Website founded by
Milan Velimirović
in 2006

3:30 UTC
ISC 2020



Remember me

Forgot your
Click here!
to create your account if you don't already have one.

Rating lists


MatPlus.Net Forum General Isardam ambiguity & Fairy Units
You can only view this page!
(1) Posted by Kevin Begley [Sunday, Dec 19, 2010 15:03]; edited by Kevin Begley [10-12-19]

Isardam ambiguity & Fairy Units

Yet another ambiguous Isardam rule (though, I am aware of no problem which specifically exploits this).
[note: this might also constitute an ambiguous Madrasi rule.]

(= 2+2 )

#1? (or is it =1?)

= Chameleon Rook

a) = Orthodox Knight
b) = Chameleon Knight

The question here is: do equivalent powers form the basis for Madrasi Paralysis, or must the units be exactly equivalent?

a) 1.cRh8[=cQ]# is legal mate in Win Chloe, but popeye v4.41 does not agree (it considers this stalemate).
b) 1.cRh8[=cQ]= is neither check, nor mate (here, both programs agree this is stalemate).

I would presume virtually everyone agrees with Win Chloe's interpretation.
If Chameleon Rooks can not castle, they must be inherently different pieces; and, this principle should extend to all chameleons.
The matter becomes considerably more complicated when extended to Orphans (and pieces of this type).

I believe it is always in our best interest to locate & challenge such ambiguities (in problem form); and, we should reward this effort.
We celebrate a rich history of exploiting loopholes in FIDE's Chess rules -- in so doing, composers helped improve upon FIDE's rules.
I find it ironic that some despise the exploitation of our own loopholes -- let's be sporting, even when the joke is on us.
But, let's also be responsible (and responsive) in improving upon our own rules.

Chess has an official authority responsible for drafting its rules, whereas Fairy Chess has none.
In this capacity, FIDE has always been fairly responsive in removing any ambiguities demonstrated by problemists.

It is true that Fairy Chess has vast rules.
As such, many have suggested that the task of achieving a Fairy Codex is beyond us.
Some have even gone so far as to suggest (without any proof) that Godel's Theorem might render hopeless any such effort.

I do not subscribe to the prevailing pessimism.
First, begin from the bottom up (from aims to fairy elements), precisely defining our fundamental terms.
Second, establish consistent principles to resolve special cases (e.g., how do pawns function when upon their own 1st rank?).
Third, consider granting "official sanction" to elementary fairy elements which are fully consistent with all things sanctioned.
[Before ruling, encourage colleagues to present problems which exploit ambiguities in any fairy elements being considered.]

Concurrently, some fairy-element version control must be established (starting from the simplest, and longest established).
Along with efforts to recast such problems, version control must cover problems depending upon unsanctioned fairy elements.
As time goes on, we may begin to consider more complex fairy elements.

This is how we can begin building a significant Fairy Codex, which would serve us all well.
And, it would allow future generations to build upon it -- either we pass this as a gift, or we leave it as a burden.
Annual suggestions by the Fairy Rules Council should require ratification by the complete delegation.

It is really not so difficult to remove ambiguity.
I myself had the foresight to explicitly remove some Madrasi/Isardam ambiguities, when introducing my own fairy invention (psycho-units).
It is a simple matter of using precise wording, keeping an eye on the bigger picture, and encouraging colleagues to help.
[edit: and also leaving room for revision!]

Our rules can no longer default upon the full agreement of popeye, Win Chloe, and Fairybadix...
Though the opinions of programmers should be sought (and highly regarded), it is our own responsibility to establish precise, unambiguous rules, and to protest any element which does not fully align.
(Read Only)pid=6492
(2) Posted by Hauke Reddmann [Sunday, Dec 19, 2010 15:37]

Funny you mention Gödel since them dude was the first
I thought of in the Big MPF Rule Wars :-)
But it's easy to see that he can't apply:
Chess is a completely finite game. All fairy pieces
known at a time are finitely many too. All versions
of chess can be reduced to very long list of allowed
state1->state2 alterations (unless the board is
infinite, or there is the possibility of history-
dependent moves. E.g. Castling falls under the latter,
but this adds just one bit to the description - you
"need" infinite history-dependency).
Summing up, chess or almost any fairy chess version
can be defined as a very long, but finite, list
of allowed moves. No undecidably here except that
you have to agree WHICH list.

(Read Only)pid=6493
(3) Posted by Bojan Basic [Sunday, Dec 19, 2010 15:38]

Could you check what WinChloe says for:

(= 2+2 )

Is c) 1.cRh8[=cQ] mate or stalemate? That is, does chameleon always paralyse chameleon, or they must be in the same phase?
(Read Only)pid=6494
(4) Posted by Kevin Begley [Sunday, Dec 19, 2010 16:26]; edited by Kevin Begley [10-12-19]

Yes, Bojan, I had considered this too.

Note: Win Chloe implements each chameleon as a separate unit.
Thus, Madrasi-Parlaysis, and rules of Isardam, follow directly from this implementation (which I consider to be correct).

1.cRh8[=cQ], is simply checkmate (win chloe agrees), but remember the cR becomes cQ (leaving no Isardam issue between h1 and h8).

Perhaps you mean with a chameleon Queen on h1...

[edit: some revision was required.]
Using black Chameleon Qh1, Win chloe correctly considers 1.cRh8[=cQ] to be checkmate.
At first glance it appears that no check is given -- but this is false (the threat is 2.cQxh6[=cS], and there is no Isardam defense).

If you move the black cQ from h1 to h4, then 1.cRh8[=cQ]+ is only answered by the Isardam defense: 1...cQg4[=cS]!

(= 2+2 )

#1? Isardam

chameleons a1, h1

No mate.
1.cBh8[=cR] <== This is not checkmate, because 1...cRc1[=cQ] (or 1...cRh~[=cQ]) defends against the threat (whereas 1...Kg7? is clearly illegal).
(Read Only)pid=6496
(5) Posted by Neal Turner [Sunday, Dec 19, 2010 16:42]

My beef with Isardam is the idea that a check can be cancelled by defending the king with the same type of piece as the one giving the check.
The logic seems to be that it is now impossible for the checking piece to capture the king.
But in chess the king is never captured and and giving check is not threatening to capture the king - in fact FIDE now have a rule in blitz chess stating that capturing the opponent's king is an illegal move leading to loss of the game!
(Read Only)pid=6497
(6) Posted by Kevin Begley [Sunday, Dec 19, 2010 16:46]; edited by Kevin Begley [10-12-19]


>My beef with Isardam is the idea that a check can be cancelled by defending the king with the same type of piece as the one giving the check.

Perhaps you might consider Isardam Type B -- where check cannot be canceled in this manner.
This is fully implemented in Win Chloe, though I'm not sure about the latest version of popeye.

There are a few (but very few) problems composed using this less complex version.
(Read Only)pid=6498
(7) Posted by Juraj Lörinc [Monday, Dec 20, 2010 13:39]

For me, this thread does touch upon two different issues. One very specific - Isardam & chameleons, indeed very current issue due to the tourney. One very general - Fairy Codex. Let me react to both in separate paragraphs.

Isardam & chameleons: For me also WinChloe interpretation seems the right one. In fairy chess the check/mate presence is usually tested by virtual capture of the king to see whether it is indeed legal. There are some fairy elements explicitly stating that check/mate is orthodox or considered other way, but Isardam clearly is not the case.

What I however find interesting that Chameleons have to be in the same phase to be paralysed in Madrasi. It is not self-evident in my view, althogh as I already have written, I welcome this intepretation. Why?

Because once, when I was just starting with fairy chess I have asked about possible explanation/interpretation of fairy rule more experienced man. I no longer remember the specific rule, but I remember the answer that was very general. "When you are creating the new fairy element or you have to interpret the unusual and explicitly undefined case of fairy elements combination, always try to choose the most fruitful rule. The one promising the most interesting problems." It requires some foresight, but it usually comes with experience. Of course, it is always better to have rules written, but is it conceivable?

This brings me to...

Fairy Codex: It would be nice to have, but I do not think it is possible we will something like that in my lifetime. Even now authors differ in the interpretations and imagine writing a kind of one definitive text. That would lead to endless discussions and controversies. Just remember how long it took to publish "normal Codex".

I do not say it is impossible, but my guess is that most people would say it is not worth the time and effort invested. I will just be happy if a group of people proves me wrong.
(Read Only)pid=6512
(8) Posted by Joost de Heer [Monday, Dec 20, 2010 14:02]

My interpretation is that 'chameleon' is a property of a piece, just like 'magic' or 'patrolling'. A magic rook paralyses a normal rook in Madrasi, so why wouldn't a chameleon rook paralyse a normal rook (or a magic rook) in Madrasi?
(Read Only)pid=6514
(9) Posted by Kevin Begley [Monday, Dec 20, 2010 16:28]; edited by Kevin Begley [10-12-20]


Good point, but if 'chameleon' is a property of a piece, then what prevents me from castling with a Chameleon-Rook?

Further, suppose I have a unit which moves as a Bishop, but attacks as a Rook.
What orthodox units (Bishops/Rooks/Queen?) can madrasi-paralyze this unit?
(Read Only)pid=6516
(10) Posted by Joost de Heer [Monday, Dec 20, 2010 17:10]


Further, suppose I have a unit which moves as a Bishop, but attacks as a Rook.
What orthodox units (Bishops/Rooks/Queen?) can madrasi-paralyze this unit?

It can only be paralysed by a piece that moves like a bishop, but attacks as a rook. Something similar happens with Chinese pieces: A pao on a5 will only paralyse a pao on h5 if there is a piece in between. A rook-lion won't paralyse a pao, even though their 'capture-movement' is equal.
(Read Only)pid=6517
(11) Posted by Kevin Begley [Tuesday, Dec 21, 2010 16:26]; edited by Kevin Begley [10-12-21]


>My interpretation is that 'chameleon' is a property of a piece, just like 'magic' or 'patrolling'.
>A magic rook paralyses a normal rook in Madrasi, so why wouldn't a chameleon rook paralyse a normal rook (or a magic rook) in Madrasi?

I presume we can agree that neutral (and half-neutral) are properties (certainly, these are not colors).
And, I will concede that 'magic' and 'patrolling' are similarly properties.
However, there is evidence that chameleons are different from each of the above properties.

First, as already mentioned, castling is expressly forbidden with a chameleon Rook (and this spans all conditions).
Under the right conditions (e.g., anti-Kings), it should be possible to castle with both a neutral Rook, and a half-neutral Rook).
It is already possible to castle with a magic-Rook, or patrolling-Rook.
This affects all chameleons, by logical extension.

Second, in Circe, a chameleon-rook is not reborn to the square that an orthodox rook would be -- it rebirths as a fairy unit (promotion square).
This is established in several published problems, and is testable both with popeye, and Win Chloe.
For any other property (neutrals, half-neutrals, magic-units, and patrolling-units), rebirth is equivalent to the orthodox unit.

These might be unfortunate inconsistencies (I'll be the first to concede that these things do happen).

But, if 'chameleon' is a property of a unit, what exactly is the fundamental nature of the unit?
It would be entirely dependent upon its present phase, within the solution (it would be a blank canvas, a unit made of clay).
Unlike all other properties cited (in which a unit's fundamental nature remains intact), the 'chameleon' property would cause an alteration at the core of a unit's nature.
Thus, Madrasi paralysis would depend upon the property and phase of a unit, without any regard for its fundamental nature.

But, what I find most troubling is that your interpretation provides no means for determining what is, and what is not, a property.
What is the principle at work, so that I might determine what is a 'unique' unit, and what are merely the 'properties' of a unit?
If a grasshopper/andernach-grasshopper are mutually attacking, how can I determine whether madrasi-paralysis ensues?

Is it based upon an equivalent movement/attack pattern?
If so, then a Fersrider -- (1,1)Rider, like bishop, but different rebirth square -- must paralyze an attacked Bishop, in Madrasi.
Presently, in popeye, mutual-attacks by such units will not result in Madrasi paralysis.
It is unreasonable to expect a program to calculate movement and attack patterns.
Particularly considering that this would require an infinite board -- on an 8x8 board, a (4,4)Leaper seems equivalent to a (4,4)Rider.

What about Reflecting-Units (or Transmuting-Units) -- can they be attacked in Isardam?
Can they only be double-attacked?
For Madrasi Paralysis to occur with two mutual-attacking Orphans, must the phase of each exactly match?
Here are units, like chameleons, where their fundamental nature may wildly change, during the course of the solution.

As I see it, any property which would alter the fundamental nature of a unit, can not be a property (it must be a fundamentally unique unit).
Nevertheless, I do not profess to know how to resolve the ambiguities involved with units of changing nature (as mentioned above).

It is amusing how little is certain about Madrasi/Isardam -- given that their rules are far more elegantly stated than most conditions.
Meanwhile, our two prominent fairy "schools" continue stimulating zoo constructions, in clouds settled upon shifting sand.
If the certain structural tragedy were not so comical, it would offend humanity. :-)
(Read Only)pid=6527

No more posts

MatPlus.Net Forum General Isardam ambiguity & Fairy Units