MatPlus.Net

 Website founded by
Milan Velimirović
in 2006

19:35 UTC
ISC 2024
 
  Forum*
 
 
 
 

Username:

Password:

Remember me

 
Forgot your
password?
Click here!
SIGN IN
to create your account if you don't already have one.
CHESS
SOLVING

Tournaments
Rating lists
1-Jan-2024

B P C F





 
 
MatPlus.Net Forum General What is a "line pin" and are there undefined pins?
 
You can only view this page!
Page: [Previous] [Next] 1 2 3 4 5 6 7 8 9 10
(121) Posted by Nikola Predrag [Saturday, Oct 18, 2014 14:25]

Georgy (and Hauke),
"...So the word "removal" instead, or "absence", if you prefer it, is better, as it does not allow these questions to appear.
..."
"...Thus I find it more natural to resort to a remove-X definition
than a move-X-definition..."

Simple removal of a piece is absurd/meaningless with regards to the rules.
Removal/absence should be considered as a result of some hypothetical move. A completed move includes the departure, arrival and other effects. Interpretation of the "pin" depends on the effects which are extracted as relevant but DISREGARDING THE IRRELEVANT EFFECTS.

The main cause for wrong and dogmatic interpretations is a lack of care about determining the irrelevant features.
In many cases, we might say that for the Pin, all effects of a hypothetical move are irrelevant, except those which could be interpreted as "removal/absence of a piece".
But that is not good enough for generalization. Pin is an interpretation of the square-properties. Sentinels is defined by the additional departure-effect which is not irrelevant for the Pin.

Neal,
battery and pin are not the rules(laws) but 'board', 'piece', 'move', 'capture', 'check' etc. are.
If we can define any set of general rules of "G(eneral)-chess" which would be obeyed by Orthodox chess, we might say that Orthodox chess is just a variant of G-chess.
Different variants of G-chess could be considered as different "games", as you say, but still having something specific in common.
The point is to determine which feature(s) will distinct chess from non-chess (e.g. football).
If we say that the general rules are "there are two opposing sides which both try to win or at least not to lose", that obviously won't be of any help.
 
   
(Read Only)pid=12857
(122) Posted by Nikola Predrag [Saturday, Oct 18, 2014 15:02]

"...Concerning your interpretation of position as a set of squares with different properties, it is, of course, possible, but seems even more difficult to generalize. So traditional approach with pieces seems more fruitful)) to me...."

That is certainly true, I can't imagine to formulate the chess rules without using the concept of pieces.
But that's not the point, we just should be aware that some properties of the pieces could be overlooked.
Either a Queen is not exactly the same piece in case of different conditions, or it is always the same but the condition changes the square-properties.

Do the Pawns occur in Sentinels as the "eggs" left by a piece or they pop-up out of a vacated square, or anything else?
Whatever explanation might be offered, it should be consistent and useful.

Square-properties could serve as a tool, I don't claim that they must be the basis for definitions.
However, in my speculations about the definitions of check and pins&batteries, the square-properties have appeared as a rather promising tool.
 
 
(Read Only)pid=12858
(123) Posted by Kevin Begley [Saturday, Oct 18, 2014 15:23]; edited by Kevin Begley [14-10-18]

@Georgy,

SPIKE is a nonsensical term, which lacks any precise definition -- I refuse to use such terms, because they only translate to "goo goo ga ga".
(note: I intend no personal offense by this phrase -- which is defined in the Urban Dictionary -- it is merely an honest translation of terminology used, absent any credible definition).

Yes, some of my definitions were intentionally described in a rigorous manner (based strictly upon the analysis of an objective, determinate algorithm).
That is the ever-loving point -- to provide a methodology (rather than persuade gullible members of the community to swallow more pretense, masquerading as a definition).

There are times when you want to define a thing -- e.g., you may define a variable (p) to be prime (and the mere assertion makes it so).
Then, there are times when you want to describe the analytic meaning of a thing's definition -- e.g., you may offer a mathematical statement describing the meaning of the variable (p).
It should be no secret that my intent was the latter.

I appreciate that this may seem an excessive complication, for some; and, I have no quarrel with the claim that I might have stated this in simpler terms.
However, you may presume that my intended audience must not include those unable to follow a relatively straightforward analysis.
Yes, I pay a small price for this exclusion; but, WE ALL PAY a much higher price for the bogus definitions which have been routinely accepted (and passed down).

I now return you to your semantics argument (for which you have established no legitimate arbiter), about the proper label for a "simple" term, which you can not define.
 
   
(Read Only)pid=12859
(124) Posted by Kevin Begley [Saturday, Oct 18, 2014 16:14]; edited by Kevin Begley [14-10-18]

@Nikola,

I like your G(eneral)-Chess idea (very much!) -- if done smartly (in the most general and logical form, which not necessarily the most orthodox), it could prove very useful.

I do not agree with the claims about "rules" ...
Check is not a rule (it is a defined state of a position)... checkmate, in fact, is also a state of a position, derived from CHECK-STATE, plus LEGAL-MOVE-STATE.
Of course, this only holds for a DETERMINATE class of chess problems (where all states can be determined from a position -- however you define position).
In orthodox chess, position has many subtle features (unfortunately, it requires inclusion of the entire game history).

-capture- is a feature of a move (same goes for a number of similar aims: -castle-, -ep-, -promote-, etc).
note: -annihilation- can be different from capture, and where it is, it might make for an interesting aim (though I don't recall seeing it).

Nor is the square a fundamental element of a BOARD (oddly enough, for the chessboard, the square is not square-one).
You have to go one level beneath the square -- to consider: What is the fundamental nature of a square, on the chessboard?
It should come as no surprise that the answer is INFORMATION (the nature of the square is to serve as a memory storage for G-UNITS).

G-UNITS = color (white, black, neutral, traitor?, null?, etc) + type (King, Queen, Grasshopper, Hole?) + features (Kamikaze, Chameleon, etc).
note: you can not assign this G-UNITS to a square, because the square might hold multiple units.
Plus, sometimes we want to disambiguate the information storage from the board square (e.g., some circe forms can be described by a position containing hidden squares, not on the board).
We require some hidden variable... maybe define the SQUBYTE (again, ask yourself: what would Hans Kmoch do?).

The SQUBYTE would contain all G-UNIT information.
The SQUARE would contain (among other things) a set of SQUBYTES.

By other things, I mean that a square would be assigned some set of properties, such as:
1) a unique ID / location (e.g., e4),
2) a display state (e.g., hidden or visible),
3) a name (e.g., "e4"),
4) a shade (e.g., light square or dark square),
5) it might have information pertaining to its directional interrelation with other squares (as defined by the BOARD),
6) it might have information pertaining to its promotional features (as defined by the BOARD),
7) it might have information pertaining to its castling features (as defined by the BOARD),
8) it might have some defined features (as defined by the set of all RULES),
9) etc.

The BOARD would contain (among other things) a set of SQUARES.
The POSITION would contain (among other things) a set of BOARDS.
etc.

It would be nice, if you could define a fundamental unit of information, and from that alone (plus some properties), store the entire POSITION (including castling information, en passant information, 50-move information, repetition information, etc)...

Essentially, we would like to specify the general G-CHESS GAME/PROBLEM, at least for some general class, based upon "DETERMINATE" rules, with "FORMAL" stipulations.
note: I have rigorous definitions (and justifications) for the two terms above, but it is involved, and highly complicated -- I will not presently define them... for complex matters, I await a legitimate arbiter!
It should come as no surprise that this process is nearly identical to how any programmer would logically specify objects (or subdivide a task).
The difference is, we just need a spec (not even pseudo-code) -- the developer needs to make it run, and test it.

So, next time somebody says this can not be done, just tell them to stop using their solving tool.
It has been done, many times over -- we need only get WFCC involved, so that everybody is on the same page (solving tools, databases, journals, albums -- everybody who wants to play ALL of the games of football).
 
   
(Read Only)pid=12860
(125) Posted by Hauke Reddmann [Saturday, Oct 18, 2014 16:41]

@Kevin: Interestingly, the approach to explain all
phenomena from the "fundamental bit of information"
is exactly a dream (pipe dream?) of modern physics.

Since Chess (ortho at least) is a finite game, we could
sum up all allowed moves, all rules, all phenomena as
one big state transition matrix. (Works also with all
finite fairy variants as long as their definitions
are clear.)
The Nalimov tables do more or less this approach, and the
disadvantage for the human is obvious - he can't understand
what's going on. Humans want to "chunk".

tldr: I'm half of a mathematician myself, and thus I think
Kevins Grand Unification Dream is viable. But I fear
only full mathematicians would enjoy it.

Hauke
 
 
(Read Only)pid=12861
(126) Posted by Nikola Predrag [Saturday, Oct 18, 2014 17:02]

You're certainly right about the square as a memory storage.
For some conditions, it seems better to use a concept of Royal-squares than to focus only at the Royal-pieces.

As long as check has a direct influence on the legality, it belongs to the rules. If the legality could be determined without using the concept of check, then "check" would not belong to the rules. I don't claim neither of these possibilities.

This stands also for the capture.
 
   
(Read Only)pid=12862
(127) Posted by Kevin Begley [Sunday, Oct 19, 2014 00:01]; edited by Kevin Begley [14-10-19]

@Nikola,

Think of G-Chess as a state machine.
You may want to think of it as information-states (maybe G-UNITS on G-SQUARES), plus some algorithm which controls state changes (G-RULES).
And, you might want to think that G-RULES must contain checks, captures, etc.

But, this is not the most fundamental reduction of a state machine -- the more general reduction would collapse the algorithm (the G-RULES) down to information states.
CHECK is not generally-definable -- you can not escape the fact that it is a defined state of position.
Thus, any G-Chess would need to collapse check (and all G-rules) to defined states of the position.
Eliminate the presumption of an algorithm, and focus only on the information states.

By the way, it is possible to start from the most general case, and presume problems which have:
1) a G-position,
2) some glossary (mapping unit images to G-unit meanings),
3) a (formalized) G-stipulation,
4) a set of G-rules,
5) are DETERMINATE (long story),
6) etc.

This would likely cover more than 95% of published chess problems, and it could continually reach to generally formalize the alternative classes of chess problems.

And, beyond that, it will allow us to establish some proof that defined states are completely determinate, based upon the information states.
There is a class of chess problems in which the defined checkmate state (and likely others) can not be determined, from the information states of the position.
This requires an involved discussion -- but, all is not lost: we could specify a few very simple rules (pertaining to the collapsed information) which would constitute a subclass of problems, wherein this "failure" is completely eliminated.
Eventually, the information from a fully collapsed state machine might enable us to increase our understanding of what causes indeterminate states, and we can expand this subclass.

By the way, (if memory serves) Nicolas Dupont tried to convince me, some years ago, that a general form of chess should be established, beginning from a base-class (the most general form of the most fundamental elements), and progressing in a completely logical manner, toward greater complexity.
I had this same idea myself (many years prior to Nicolas expressing it -- but, I never expressed it).
By the time Nicolas was advocating this, I had decided that a proper encoding would instead tie all fairy elements to an orthodox base class (then seek to minimize usage of fairy elements).

For reasons I can hardly justify now, I was arguing for an encoding like "Morse code" (where some orthodox base would be the most common letter).
Increasingly, I am of the opinion that Nicolas had it right.
 
   
(Read Only)pid=12865
(128) Posted by Nikola Predrag [Sunday, Oct 19, 2014 01:13]

Hm... Kevin, I am not convinced that orthodox rules could be defined without check.
Pieces can move according to their properties as they wish, but if a King is checked, the legal moves are reduced to those which parry the check.
"...it is a defined state of position..." OK, but what is that, if not a rule?
 
 
(Read Only)pid=12867
(129) Posted by Kevin Begley [Sunday, Oct 19, 2014 01:50]

@Hauke,

>Interestingly, the approach to explain all phenomena from the "fundamental bit of information" is exactly a dream (pipe dream?) of modern physics.

I agree -- today's physics, and computer programming, and engineering (and ,to some degree, even mathematics) are all collapsing toward the "science of information."
Information is conserved.

Even metaphysics (which is ever eager to distort science, for their own purpose) has tried to get in on the act.
The difference is: metaphysics is a non-analytic construct, based purely upon defined terms (and rarely can they actually define anything).
Sadly, the same may be said about the modern theory of chess problems.

We would like to pretend that we can simply define terms (like PIN, SPIKE, etc) -- in a non-rigorous manner -- and expect that truth can be simply determined.
As I was telling Georgy, you can define a variable (p) to be a prime number, or you can rigorously construct a mathematical definition of primes, such that the variable (p) must be prime.
The difference is, the former will not prevent mistakes -- the latter would be objectively determinate.

When "Take & Make" was invented, I insisted this was an "anticirce form."
Several people wanted to insist that it is not, simply because it was not defined as such.
That is their present model -- all truth is derived from definitions, and contradictions can always be swept away by appending exceptions (which will subsequently be ignored).

The difference is, I had a rigorous definition for the anticirce form, and "Take & Make" met all of the criteria.
In fact, it proved impossible for anyone (even me -- and I actually tried!) to provide a general (logical) definition for the form, which would strictly exclude this new invention (while including all members which were accepted by definition).

The non-rigorous model is a complete failure.
I can define a variable (p) to be a prime, by definition, but unless I can enforce the definition of prime, somebody will eventually come along and issue a definition (like a fatwa!) which wrecks the previous definition.

For example, some fairy inventor might insist that the number one (1) is prime.
Hey, it is not evenly divisible by any whole number except itself -- and according to some translated definition, that is proof enough!
So, everybody comes to accept it (without proof -- the same way we learn to accept that C=2*pi*r, and A=pi*r^2).
Years later, somebody (suffering from a slightly better definition) stumbles onto the proof that two (2) can no longer be considered a prime number, because it is evenly divisible by a prime -- namely one (1).
What does problem chess do?
No problem -- we have tools for that: add an exception which return primes to the meaning that we have all come to accept (but, do not bother to fix the number one as non-prime, because by now, we believe that is true).
And, on it goes...

The point is clear: the definition model suffers constant erosion -- typically, from years playing the game of telephone, within the journals (which have no room for rigor).
Eventually, the terms become completely undefined, and completely meaningless (and even our titled judges/masters/delegates are reduced to spoonfeedings of "goo goo ga ga").

Eventually, nobody can define anything -- and, you even find people parroting the hopeless belief that specific definitions are impossible (because hey, if this were possible, somebody else would have already done it for them).
All attempts to correct their misunderstanding is like reasoning with an infant -- until the hopelessness of a babied community has blanketed the world.

Anybody with a mathematical background will understand that there is a more stable model (for the low, low price of rigor, and eternal vigilance, we could ween ourselves from the spoon).
Of course, pure mathematics has its share of problems (the irrationals were never properly resolved, there may exist axiomatic truths which might be completely indeterminate, etc).
And, for some, the rigor may seem like quantum physics (or any subject which employs Greek symbols).

Ironically, those same people who claim this is too hard (or impossible) will expect popeye to solve every kind of problem (in full agreement with win chloe, alybadix, etc), in seconds.

Truthfully, you could hardly find a better location to observe the tragic-comedy of bad logic (because here there is a profound incongruity: this community depends entirely upon logic).





>Since Chess (ortho at least) is a finite game, we could sum up all allowed moves, all rules, all phenomena as one big state transition matrix. (Works also with all finite fairy variants as long as their definitions are clear.) The Nalimov tables do more or less this approach, and the disadvantage for the human is obvious - he can't understand what's going on. Humans want to "chunk".

When I finished (thoroughly) reading my first chess book ("Pandolfini's Endgame Course"), I could checkmate with KBN vs k.
Back then, it took me a ridiculous amount of time -- but, for some odd reason, I was determined to understand things from such a perspective.
Today, automated tools can teach a beginner an algorithmic pattern, in a small fraction of the time.

I also knew something about KQ vs k, but I did not have the bigger chunks.
By the time I cared to build these chunks, the understanding (and thus the teaching of this elementary endgame) had considerably improved.

This can be a be a question of philosophical outlook -- I chose to believe all the chunks will eventually clump; and, I backstop this with another philosophy: should we fail, at the very least, this outlook will have produced some reason for our failure. And, this all rests upon a third turtle: HOPE (I chose to hope).


>...I'm half of a mathematician myself, and thus I think Kevins Grand Unification Dream is viable.
>But I fear only full mathematicians would enjoy it.


I do not agree -- everyone PROFITS from advancement (just as non-mathematicians profit from the advances of math/science/tech).
People will naturally orbit the biggest spoon you place in their sky.
And, if we merely ask them to contribute to a community effort, they will help in ways that can not be imagined.
 
   
(Read Only)pid=12868
(130) Posted by Kevin Begley [Sunday, Oct 19, 2014 02:06]; edited by Kevin Begley [14-10-19]

@Nikola,

> Hm... Kevin, I am not convinced that orthodox rules could be defined without check.

First, I am not concerned only with orthodox; but, orthodox is definitely a important subclass.
The truth is reversed: check is a defined state of the position, and you can not define it as a static rule (it is a variable unit of information).

>Pieces can move according to their properties as they wish, but if a King is checked, the legal moves are reduced to those which parry the check.

Legal moves are also a defined state (rooted in information of position, and information of movement patterns).
If you can determine the set of legal moves (orthodox or otherwise) from the STATE INFORMATION, then CERTAINLY, you can determine the state of CHECK/CHECKMATE.

CHECK is a much simple state -- which may be defined to depend only upon a very specific set of state information (e.g., the position itself).
By definition, then, the full state of a POSITION must include all information necessary to determine CHECK.

Check is the most elementary defined state of a position (and it may be variable) -- CHECKMATE is a derivative (from CHECK, and LEGAL MOVES).
Think of it as the answer to a YES/NO question, but the question varies (we may define this question).

Legal moves must also be determinate (perhaps even computable!) from a specific set of information (which we may similarly confine to some object -- like POSITION).


>"...it is a defined state of position..." OK, but what is that, if not a rule?

Yes, it is a rule -- but it is transformed into a state of pure information.
It becomes a fundamental concept, which can be physically constructed (of light, of matter, of energy, etc) -- and it becomes subject to logical analysis.
 
 
(Read Only)pid=12869
(131) Posted by Dupont Nicolas [Sunday, Oct 19, 2014 02:12]

Yes Kevin, I indeed trust in such a possibility, modulo an important fact: some conditions should be left outside a workable logical system, as their rules are not logical by themselves – or at least are using so exotic kinds of logic that they can’t fit into an easy-minded coherent theory, based on a few basic axioms and clear-cut definitions (like e.g. in math).

This is the main difficulty – I don’t think many of us are ready to left or to modify the rules of existing conditions, in order to put some unity in our fairy jungle. I already provided a couple of examples, here is another one for those who missed them:

Just put bQh8 and wKd1 on the board and let black playing Qa1. In Monochromatic chess, it is not a checking move, and this is a good rule, mainly because Qa1 doesn’t observe Kd1 (the 2 involved squares are of opposite colors). But in Maximummer it is a checking move! This is not a good rule (to me) for the same reason - Qa1 doesn’t observe Kd1 (the next black move will always be Qh8, the unique legal move whatever white is playing before).

The possibility of constructing a consistent and useful logical system that will embrace those two kinds of checking notions is an illusion! To my eyes the great challenge is here: accepting that fairy chess is on a bad way (mainly because the lack of a good formalization at the very beginning), and deciding to remedy…
 
   
(Read Only)pid=12870
(132) Posted by Kevin Begley [Sunday, Oct 19, 2014 02:17]

@Nicolas,

I could not agree more -- and I could not express it better.

I think we need to define a class of elementary objects, from which ALL DEFINED STATES (e.g., legal moves, check, checkmate, etc) and ALL DEFINED PROPERTIES (e.g., capture, annihilation, en passant capture, castling move, promotion move, etc) can prove both DETERMINATE, and COMPUTABLE, for the information contained within these objects.

Clearly, POSITION is one such object.
Where there is a class of rules wherein checkmate might not be determinate from a position object, we should (for now) exclude consideration of this class (and attempt to carefully bound it).
 
 
(Read Only)pid=12871
(133) Posted by Nikola Predrag [Sunday, Oct 19, 2014 11:56]

Nicolas,
Maximummer, Reflex, hs#, h#, s#, Seriesmovers etc. are the heterodox stipulations but not the fairy conditions. The rules about legal moves are orthodox.
The stipulation just determines an additional specific goal for the play. The primary goal in Maximummer is the longest move, but according to the orthodox rules.

Seriesmover in n moves, starts with a move that creates a legal twin position which can be stipulated as Seriesmover in n-1 moves, and so on.
Castling convention requires some adjustments, but there's nothing fairy in the basic rules, it's only a heterodox stipulation.
 
   
(Read Only)pid=12872
(134) Posted by Neal Turner [Sunday, Oct 19, 2014 15:29]; edited by Neal Turner [14-10-19]

@Nikola
You certainly put your head on the block with your last post!
There won't be much argument that "hs#, h#, s# are the heterodox stipulations but not the fairy conditions" - but including Maximummer, Reflex & Seriesmovers in the same category is asking for trouble.

It comes down to what we mean by 'fairy chess'.
I would say first we need a baseline, and for this I'm going with orthodox Chess.
So we say that 'fairy chess' comprises those chess variants which differ from orthodox Chess.
This in itself would exclude Maximummer, Reflex & Seriesmovers from the realm of the orthodox, but let's try to be more specific.

Now I'll put my own head on the block and try to spell out how fairy chess differs from orthodox chess.
It comes in two parts - It's fairy chess if:
1) we can make moves that we can't make in orthodox chess.
2) we can't make moves that we could make in orthodox chess.

The three conditions you quoted all fall into the fairy category on the basis of the the second part.
Maximummer: Black can't make moves shorter than the longest move.
Reflex: If a mating move is possible no other move can be made.
Seriesmovers: One side can't move at all!
The fact that the actual moves made may be orthodox is irrelevant.
 
   
(Read Only)pid=12874
(135) Posted by Dupont Nicolas [Sunday, Oct 19, 2014 15:52]

Nikola,

I don’t reject that Maximummer has its own kind of logic (although I feel it dubious). What I want to point out is that this “condition” (or whatever we are calling it) is very near from Monochromatic – it would also have been possible to consider this later as a “heterodox stipulation” (keeping that checks remain normal).

Thus the difficulty is that the 2 above mentioned “conditions” are of the same kind (the range of legal moves of a single unit on the board is strongly decreased), but are not governed by the same checking rules… So unity is missing between those 2 neighbors.
 
   
(Read Only)pid=12875
(136) Posted by Nikola Predrag [Sunday, Oct 19, 2014 20:44]

Neal,
#n means: the solution is any play that fulfills the stipulation, any other play is simply not the solution.
The rules of Maximummer and Reflex are orthodox, but the solution is only such play which completely fulfills the stipulation.
There is a specifically defined partial help-play.

Ser-#n: find a legal white move from a diagram position (twin 1) to achieve a legal twin-position 2 with White on the move, such that twin 2 can fulfil a stipulation ser-#n-1. Ultimate goal is achieving a twin position n, stipulated as #1.
One-move stipulations certainly mean that only one side moves, ser-h#n ends with the n-th twin h#1 where both sides play one move.

I find it absurd to consider the Chess composition just as an imitation of a chess game. There are chess rules and I can say that I'll mate you without moving my Queen from d1. If I move it, I lose the bet. The rules are not altered, only the deal is specific. Rules would be altered if e.g. bKd2 would not be considered as attacked by wQd1 just because I promised not to use wQ.

Stipulation is a kind of "deal" between a composer and a solver, rules don't change but the deal determines which play will be considered as the solution and which not.

Nicolas,
"...(the range of legal moves of a single unit on the board is strongly decreased), but are not governed by the same checking rules…"
If any detail of the orthodox rules is changed, it would mean a fairy condition. Monochrome would not be fairy if no detail of the ortho-rules would be changed.
Strict-Maximummer, or whatever the name, would make the logic unity with "Strict-Monochrome". The only trouble is in naming, Monochrome is strict (and thus fairy) by default, while it's the opposite in case of Maximummer.
 
   
(Read Only)pid=12877
(137) Posted by Dupont Nicolas [Sunday, Oct 19, 2014 22:23]

I understand the distinction you made between a heterodox stipulation and a fairy condition. It remains that I can't see any coherent logic while deciding that Monochromatic is fairy and Maximummer is "only" heterodox. What can be the reason, except an artificial choice from the respective inventors? Btw, both Popeye and WinChloe classify Maximummer as a condition, not as a stipulation.
 
   
(Read Only)pid=12878
(138) Posted by Sarah Hornecker [Monday, Oct 20, 2014 04:07]

For me, it is all a venture into the depths of the jungles of the faerie land, where dreams come true and bluebirds fly over the rainbow.
And it is all heterodox that is not an orthodox directmate or study.

I see not why we hold onto this anyway. Why is a composition different of another, if only of their stipulation? How many roads must a man travel down before you can call him a man? How many problems does it take to see all of the same world, only different regions. The wide beautiful jungles and steppes of faerie land, or the green grasses of orthodox land, those shall not be divided but as one.
 
 
(Read Only)pid=12879
(139) Posted by Kevin Begley [Monday, Oct 20, 2014 05:57]; edited by Kevin Begley [14-10-20]

Neal describes "how fairy chess differs from orthodox chess."
If there is to be a definition for "fairy chess," then it must differ from a "base-game" (orthodox is not a credible term -- it has no definition, and is subject to constant evolution).

Let me first say that Neal is probably wrong -- as I once was -- to prefer a base-game which is styled upon the present (undefined) perception of "orthodox chess."
The base-game is meant to define some ROOT set of rules, from which we may describe any chess game/problem, and to which any game/problem may reference.
I made the mistake of presuming that the base-game would serve as a constant default reference, when in fact, it is only the trunk of the tree.
The evolutionary nature of the "orthodox" game strongly suggests that the popularity of these references will likely shift (to branches we can not yet begin to imagine).
Further, popularity of reference is not the only matter; there are other matters which probably deserve a more primary consideration, such as:
1) we may prefer a logical root for the base-game (enabling smoother development, promoting easier understanding, reducing redundant derivatives, etc), and
2) we may prefer a base-game which provides us with a means to fairly account for the ECONOMY OF RULES (which might enable our judges to make wiser evaluations).

That said, what I really want to address is Neal's defined separation of fairies (from the base class):


>"...It's fairy chess if:
>1) we can make moves that we can't make in orthodox chess.
>2) we can't make moves that we could make in orthodox chess."


The good news: this is a big improvement upon Nikola's suggested definition (which represented a demonstrably false attempt to draw a distinction between fairy and orthodox terms).
The bad news: Neal does not quite survive the chopping block -- the above definition is demonstrably false.
The assessment: Neal's definition contains a rather minor mistake, which can easily be remedied.

Two RULE-BOOKS (for games or problems) may be demonstrated to be distinctly different, even if they share a completely equivalent set of legal moves (for all possible positions), if the interpretation of any defined GAME-STATE can be demonstrated to vary.

First, consider the most trivial case of this distinction: an identical GAME-SPACE, with non-identical outcomes.

One game (or problem) may award a half of one point for stalemate; another game (or problem) awards only a third of one point for stalemate.
The distinction here might be deemed completely irrelevant within the subset of formally-stipulated chess problems (or, it may merely require a precise definition of WIN/DRAW aims).
However, from a game perspective, different outcomes may constitute an important consideration in comparing the equivalence of rule-books.

The larger problem with Neal's definition is: even within the subset of formally-stipulated problems, a theoretical failure is evident.

Checkmate is a defined STATE, and it is not inconceivable to define this according to specific exceptions.

For example (off the top of my head, and for purely intended clarity in illustration), one may defined stalemate to include any checkmate delivered by a pawn.

According to Neal's definition, even for chess problems with "formal" stipulations, this would appear equivalent to the base-game (it would share all moves, for all positions), despite the fact that solutions for an identical problem (or the outcome of an identical game) may distinctly differ!

Therefore, I think it critical to define equivalence differently.

I think the rule-book must define some algorithm which is guaranteed to deterime & compute ALL GAME-STATES (the set of legal moves, Check, Checkmate, Stalemate, etc), from a given POSITION-STATE.

And, the POSITION-STATE must contain a basis of information necessary to guarantee that:
a) the rule-book's algorithm will be sufficient to determine & compute all of the defined GAME-STATES, and
b) any possible position may be completely expressed.

note: this includes castling info (KQkq), en passant info, 50-move count info, repetition info (perhaps the move history), all retrograde presumptions (e.g., castling presumed legal unless proven otherwise), rebirth information (e.g., bQ captured on c6, awaiting rebirth), etc -- ALL info necessary to determine every possible GAME-STATE defined (legal moves, check, etc).
Also note, you need not define checkmate, if the problem/game ends in capture.

Then, presuming somebody manages to define the term "ORTHODOX," then the meaning of "HETERODOX" would be immediately clear:
1) a necessary difference in the set of defined GAME-STATES, or
2) a demonstrable difference in any GAME-STATE, for any given POSITION-STATE.
 
   
(Read Only)pid=12880
(140) Posted by Georgy Evseev [Monday, Oct 20, 2014 09:04]

Some short comments concerning the last turn of discussion.

1. "Programmer's approach" to chess definitions is wrong. In programming defining something means making it to show needed behavior. In real life defining means emphasizing existing specific properties. These two meanings of "defining" are in fact quite different.

2. I do not believe it is possible to have a stable and solid definitions for fairy realm. A soon as you consider your work finished, some people will test the boundaries and propose new pieces/conditions/stipulations that will require updating your concepts. So I think there is some sense is leaving everything as it is. Orthodox problems based upon unchanged rules of game of chess which has not changed for at least 200 years (do not mix them will tournament chess rules, which may be taken into account or ignored - 50-moves rule, dead reckoning, etc).

So the orthodox chess become the eye of typhoon, the island of stability around which we have a slightly ordered chaos of fairy chess. So this chaos will naturally resist to attempts of excessive ordering.

3. We _can_ live without strict definition of pin, because the pin is not a part of the rules. If you use pin, spike, paralysis, nail or goo goo ga ga as a technical feature or artful effect, it does not really matter, how it called if called at all. A descriptive explanation should be enough.

4. There are to known kinds of Maximummer. In a more traditional kind, one first creates a set of legal moves and then only longest of them are valid. In another kind (called Exact Maximummer or Ultra Maximummmer) we first select the longest moves and then check if there legal ones between them. I am sure that thoughtful search will show many kinds of non-commutative conditions (for instance, Take&Make and Anti Take&Make were used to show similar difference).
 
   
(Read Only)pid=12881

Read more...
Page: [Previous] [Next] 1 2 3 4 5 6 7 8 9 10

MatPlus.Net Forum General What is a "line pin" and are there undefined pins?