MatPlus.Net

 Website founded by
Milan Velimirović
in 2006

8:47 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 Retro/Math A proof game question
 
You can only view this page!
Page: [Previous] [Next] 1 2 3
(41) Posted by Andrew Buchanan [Saturday, Oct 1, 2022 04:03]

-1 Seetharaman for his ad hominem :(

(Surprised he found my contribution particularly long. It's easy to pose short questions, or superficial non-answers - it takes longer to tease apart these wretched issues. Thanks to others for diligently working forwards (with longer replies) while I was silent.)

Moving on...

(1) Dealing with the .0/.5 distraction

There is no particular merit in "standards across genres". Rightly, I see no campaign to replace "#2" with "d#1.5". In #n, s#n, h#n etc, the identity of the final mover is always known. A decimal digit in h# indicates who makes the *first* move.

In a PG on the other hand, the first mover is certain. The .0 or .5 instead indicates the *last* player, but to me there's no reason why this digit is any more binding than the integer *before* the decimal point.

(2) We know we do not have universal agreement. So what do we actually do? We have to be charitable.

* LCD: The lowest common denominator is that "PG in x.y" is interpreted as "PG in exactly x.y". Otherwise old problems are unfairly designated as cooked.

* THE OLD: In databases alongside the original stipulation, we may possibly over time add a canonical stipulation: "SPG in x.y" or "PG in exactly x.y", but it's not really necessary. If an old PG admits a shorter solution, it may be fun to add a new version, but the original remains correct (unless the original composer wants to change it).

* THE NEW: Composers of new PGs would be advised to avoid ambiguity. I always verify that no short solution of -0.5, -1.0 or -1.5 exists. Usually quicker than the main C+ run, this is sufficient to guarantee SPG status. I think if someone has taken the trouble to do this, and chooses to publish a problem for which a shorter solution exists, they would always put in "exact". Because it's obvious that otherwise discord may result.
 
   
(Read Only)pid=23888
(42) Posted by seetharaman kalyan [Saturday, Oct 1, 2022 04:54]

Obviously i am not the expert in PGs. Dummies like me can only ask questions. When experts refuse to answer and write 30 or 50 lines, what should one do? 🤔 😢

In fact, I didn't ask for standatds across genres.

Added later: There is no system of decimal digits in S#.
 
 
(Read Only)pid=23889
(43) Posted by shankar ram [Saturday, Oct 1, 2022 05:18]

Deleted
 
   
(Read Only)pid=23890
(44) Posted by seetharaman kalyan [Saturday, Oct 1, 2022 05:36]

That is great Shankar. You are free to stir the hornet nest.
 
   
(Read Only)pid=23891
(45) Posted by Andrew Buchanan [Saturday, Oct 1, 2022 06:01]

You don't get this kind of rude behaviour in the Discord problem group, even though most are much younger. Please behave yourself, Seetharaman
 
   
(Read Only)pid=23892
(46) Posted by seetharaman kalyan [Saturday, Oct 1, 2022 07:10]

Calling me and Shankar as ignorant of PG and asking us politely to shut up --- and then deleting it. Is it not insulting? Do I need a certificate from you before I can even ask a question? You are the one who talked about hornet's nest and Shankar's ignorance.

Till now I had been very polite only. Merely commented on the length of posts by many including you.

Writing long lines doesn't make anybody an expert and entitle them to call others ignorant.
 
   
(Read Only)pid=23893
(47) Posted by Kevin Begley [Sunday, Oct 2, 2022 00:48]

Everyone's opinion is welcomed here.

@Andrew,

 QUOTE 

There is no particular merit in "standards across genres".


If you are suggesting that a more universal stipulation standard would have no benefit, I respectfully disagree. Vehemently. That is simply incorrect (and it ignores historical precedent).

Problem enthusiasts have already made one costly shift to a more universal stipulation standard.
As you are no doubt aware, not all directmate threemovers you see in the problem databases were originally stipulated as "#3".
Plenty had to be read (in their original language): "mate in three", "black mates in three ", "white to move and force mate in three, but not two", etc.
The benefits of this more universal stipulation standard are undeniable (in most cases, enthusiasts will understand the stipulation without any need to translate Chinese, Russian, etc).

For many years, I edited a problem/puzzle database (for a robot player) on the Internet Chess Club.
I have some firsthand idea what a time consuming enterprise must have been required to edit every previous problem (to obtain a more universal standard), but I constantly invested time toward that end, because I well understood that the benefits would far outweigh the cost.

You might also take note that the original stipulations are not preserved in the databases -- only their meaning is preserved (to conform to our more universal standard).
So, you can hardly argue that the original meaning of the "PGn" must be preserved at all costs.
The only thing that must be preserved is the original intent of the author (and that is ALWAYS BEST expressed in a more universal stipulation standard).

Our stipulation standard today is could be more universal. Some argue it is badly in need of repair (others may consider it broken).
Even if you think it functional, you can not deny that it could be improved, and improvements would accrue some undeniable benefits.
This goes without saying: the more easily newcomers are equipped to understand the stipulation in an unfamiliar genre, the more likely they are to feel invited.

All that said, I will readily concede the following:
1) the cost of establishing a more universal stipulation standard is FAR more onerous today (due to a dramatic expansion in problem stipulations, problem quantities, and problem software).
Such a major undertaking would tower over all previous standardization work, and the upheaval can not be overstated. We are fast approaching a state where our standards are "too big to fix".

2) maintaining a more universal standard is a never ending process, with never ending costs (future generations will always discover a new need for further refinements).

3) the benefits accrued from a more universal stipulation standard will yield definite benefits, but there are diminishing returns for every major revolution of the stipulation standard.

We can hardly afford to update our stipulation standard every century. This should only be done when the benefits far outweigh the injuries caused by the present system.
I would argue the present standard is clearly broken (as evidenced by the fact we can't even establish a correct classification primer, and we can't agree what is meant by "PGn").
Even so, I can't deny there remains a lingering question: how much longer can we hold off?

It's like counting the census every ten years -- if you're going to do it, if you're going to pay for it, you want to be sure you don't get it wrong (and it holds up to future scrutiny).
To do this right, would require a team of courageous revolutionaries (timid people need not apply -- this requires people unafraid to scrape away ALL of the bad precedent we have accumulated).
You can't abide people who defend bad precedent for the sake of precedent (in such an endeavor, that attitude amounts to nihilism).
You need to establish unambiguous definitions for every element of a standard chess problem, and imagine stipulations that are not yet in common use.
Brevity will not suffice -- to establish the most simplified universal standard, you need a team capable of producing a tome (purely for internal use).

 QUOTE 

Rightly, I see no campaign to replace "#2" with "d#1.5".


Direct is a Mode of Play in a chess problem (which specifies the side not tasked to achieve the aim will resist, and the side tasked to achieve the aim will be aware of this resistance).
The "d" in directmate is implied by default -- same goes for the "d" in direct-capture problems (xn), and the "d" in direct-check problems (+n), and all other problem stipulations which provide an aim and a number of moves but no mode of play.
Why would anyone campaign to change the standard default?

As for the move numbering -- yes, that will eventually change (it will become universal, whether you like it or not).
The only question is: are you willing to oversee that change, or do you prefer to leave that for others?

 QUOTE 

In #n, s#n, h#n etc, the identity of the final mover is always known. A decimal digit in h# indicates who makes the *first* move.


That's false. The standard for "#n" only defaults to "for white" (it can be stipulated -- commonly seen in twins -- that the task of achieving the stated aim is "for black").
What does it mean when we state "#n for black"?
Actually, you have to disambiguate two separate concepts (which are tragically not the same under our present -- broken! -- stipulation standard):
1) most fundamentally, the addition of "for black" means it is for black (not white -- which is the default) to realize the AIM (#) of the problem (since it is direct-play, white will seek to prevent the aim).
2) the addition of "for black" also defines who plays the last move (and this, we will see, has implications about what constitutes a dual).

To illustrate how poor is this term in our present standard, consider the addition of the word "for black" in a selfmate.
The intent, clearly, is to reverse which player must realize the aim. But black already plays the last move of the selfmate problem.
And, the "s#n" stipulation FAILS to provide an efficient mechanism to convey whether (or not) multiple black mates should be considered duals.
White is tasked to achieve the aim in a s#n -- so why doesn't the problem stop when white has achieved a state wherein black has no moves but mate?
If you said, "the selfmate problem goes on to mate because of the dual convention," you'd be correct.
And if you are beginning to recognize that the composer must be allowed to specify when the problem ends (what should be considered a dual), congratulations -- you're seeing just the start of what benefits a more universal problem stipulation standard could provide.

A more universal stipulation standard would express exactly what the author is asking, in the most efficient way possible, in a manner which is maximally consistent across all problem genres.
And the old problems will be translated in the most charitable way possible -- just as ancient problems had their stipulations updated to "#3" -- to conform to our updated standard.
The only difference: you will have no doubt what was intended by the author (and if you can read one problem genre, there's a good chance you can read them all).

 QUOTE 

(2) We know we do not have universal agreement. So what do we actually do? We have to be charitable.

* LCD: The lowest common denominator is that "PG in x.y" is interpreted as "PG in exactly x.y". Otherwise old problems are unfairly designated as cooked.


To whom are you being charitable?
Yes, we should be charitable when translating old problems into newer standards. However, it's uncharitable to the future to impose those old stipulations.

Is it unfair that old problems which stipulated "white to move mates in seven moves, but not six" was designated as unreadable?
If the original stipulation must be maintained, lots of luck scouring every library on Earth, and maintaining a search function (and a solving tool capable of reading Chinese).

If you want to be charitable to the future, you standardize as many problems as you can, in the most concise, most universal way possible.
You can do this, and be MORE charitable to the old problem composers (who do not want to be left behind, who want their intent clearly stated).

The intended meaning for the VAST majority of all "PGn" problems (composed yesterday and today) is not "PG in exactly n moves" (the VAST majority crafted their problem with the intent that any shorter solution constitutes a cook, and you can not call it charity if you are going to decimate their intent, because you want to be charitable to those problems which failed to note the distinction and the precedent).
We must be charitable to both sets of composers -- which means we clearly CAN NO LONGER abide today's broken standard.
For those few (very, very few) which intended that the word "exactly" should be assumed, we move to a better standard and preserve their original stipulations (the most charitable interpretation of what they had intended), so their problems are not left behind.

"PGn" and "PG in exactly n" are two completely different stipulations -- they CAN NOT be expressed by one stipulation.
That's not charitable to the vast majority of PG composers (who deserve a stipulation standard which clearly conveys their intent), it's not charitable to the solver (who must guess what the composer's intended for cook possibilities), and it's not charitable to the database users (who may want to search exclusively for "PG in exactly n" problems, but have no means to distinguish between the two stipulations).

The charitable standard is indefensible as a standard.
This was known by Tibor Orbán, way back in 1976, when he created the first "PGn" problem which had unwanted solutions in fewer than n moves.
Orbán was well aware his problem required a departure from the standard stipulation (and the standard default, for all problem genres, that a shorter solution should be considered a cook, unless expressly stated otherwise) -- that's why he expressed his stipulation using the word "exactly."

That's now the general standard for other genres, as well.
A universal standard would not only provide a generalized way to express the elements or a problem, it would also be shared by all problem genres, and the elements would be expressed as efficiently as possible:

Elements such as:
1) Who moves first? White? Black? Either player? Figure it out by retro-analysis?
2) When is the aim achieved (when does the problem end)?
3) Who can realize the aim? White? Black? Either player?
4) Who can move last? White? Black? Either player?
5) What constitutes a cook (or a dual)?
6) Can the problem be solved faster than the number of moves specified, or must it be exactly that number?
etc.

I have argued for a stipulation standard based on recursive goals, wherein today's standard s#n would contain an inner-goal (#1 for black -- essential to express where the problem ends, and what constitutes a dual), within the outer-goal (white forces black into a position having only legal moves which checkmate white).
This way, if you want to express that multiple mating moves are not part of the solution (and do not count as a dual), you expressly indicate that the inner stipulation is not part of the solution (and the problem stops after white's last move).

A dual may occur in any move required to meet a clearly stipulated goal (and no move that goes beyond the solution).
So, if you stipulate to play the inner-goal (#1 for black) in a selfmate, then that inner-goal is expected to be unique (unless expressly stated that black must have 2 solutions).

Think in terms of what is the core aim of the problem -- everything revolves around that (the mode of play, and the goal):
There are state aims we may ask a solver to achieve (#, =, ==, etc).
There are move feature aims we may ask a solver to achieve (+, x, ++, ep, roque, etc).
There are combinational aims we may ask a solver to achieve (++ and #, # or =, etc).

So, the goal is to achieve an aim in a given number of moves (with the above elements defined, or provided by a standard set of defaults).
But, recursion allows a sub-goal to be the aim of the larger goal.
In a selfmate, white forces a position wherein the sub-goal is a "#1 for black", and black resists the outer goal.
In a help-selfmate, you get more recursion (the selfmate becomes the inner goal, and in the outer goal, both sides help to achieve it).
Once you understand this, you can create a number of interesting stipulations you had never imagined (in fact, I used this to create a number of new stipulations, only to discover there were often problems composed which already used these stipulations; but those problems could not be stipulated in a standardized form, because our poor standard failed to imagine they could be expressed).

My recursive idea efficiently covers a vast number of stipulations (which were not considered standard previously) -- such as problems where the strategy changes -- e.g., black goes from helping white achieve the aim to defending against the aim at some point.
I would need to produce a very large book to demonstrate the greater coverage this would provide (and how the number of characters in stipulations can actually decrease), but it covers many more stipulations that only the future can hope to imagine.

Best of all, it helps composers to identify new stipulations they had never imagined.
With a more universal standard, you vastly expand the possibilities of our artform.
To dare suggest this is not beneficial amounts to gross negligence.

Someday, I may write a very long article about it. Even knowing that most problemists today will refuse to accept it, I suspect future problemists may find it highly profitable, when they're ready to put all our bad practices behind them.

It's difficult to describe a fishbowl to those who are living in it. Nowhere is this more tragically absurd than in problem chess.
Problemists have meticulously classified every conceivable type of interference pattern, yet they can not provide even a functional definition distinguishing a stipulation from a fairy condition.
If they might benefit in the award from reciting defined elements of a standard chess problem, as they do from reciting the defined theme, we'd witness a more universal stipulation standard in days.

How does this impact proofgames? Like any other standard chess problem, Proofgames can be expressed as an aim, a mode of play, a number of moves, and the standard set of elements.
A proofgame is just an A->B problem with a default starting position (wherein white starts, by default).
But, how do you specify an A->B problem wherein black starts (for example, in Progressive Chess, where parity may be of concern)?
We needn't ponder such questions, if we had a universal stipulation standard. There would be no doubt what is stipulated (even for those unfamiliar with a given genre).

The aim of an A->B problem (or PGn) is to achieve the diagram position (let's call it "dia"), but we fail to explicitly specify who is tasked to achieve the aim (either anyone can achieve it, or it must be achieved by a specific side, according to the number of moves provided). This is a bad practice. The stipulation should be clear. It should default to white seeks to achieve the aim, and explicitly state when black must achieve the aim (or allow that either side may achieve the aim).

Generally, the mode of play in an A->B problem (or PGn) is HELP-play (h-dia-n), but in highly constrained conditions/diagrams, it may be direct- (dia-n), or self- (s-dia-n).
Under the Charitable Standard you advocate, how would you stipulate a help-self-proofgame?
You'd probably use words ("white to move and blah blah blah").
With a universal stipulation standard, you'd simply stipulate the mode of play (hs-dia-n), as you always do, and everybody would understand the task (and all the elements, and all the default conventions).

That's how easy it is to translate PGn and A->Bn problems into a universal standard.
There's no reason to reinvent the wheel. The Proofgame can follow the universal stipulation standards (and these can be applied to all genres of standard chess problems).
We need only inflate the tires now and then.

Today, we are riding on four flat spare tires. We need to remedy the gross failures in our present standard (make it more universal), and it starts by clearly defining the elements of a problem.
Damn the old notation -- just like past generations decided to alter every problem which was stipulated by non-standardized words (e.g., "White to move and helpmate in three moves, but not two").
The original stipulation is not what we want to preserve. By standardizing more stipulations, we can better express the original stipulation that the author had intended.

When the old standard is not logical (or broken), we have a duty to admit its failures.
The charitable standard doesn't work. Nobody can defend it. If "PGn" means be charitable, then no solver should care about that stipulation (because it's not a clear expression of what's expected).
Give solvers one standard, which applies to everything, and they'll never need to ask what's a cook, or what's a dual.
Failure to provide a universal standard amounts to turf protection (it only inhibits newcomers from entering unfamiliar problem genres).

Most problemists think only of their own short-term self-interest.
Composers of Reflexmates (a fairy condition) want to be judged with orthodox selfmates.
Composers of Series-Movers (ditto) want to be judged in their own section (pretending, or fooled into thinking, they are pseudo-orthodox because their fairy condition is expressed as a stipulation).
Creators of new fairy conditions (like parry-series) want their invention to be deemed a stipulation (not a fairy condition).
When ambiguities are revealed in a fairy condition, and the condition must fragment into two interpretations, every composer wants the default version to be charitable to their compositions.
And composers don't want to add "exact" to their Proofgame, despite the fact that their stipulation is not the same as the standard proofgame (nobody wants some rare stipulation to exclude their problem from the larger group).

On and on it goes. Almost nobody asks, "what's best for the future of our beloved artform?"

Those who are really self-interested (and capable of taking the long view) will know they should be charitable to our artform, first and foremost.
The best way to ensure your problems will endure is to provide our future enthusiasts with better standards.
If we lose our future problem enthusiasts, we lose everything -- we want to ENCOURAGE participation in proofgames (by clarifying our stipulations, and adhering to more universal standards).

Your greatest charity should go to the future (newcomers), and only secondarily to preserving the original intent (in all future stipulation standards) so the works of past composers are not forgotten.
You can do both, simultaneously; but you can not do either if you fight to preserve the Charitable Standard (which is demonstrably broken and indefensible).

Because our standard is undeniably broken, we have a rare opportunity to forge a better, more universal standard -- across the board.
If we look ahead to what future stipulations (which may not even exist today) could be standardized, we can create a standard that will endure for a very long time (and will help us to open up to new stipulations we might never have dreamed possible).
More importantly, we can make problem stipulations more readily understood by a wider audience.

But, we will continue to achieve nothing if we remain constantly reduced to the LOWEST COMMON DENOMINATOR (do nothing, avoid upheaval at all costs, fight for short-term self-interest, etc).
How many new adherents has problem chess gained? Would you say it's going up? Would you say it's healthy?
My guess: things could be much better, if we could focus on the beauty of chess problems, rather than constantly fighting the war of self-interests (to preserve our failed standards, and deny there's a problem).

If our bad standards discourage future enthusiasts, every problem we ever made may be forgotten. Imagine every great work of chess problem art lost forever, because we didn't want to provide an update.
And if, many thousands of years later, some group of new enthusiasts should stumble upon an old database of proofgames, do you really imagine they will feel obligated to extend charity to those who refused to properly stipulate the task they had intended?
Do you think they would extend their patience to learn our divergent standards (based on genre turf)?
If we can't provide them a good standard, why should they keep reading? That's game over. We owe them a reason to keep reading, and that starts with the way we define aims, stipulations, fairy conditions -- all the elements of a standard problem should be crisply explained (so they never need ask, or entertain the slightest doubt that they might not ALREADY KNOW, what is a cook, what is a fairy condition, etc).

I imagine they'd think very little of us, based upon our present standards; and even less should they encounter all our stupid efforts to preserve a broken system (should future prospects ever stumble upon this forum after a prolonged period of inactivity, they would be appalled by our adamant refusal to intelligently fashion our stipulations -- when the solution is elementary obvious).
They might think more highly of us if we provide them an opportunity to solve some of our problems, but that would require us to provide them a clear stipulation (and we'd need to stop asking charity from them).

If we build a universal stipulation standard, they will come.
 
 
(Read Only)pid=23911

No more posts
Page: [Previous] [Next] 1 2 3

MatPlus.Net Forum Retro/Math A proof game question