22:51 CET

Remember me

 Rating ListJanuary 1st2013

MatPlus.Net Forum General Circe Coucou versus Circe Couscous

### Circe Coucou versus Circe Couscous

Do I have this straight...

Circe Coucou - (invented by Pierre Monreal, circa 1980?)
It's the same as Circe, except the captured unit is reborn onto the "circe rebirth square" of the capturing unit.
Pawns captured by a piece are reborn into immediate promotion, and the promotion is chosen by the capturing side.
Pieces captured by promoting pawns are reborn according to the promotion chosen (as if captured by the promoted piece).

Circe Couscous - (invented by ?, circa ?)
It is exactly the same as Circe Coucou, except for two important difference:
1) the promotion is chosen by the promoting side (rather than the capturing side), and
2) Pieces captured by promoting pawns are reborn as if captured by the pawn (onto the same file, independent of promotion choice).

Do I have this right?
Are there any other subtle differences?

The capturing side chooses the promotion in Circé Couscous, while the promoting side chooses the promotion in Circé Coucou (not vice versa).

At http://jurajlorinc.tripod.com/chess/qpromo02.htm#uloha7, Juraj Lörinc does not mention the second difference of yours.

Thanks Bojan -- indeed, I had those rules reversed.

This second difference is not often mentioned (neither does it appear on Wiki's page)...
Unfortunately, the rules of some circe forms would require considerably more space to explain.
I found this difference by browsing problems in Win Chloe's database:

René Jean Millour
1st Prize, The Problemist, 2003

h=3 0.4.1... Circe Coucou (C?)
b) bKe3>e2

a)
1…fxe8=Q(Rd1) 2.bxc1=Q(Rd8)+ Sxc1 3.Rd4 Rxd4=
1…fxe8=R(Rh1) 2.bxc1=R(Rh8) Sxc1 3.Rh4 Rxh4=
1…fxe8=B(Rf1) 2.bxc1=B(Rf8)+ Sxc1 3.Rf4 Rxf4=
1…fxe8=S(Rb1) 2.bxc1=S(Rb8) Sxc1 3.Rb4 Rxb4=
b)
1…fxe8=Q(Rd1) 2.bxa1=R(Bh8)+ Sxa1 3.Rd4 Bxd4=
1…fxe8=R(Rh1) 2.bxa1=B(Bf8) Sxa1 3.Rh6 Bxh6=
1…fxe8=B(Rf1) 2.bxa1=S(Bb8) Sxa1 3.Rf4 Bxf4=
1…fxe8=S(Rb1) 2.bxa1=Q(Bd8)+ Sxa1 3.Rb6 Bxb6=

Here is Coucou:
Black's first move shows a captured bR reborn onto different squares (d1, h1, f1, or b1) according to promotion choice.

Uri Avner
2nd Place, Nunspeet T.T., 2006

s#3 Circe Couscous (C+)

1.Rg7! (zz)
1…Ra8 2.dxe8=B(Bf1)+ Kxe8 3.bxa8=R(Rh1)+ Bxg2(Sc8)#
1…Rb8 2.dxe8=Q(Bd1)+ Kxe8 3.axb8=R(Ra1)+ Bxc2(Bc8)#
1…Bxd7 2.Rh7 [> 3.Rxh8(Sa1)+ Sxc2(Bg8)#]

Here is Couscous -- consider the first two variations:
White's second move (capturing bBe8) does not result in rebirth (on d1 or f1) -- the rebirth square (e2) is occupied.
White's third move (capturing bRa8) does not result in rebirth (on h1) -- the rebirth square (a2) is occupied.

By the way, Win Chloe tests only the latter form (Couscous).
I suspect some helpmate problems (many could function under either condition) are called Couscous for testing purposes.

QUOTE
Here is Couscous -- consider the first two variations:
White's second move (capturing bBe8) does not result in rebirth (on d1 or f1) -- the rebirth square (e2) is occupied.
White's third move (capturing bRa8) does not result in rebirth (on h1) -- the rebirth square (a2) is occupied.

Umm...

QUOTE
1.Rg7! (zz)
1…Ra8 2.dxe8=B(Bf1)+ Kxe8 3.bxa8=R(Rh1)+ Bxg2(Sc8)#
1…Rb8 2.dxe8=Q(Bd1)+ Kxe8 3.axb8=R(Ra1)+ Bxc2(Bc8)#
1…Bxd7 2.Rh7 [> 3.Rxh8(Sa1)+ Sxc2(Bg8)#]

Err, disregard... I seem to have this confused -- I now see that I have made a case against my own claim.

For the record, beware of Win Chloe's Redaction Mode feature (which does not show rebirths on d1/f1 or a1/h1)!
If you're not careful, it can cause serious delusions.

Hold the phone -- disregard that "disregard" (earlier)!

If I change the first problem's (Rene's h=2.5) condition, from Coucou to Couscous, and I have WC solve it...

Here is one line (of many) it gives (translated into English notation): 1…g4 2.Ke2 fxe8=Q 3.bxa1=Q(Ba7)+ Sxa1=
So, why was the bR not reborn onto d1 (following it's capture by a promoted Queen)? [It seems e2 is used, instead.]
Why was the wB reborn onto a7 (not d8)?

Is this is a bug?

Strangely enough, Problemesis defines the two conditions as follows:

Circe Couscous:
When a capture is made, the captured unit (except a King) is replaced on the capturing unit rebirth square (according to the "modalités Circé") if it is empty : otherwise, the captured unit vanishes.
Unless otherwise stated, when a promotion by capture occurs, the rebirth square is the Pawn one.

Circe Coucou:
Like Circé couscous, but the capture of a Pawn is forbidden if it must promote and if one of the possible promotions is a self-check.

Setting aside the strange definition of Coucou...
In neither of the quoted problems do I find it "otherwise stated."
So, why is rebirth listed according to the promotion choice (rather than the pawn)?

Furthermore, Uri's problem is listed as C+, but neither Win Chloe (v3.06) nor popeye (v4.41) find a solution.

C. Poisson says (http://christian.poisson.free.fr/problemesis/condra.html):

QUOTE
Circé couscous: When a capture is made, the captured unit (except a King) is replaced on the capturing unit rebirth square (according to the "modalités Circé") if it is empty : otherwise, the captured unit vanishes.
Unless otherwise stated, when a promotion by capture occurs, the rebirth square is the Pawn one.

QUOTE
Circé coucou: Like Circé couscous, but the capture of a Pawn is forbidden if it must promote and if one of the possible promotions is a self-check.

What can we see? First of all, Christian agrees that the only difference between Circé couscous and Circé coucou concerns capturing a pawn which promotes on reborn (although his definition of Circé coucou is pretty sloppy, we know the meaning). However, he states that the default rebirth square after promotion by capture is the pawn one (in Circé couscous, but surely also in Circé coucou, by his interpretation), and that it may be stated otherwise. (I see no problem so far—since, if both conventions really exists, one of them had to be chosen as the default one in Winchloe, and the other one then had to be stated explicitly.) Could you check whether something like that can be stated in "complementary conditions" (I am not Winchloe owner)?

In Popeye, the default rebirth square is the one of the promoted piece. Popeye v4.55 solves Uri’s problem perfectly.

Correction: Popeye 4.41 does solve Uri's s#3.
The difference appears to be that in popeye, rebirth is based upon the promotion choice.
In Win Chloe, rebirth is based upon the Pawn.

scheme

h#1 Circe Couscous

Win Chloe 3.06: 2 solutions

1.bxc1=B(Rc7)+ Rxc1(Ba1)#
1.bxc1=S(Rc7) Rxc1(Sa1)#

Popeye 4.41: 1 solution

1.b2*c1=R [+wRh8] Rh8-h1 #

As for "complementary conditions," I am not aware of such a complimentary condition.
Win Chloe does also allow for "Circe Couscous malefique" -- but this is not what is being used here.
And, for the record, under "Circe Couscous malefique," the scheme has 4 solutions: 1.bxc1=~(Rc2)+ Rxc1(~h8)#

"Rules by Default," according to Problemesis, says:
"When a promotion by capture occurs, the promotion precedes the capture."

So, Christian went out of his way to implement the alternative rules for rebirth in Circe Couscous (capture first, then promote).
This suggests the inventor of Circe Couscous might have specified the rule (as contrary to the default).
Can anybody confirm this?

I've discovered a 5-step program to really help people understand Fairy Chess:

Step 2: type in the first rule for Fairy Chess: "Rules must not be universal."
Step 3: translate this from English into Greek: => "Οι κανόνες δεν πρέπει να είναι καθολική."
Step 4: copy & paste the Greek text into the translation window
Step 5: translate this back from Greek into English: => "The rules should be universal."

This is why the first rule of Fairy Chess is: The rules should be universal.

After that h#1 exmple, I wonder whether magazines would accept a problem showing AUW or Babson with a twinning like:
a) winchloe
b) popeye
as the simpliest way to specify preciselly a fairy condition.

(11) Posted by Jan Hein Verduin [Wednesday, Jan 12, 2011 20:47]

As i understand it there’s another difference between coucou and couscous, one that follows from the definition: in coucou the capture of a pawn is illegal if any piece that pawn could promote to would give check (in couscous of course the capturing side may not select a checking promotion). Presumably this goes for helpplay type of problems too (including SPG’s), even if the ‘helping’ side would not choose a promotion that checks; what matters is he can choose a promotion that checks. So my initial thought –that in helpproblems the difference between coucou and couscous is academic since it's irrelevant who decides what to promote to- turns out to be wrong. Here's an example that illustrates what i mean - that is, if i'm right!

Original

h±1
a) couscous
b) coucou

a) 1.Kxf6(+Qe8),Qg6± (after 1.Ke8?,f7 2.Kxf7(+R/Se8) parries the check)
b) 1.Ke8,f7± (1.Kxf6 is selfcheck; white could choose a Se8)

QUOTE
After that h#1 exmple, I wonder whether magazines would accept a problem showing AUW or Babson with a twinning like:
a) winchloe
b) popeye
as the simpliest way to specify preciselly a fairy condition.

Here is a similar example. These two problems are published next to each other.

Peter Harris
h=3.5 (C+ by Popeye)
lions
a) AntiCircé Isardam

Peter Harris & Guy Sobrecases
h=3.5 (C+ by Winchloe)
lions
a) AntiCircé Isardam

However, I don’t think that a twinning like a) Winchloe, b) Popeye is really the best way to present the problem. The difference between the two programs always comes down to a different interpretation of the rules in some specific situation (unless one of the two programs has a bug, in which case there is no sense in the mentioned twinning), and I believe it would be much more clear to simply describe the interpretations used in corresponding twins (like, in Kevin’s example: a) rebirth precedes promotion, b) promotion precedes rebirth). It cannot be expected that an average solver is aware of all the specific differences between solving programs (and I don’t even mention the possibility that a program is later recoded to behave differently, or cease to exist!). With describing the interpretations explicitly, everything becomes much more sensible. Actually, this is precisely how things already work with some fairy conditions having dubious interpretations in specific situations (and without explicit names for the possible interpretations—unlike, e.g. AntiCircé type Cheylan/Calvet): where is a locust-like piece reborn in AntiCircé?; is promotion to imitator allowed?; do pawns promote on the last rank of horizontal cylinder?; does joker promote when reaches the last rank while moving as a pawn? etc. Answers to these questions, if needed, usually are stated explicitly under the diagram, and I see no problem with that. Interestingly, considering the first of this questions, Popeye and Winchloe had different opinions about it (and they still do, although WinChloe later offered additional condition "Une piece capturante renait relativement a la case de capture," to cover the other convention, too). :)

Bojan,

I almost entirely agree...

>I believe it would be much more clear to simply describe the interpretations used in corresponding twins (like, in Kevin’s example: a) rebirth precedes promotion, b) promotion precedes rebirth). It cannot be expected that an average solver is aware of all the specific differences between solving programs (and I don’t even mention the possibility that a program is later recoded to behave differently, or cease to exist!).

The above twinning (by Harris & Sobrecases) is short-sighted (if not damn sloppy).
I don't suppose they intended these problems to endure beyond the next version of these programs.

However, there is serious a flaw in the twinning form used in my example -- it provides the solver too much information.
I might as well put in red letters, "promotion will occur here."
While promotion is completely obvious in the example I gave, it is unfortunate if such information will only be provided when it specifically relates to the solution.
Providing such information may not always be "sensible."

For example, there has been some disagreement whether extra-long castling is a possibility in Vertical Cylinder problems.
If it were not a possibility, and the composer must instantiate this (e.g., Vertical Cylinder Type ExtraOO"), the solver can not avoid the logical deduction: castling must be a part of the solution (otherwise, normal Vertical Cylinder would have been used).

I'll not get into the case for (or against) extra-long castling... except to say that the rules for V-Cylinder did not explicitly exclude the possibility, nor did they explicitly endorse it.
It is unfortunate that the inventor's failure what not observed -- and called out! -- sooner.

We have no "official" Fairy Codex... there are no "default" laws to be found here... and there's nobody to uphold them, anyway.
Almost nobody is happy with the situation -- and we are all familiar with the difficulty in resolving ambiguous rules after the fact.
Yet, the sloppy inventions never cease...

So, one might expect that editors, and journals, and programmers would do their part, to collectively establish default rules, counsel fairy inventors, and demand that ambiguities be eliminated (prior to any publication/implementation).
You might expect that all parties would have demanded the official governing body (now called WFCC) provide some oversight (in the interest of consistency).
But, your behavior model would be centuries off what really has happened -- applying a wild-west mentality to all of the principal characters will generally provide the best results (as a predictor of history).

So, as good citizens of the Fairy Genre, we must escort these characters in front of a mirror.
If an editor prints a half-assed set of rules in some journal, call the entire journal out for its failure.
Journals dropped the ball, plenty, in the past -- it can not continue. They can no longer blame space considerations (today, they have ample space online).
If they can not meet their basic obligations to this genre, they should get out of the fairy chess business.
If journals were smart, they'd collaborate, to provide universal rules (spanning many languages).

If two programs have implemented something entirely different, because the original rules were ambiguous, both programs have failed to insist upon a complete specification!
I fully believe there is nobody better suited to finding ambiguities, than programmers... it's a pity they don't do this!

>Where is a locust-like piece reborn in AntiCircé?; is promotion to imitator allowed?; do pawns promote on the last rank of horizontal cylinder?; does joker promote when reaches the last rank while moving as a pawn? etc. Answers to these questions, if needed, usually are stated explicitly under the diagram, and I see no problem with that.

As I noted above, there are problems with this technique (excessive information).
It is better to encapsulate the rules (and try to minimize, through greater consistency, the need for excessive condition-types, such as Anticirce Cheylan/Calvet, or Isardam normal/Type B/Rex Inclusiv/more to come?).

>Interestingly, considering the first of this questions, Popeye and Winchloe had different opinions about it (and they still do, although WinChloe later offered additional condition "Une piece capturante renait relativement a la case de capture," to cover the other convention, too). :)

They interpret many things differently.
How do pawns on the 1st behave in Circe Equipollents?
* Win Chloe defaults to: 1-square forward plus captures, as in Circe Parrain,
* popeye defaults to: immobile as in PWC.

The original rules of Circe Equipollents did not explicitly say.
Either the inventor failed us, or the journal did (which printed these incomplete rules).
The programmers should have insisted upon a complete (and public!) specification of the rules.

But, ultimately, programmers are not responsible to officiate these matters -- this duty must fall upon the governing body.
WFCC must establish a sanctioning agency, to oversee the fairy rules (naturally, programs will desire to boast of their compliance).

Nothing today is certain in Circe Equipollents.
Popeye's old interpretation of en passant seems to me a bug.
Neither set of rules, for pawns on the 1st, can be considered "official."

Thus, the condition is not universally stable -- many details must be explicitly stated, beneath the diagram.

And, from these divergent sets of rules, new rules are constantly emerging...
How will AntiCirce-Equipollents ever seem logical for those who trusted popeye's implementation of Circe Equipolents?

No matter whose rules you believe, the inconsistencies in this genre are a self-replicating fiasco.
And, nobody has a plan to improve the situation.

It is quite possible that WFCC has no authority in this matter.
If WFCC is beneath the FIDE umbrella, how can they have authority beyond Chess (a single variant of Shantraj)?
How can WFCC award a FIDE title to Chinese Chess problem composers (when their point system is biased to favor orthodox Chess)?
Chinese Chess has its own federation.

Perhaps the responsibility of providing a Fairy Codex must fall to another organization -- one with a broader charter, unbiased towards any single Shantraj variant, and out from under FIDE's wing.

Thank you, Bojan, for showing these problems. Nice memories with Peter! Thank you also Kevin for your interest. These two problems do not deserve so many applauses. These are two different problems, and one can notice that there is here no twinning a) Winchloe b) Popeye here. Indeed, I never saw such twinning. On ProblemOnline, I could see a cyber-collaboration between these two softwares... but that's another matter!
http://www.nikola-predrag.from.hr/problemonline_2004-2008/miscellaneous_7-12.htm

QUOTE
"Rules by Default," according to Problemesis, says:
"When a promotion by capture occurs, the promotion precedes the capture."
So, Christian went out of his way to implement the alternative rules for rebirth in Circe Couscous (capture first, then promote).
This suggests the inventor of Circe Couscous might have specified the rule (as contrary to the default).
Can anybody confirm this?

I asked Christian, who replied as follows:
It should not be in Rules by Default as it only applies for Anticirce.
In WinChloe, you must add the condition "La promotion précède la capture" (not programmed).

Geoff,

Christian's accounting is incorrect -- the importance of explicitly establishing default promotion/capture priorities is in no way exclusive to Anticirce forms.
Obviously, it would also affect Circe-Coucou/Circe-Couscous (one of which follows the default rule, the other must explicitly state its non-compliance).
It also affects the reversed form of "Take & Make" (of course, T&M is just a non-compliant form of Anticirce, though not yet widely recognized as such), which I called Circe T&M (described in another post on this forum).

And, it affects other conditions which Christian is not accounting for...
Consider a simple condition, such as "no doubled pawns."
The default capture/promotion priority would establish whether a player, having a pawn at home on the h-file, can legally play gxh=Q/R/B/S.
If there were no default rule, the inventor of "no doubled pawns" must explicitly state this rule (in order to remove the ambiguity); whereas with a default rule, one need only explicitly state non-compliance.
Either way, this default rule certainly affects more than Circe & Anticirce Forms.

And, when fairy elements are combined, this default becomes especially important.
Without the support of intelligent, overarching default rules, combinations of fairy conditions lead to ambiguity, and collapse.
Imagine "single-box" and "Isardam" on a "Horizontal Cylinder."
You'd need a paragraph (explaining capture/promotion priorities), under the diagram.

Instead, we could all agree upon a small set of intelligent rules by default -- this would save space, reduce ambiguities (and confusion -- even for fairy elements yet to be invented), and provide greater overall stability for fairy chess (especially when combinations of these "elements" are employed).

Fairy elements should be viewed as elementary (atomic).
Default rules help to more elegantly describe the physics/chemistry by which these elements behave/combine, for any given space (whatever the shape of the chess board universe), over time (each quantum tempi).
Beyond the struggle, and the art/science, of a constantly changing version of an ancient board, Chess should be viewed as the underlying mathematics, which binds all that Mars gave to Caissa.