, as long as the EP is still a legal possibility (because a definitive arrival to, and occupation of e4, would exclude the possibility of the EP).
In order to make the EP legal, the formulation “may advance” allows wP to REALLY occupy e3 after advancing (but only IF Black WILL REALLY execute the EP, as defined by the rules).

However, there’s nothing in the rules defining what allows wP to REALLY occupy e4 as an unconditional result of “advancing”.
Actually, it is not the question whether wP definitively occupies e4, for as long as the EP is still a legal possibility, obviously IT DOES NOT REALLY occupy e4.

The serious question is whether wP could EVER REALLY occupy e4 after advancing from e2 to e4.
This is what we have to speculate about. The rules don’t provide a definition, as far as I can see.

If we wish, we may speculate within the limits allowed by the incomplete definition of “may advance”.
And we certainly wish it, if we want to continue the play properly, after the “advancing”.

– One possible reality after the advancing is “wP occupies e3 if Black will execute the EP”.
– If Black does not execute the EP, the legality of “advancing” requires that at least in one case, that “wP occupies e4 if Black will execute some other move”.
So, we might suppose that “in all cases, other than the execution of EP, wP REALLY occupies e4”.

Strangely enough, it seems that nothing in the rules requires the “advancing pawn” to occupy e4 or any square, except in the special cases.

My interpretation of “may advance” allows two, complementary conditional, cases:
– in case of the EP, wP occupies e3 “as though it had been moved to e3”
– in all other cases, wP occupies e4 “as though it had been moved to e4”

The reasoning is not very simple and I might be wrong, but any other belief should be supported by better reasoning.
So far, nobody else has offered anything serious enough to be even questioned.

Proofs based on the personal presumptions (or prejudices), might be valid only if the presumptions themselves could be proved by the rules and certainly not by a delusive perception of OTB players which should play by the same rules.

Georgy Evseev
Georgy Evseev
November 6, 2014 11:01

This is my last post in this thread, though I may return to this dispute if the question will be raised again in another thread. I’ll try to put here together my thoughts, including those from different posts.

1. The language is very versatile instrument. It is very easy to tell the same thing in many different ways. This allows to read between the lines or behind the letters. We should understand that such a reading better reflects initial reader’s opinion than real meaning of the text.

2. The rules of chess are formal. So, anything in the rules, concerning the movement just explains the change in position “before the move” and “after the move”. Everything else is interpretation, and this interpretation is not part of the rules.

3. There should be no real doubt that a pawn after a “long” move stands on the fourth rank. It has all of the abilities of piece standing on the fourth rank and no abilities of the piece standing on the third rank.

4. The only exception is an en passant case. And I think that it is better if it treated exactly like exception. There are two reasons for it. The first is the third/fourth controversy discussed here; and the second is that the possible moves may depend on the history of position. Trying to interpret en passant capture according to traditional approach rather creates additional difficulties than solves existing ones.

5. There is no wrong interpretation of anything in fairy chess. Any chess rule may be interpreted the way the author wants it. So, in case of any controversy or difference of opinions I think that generalization is better than limitation.

6. Chess composition is very much precedent-based. So one real problem or schematic example is generally worth much more than ten pages of theoretic argumentation.

Dominique Forlot
Dominique Forlot
November 6, 2014 13:18

– in case of the EP, wP occupies e3 “as though it had been moved to e3″

I am afraid that it infers a paradox in some orthodox cases
Let us consider the following position between Achille and Zenon:

FEN: 7b/8/8/8/4p3/8/P2P4/KB6
Achille: Pd2 Pa2 Ka1 Bb1
Zenon : Pe5 Bh8

Zenon : ” 1…e5-e4! Checkmate Achille, because you will never reached the square d4! I shall have captured your pawn before. And thus this move 2.d2-d4 will never be!

Bewildered, Achille nevertheless has no choice! he Plays 2.d2-d4! And say:

Achille : “your turn”.

Nikola Predrag
Nikola Predrag
November 7, 2014 05:20

OK Georgy there is indeed the crucial point in your 6 points:

1. ” It is very easy to tell the same thing in many different ways”
Yes, we should say the same thing.

2. Yes, AFTER e2-e4 wP is still not unconditionally/definitively on e4, in new position wP is conditionally on e4 or e3.
BEFORE the EP capture wP is on e3, BEFORE any other black move it’s on e4.

3. Do you seriously think that wKf5,wPg4,bKh5,bPh4h6, must be a mate even after a “long” move, because wP ” has all of the abilities of piece standing on the fourth rank and no abilities of the piece standing on the third rank”?
I think that abilities are not the same as in case after some other move and the result of the game might be changed quite a bit.

4. The exceptions which corrupt the organic principles of a system are caused by the inadequate point of view. Corrupting exceptions in the celestial movements were caused by geocentric viewpoint.
The potentially corrupting exception was allowing the “long” move but the EP is a beautiful amendment which preserves the organic principles .
FIDE rules avoid a corrupting interpretation that bP on forth rank may capture wP standing on 4th rank by moving to 3rd rank.
There are indeed chess rules which rely on such absurdly corrupting exceptions.
I don’t insist on FIDE rules, but only if better rules might be offered.

5. “Any chess rule may be interpreted the way the author wants it”, providing a consistent definition for everything that violates the standard rules. Otherwise, the interpretation is not detectable for a viewer.
And any viewer may interpret how significantly the chess principles are corrupted in such case.

6. Is a “real problem or schematic example generally worth” at all, if the rules or a change of standard rules are not defined?

I also say it for the last time:
When a wP definitively and unconditionally arrives on 4th rank, there is no, even exceptionally miraculous way, that bP on 4th rank may capture it. That is one of the organic principles of standard chess. Whoever changes this, at least should define and announce the change.

Dominic, when the “double-step” and ep. capture were added to the previous chess rules, double-step was sometimes even not allowed to parry the check.
That interpretation would support Zenon in your example.
So, double-step was sometimes interpreted as moving to d3 first and if the opponent alllows (refusing to apply ep. capture), wP may continue to d4. Achille’s pawn wouldn’t be allowed to jump on d4 and moving to d3 first, would be illegal.
It was not easy to change the viewpoint, many centuries ago.

After Achille has made its move d2-d4, wP conditionally occupies d3 or d4. If Zenon plays exd3ep., wP was unconditionally and definitively on d3. If Zenon plays any other move, wP was unconditionally and definitively on d4. So, e.g. Bh8-c3+ was not a legal possibility, as well as the attack on a1.
While Zenon was thinking about the EP, there was wPd3. While Zenon was thinking about any other move, there was wPd4. The rules do not allow bBh8 to attack beyond d4 as though wP had been moved to d4, but they allow exd3 as though wP had been moved to d3.

The organic principles of chess are thus preserved by conditional occupation of the squares which depends on determining the next (opponent’s) move. When the next move is determined, it just happens and the play continues.

If this viewpoint looks strange to you, then capturing wPd4 by moving bP from e4 to the empty square d3 must surely look to you as a complete nonsense.
You should claim that ep. capture is not possible at all and that it must be deleted from the rules. Then the double-step should be deleted too, because it would allow wPd2 to avoid the attack by bPe4.
And welcome to the twelfth century, or so.

Georgy Evseev
Georgy Evseev
November 7, 2014 10:15


As I cannot agree with you by several positions, I propose to switch to Forum where we may continue this discussion. Please copy you last post to Forum, or simply agree here and I’ll create a Forum thread myself.

Nikola Predrag
Nikola Predrag
November 7, 2014 10:32

OK, please open a thread, I’m clumsy with computer stuff.
There will be probably many other questions about generalizing the rules.

November 8, 2014 18:46

Christian Poisson provided me his own theoretical explanation of the ep-capture, which fits Geoff approach (see how he marked his solution of 636) : in the setting wPa2 bPb4, a2-a4 is normal and then wPa4 is captured on a4 by bPb4, which ends his move on a3. It means the black Pawn is capturing “locust-like” – the square where the capture occurs is different from the square where stands the capturing piece.

Nikola Predrag
Nikola Predrag
November 8, 2014 21:56

Yes, such Locust-Eagle behavior of a Pawn was suspected early in the discussion about No.627.
Where is a definition of that fairy piece (fairy-Pawn)?

It was not mentioned within the definition of Symmetry Circe or default Circe , so any Pawn should obey the standard definition of the ep. capture which doesn’t mention anything similar.

But the Locust-Eagle treatment at least explains why Winchloe doesn’t solve the Begley/Caillaud problem.
It was still not told which hole in the rules might justify that problem.

Diyan Kostadinov
November 24, 2014 12:08

I am sorry to say that but the way how the solution is written now is COMPLETELY WRONG and it is against the chess ruless, orthodox canons and the nature of the pawns!!
The ONLY correct way to be presented the en passant capture in the case above is 2.axb3(wPb5)e.p.!

The square of capture is b3 but the vanishing square for the wP is b4 and according the definition of Anti Take & Make (see it above) the “captured” piece “must move from its square of vanish”. What is not clear here? Why we should mess this clear situation and make such a change and discusion?

In my opinion (and I believe it is the correct one) it is very wrong practise clear rules which are followed from decades to be changed in such a way. The pawns are not Locust type pieces and can not capture like it is presented in the solution of the problem now. Also they can not capture on the square on their left/right side. They can ONLY capturing diagonally and in case of en passant they do it the same way but in this case they have the option to capture the oposite pawn on the square which is jumped over. Where is the reason to change this?

November 24, 2014 18:29

You are right of course Diyan. The Locust came into discussion only to show that the square of vanish (for reappearance) is different from square of capture !

Nikola Predrag
Nikola Predrag
November 24, 2014 19:49

“… If a piece moves to a square occupied by an opponent’s piece the latter is captured and removed from the chessboard as part of the same move.”

“The latter” is exactly the piece which occupies the square of capture.
I can’t find in the rules that the piece absurdely “vanishes” (or is removed) from some other square.

We “imitate” the moves on a material board with the material pieces.
It’s a CONVENTION that the move called “advancing two squares” is imitated AS THOUGH the pawn “has been moved two squares”.
We are expected to be aware that “imitation” is not fully equivalent to the true play (by the rules), in case of the exceptional “advancing” move.

Those, who don’t care about the rules and laws of this universe, give the prevailing importance to the imitation they’re able to see. And the whole universe is so obviously spinning around our great minds as we lay immobile watching it.
Of course, if our minds are also immobile 🙁

Just as any other piece, a pawn is a phenomenon spread across the board.
By making the move, the phenomenon of a particular piece becomes restricted/focused to a single square.
Exceptionally, the phenomenon of an “advancing pawn” remains spread two squares along the line when the move is made. The pawn was NOT MOVED to a particular single square.
It occupies 1st square as though it has been moved there or alternatively, it occupies 2nd square as though it has been moved there.
Occupying the 1st square may be the case only if it’s immediately followed by the ep. capture.

Nicolas Dupont
Nicolas Dupont
November 26, 2014 02:01


“A Pawn captures only diagonally is a canon”, true. But “a Pawn can’t capture on an empty square” is a canon too… So anyway there is something broken in the en-passant capture. Nikola’s interpretation is to keep those 2 canons, but he is breaking another one, the fact that a given move is fully determinate just before the next one is played.

So I don’t think it is possible to logically define the en-passant capture, as you more or less pretend to do. My own position is purely pragmatic – I follow how solving programs are working with it. As a matter of fact, this is also your own way, but nevertheless it can’t be considered as the ultimate truth, just as a convention.

Nikola Predrag
Nikola Predrag
November 26, 2014 05:03

Nicolas, I would perhaps understand your “pragmatic” position if would stand on my head.

You have turned the time upside-down. First there were programmers, then the inventors of fairy conditions and far away, more than 5 centuries in the “future”, there were the inventors of “double step” and en passant. And about a millennium after that, someone will invent the chess game.
And that someone will surely accept the logic of those programs which are absolutely impeccable, as we may see from some other discussions.

The “traditional” direction of time would, on the contrary, reveal that “double step” is troublesome for logical defining.

It was perfectly clear in 15th century, that a pawn MUST NOT be allowed to MOVE 2 squares ahead the same way as the line-pieces are allowed.
Because that would NOT BE CHESS!
“Double-step” requires an exceptional definition because “double-step” is the exceptional ALTERNATIVE to a natural pawn’s move.
En passant capture is a NATURAL move and such possibility simply had to be preserved by all means.
A speculation that wP may be captured on b4 by bPa4 is by far more UNNATURAL than the “conditionally determined” arrival of an “advancing” pawn.
Because the pawn-nature is defined as moving one-square forward, orthogonally to an empty square and diagonally to an occupied square.
The given name does not actually emphasize the exceptional character of en passant capture, it is the natural right of a pawn to capture a piece on the first diagonal square.
The name emphasizes that the other pawn was captured while passing through that square. Advancing is not moving to a uniquely determined square.
It was clear so long ago but 21st century schizoid man finds it incomprehensible.

And instead, the proposed 21st century logic is:
You may exceptionally count (a+X)-1=a+X, as though (exclusively on the right side of the former equation) X=(X-1).
And in the end it is only logical to apply a+X=a+X as though a+X=a+(X-1).
Isn’t that more natural than the “simple-minded” 15th century reasoning?

Nicolas Dupont
Nicolas Dupont
November 26, 2014 05:58

At least I appreciate your allusion to King Crimson!

Nikola Predrag
Nikola Predrag
November 27, 2014 02:16
Reply to  Nicolas Dupont

Nicolas, I’m very happy that at least one point was understood, so I’m perhaps not completely lost and hopelessly lonely “case” 🙂 🙂 🙂

Diyan Kostadinov
November 26, 2014 08:52

First of all I should say that I hate long comments, but now I need to do it, because probably some composers still do not understand how WRONG and DANGEROUS are these precedents in the type of presenting the solutions in originals 627 and 636!
I was keep quiet when I saw the first one (in 627), but later this was repeated in 636, and later I even started to receive problems for publications in my website from some composers in the same way… So I can not continue to be quiet and express my opinion because THIS SHOULD BE STOPPED IMMEDIATELY! Why? Let I explain again and to include some examples.

Let’s start with two wrong points in the Dupont’s comment:

1. Dupont: “A Pawn captures only diagonally is a canon”, true. But “a Pawn can’t capture on an empty square” is a canon too… ”

NO!!! The second sentence is not true! In the official chess rules by FIDE we can see that the Pawn CAN capture on an empty square in case of en passant (rule 3.7d)!

2. Dupont: “My own position is purely pragmatic – I follow how solving programs are working with it.”

It’s your right of course and many composers does the same (including me sometime). But this is not useful in all cases. The aim of the chess solving programs is not to show the most correct way of writing the solution. Their aim is just to find the final result – how many and what are the solutions. The programs can not write in all cases the full and accepted way of presenting the solutions. For example the way of presenting solutions differ in many cases between different programs, but I repeat that the program’s aim is another, so the writing solution is not the most important there. But the way of writing solutions in chess magazines (websites) is important, because they all should be clear and correct to be understandable from all (even beginners).
Sometimes not only the writing of solutions in programs is not the same with the “selected best possible” way of presenting solutions, but also the algorithm of the calculation process may differ according to the way how the program is created. So the same move can be calculated in different way in diferent program according their algorithms and this finaly can give the same correct result. But as I said this way can differ from the “humans” way of calculating or accepting the move. So sometime it is not useful to see the program’s way of calculation to understand some rules.

So here are some points why the solutions of problems 627 and 636 are presented in wrong way:

1. We should don’t forget that the the Pawn is very well defined orthodox piece from decades. According the official rules it can capture ONLY on the square which is diagonally in front of it on an adjacent file if there IS a piece OR in case of En Passat if this “square is crossed by an opponent’s pawn which has advanced two squares in one move from its original square … as though the latter had been moved only one square.” Is this not clear enough? This is an official basic chess rule from decades!
2. The pawn can not capture on its left/right square!
3. The pawn can not capture like Locust type piece (as is presented in the problem above axb4-b3!??). If you use this how will write the capturing move in following example: wRa1, bPb4 En Passant Chess: After 1.Ra1-a8 how the bP can capture? 1…bxa3 e.p. or 1…bxa8-a3 with the black pawn returns from 4th rank to its 8th rank and finished on the 3rd rank… :)) Come on… Let’s be serious 😉

4.In some fairy conditions (such as Anti Take & Make, Symetry Circe etc.) is important from which square the “captured” piece disappear. Some time (for example in case of en passant) the capturing square and the square of vanish is not the same. That’s why in the definition of these fairy conditions is noted that the following reactions (such as reborn or moving) of the captured pieces should start from its square of vanish (not from the capturing square). And this easy explain all specific situations in case of en passant!

So I ask again – when we have a clear defined rules and chess cannons why we should change them in such a quick and unneeded way, which lead to many misunderstands and may mess all? This is against the official rules of chess, this is not logical and should be stopped immediately in my opinion.
We don’t need chess composition to have orthodox rules which differ from the official FIDE rules! I never saw in some chess game en passant capture to be written in the same way as of the problem above. And you?
Please let’s we do not mess it up now…

Sorry again for my long comment and probably not so kind words. I respect all composers and all different opinions but I can not accept such a precedents which can lead to some very bad results… Hope you will understand me and what I mean.

Would love your thoughts, please comment.x