MatPlus.Net

 Website founded by
Milan Velimirović
in 2006

23:34 UTC
ISC 2020
 
  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-Jul-2020

B P C F





 
 
MatPlus.Net Forum General Composing & Computers
 
You can only view this page!
(1) Posted by Neal Turner [Wednesday, Oct 2, 2013 17:03]

Composing & Computers


So at Marjan's prompting I've started this thread.

I've been guilty of 'computer crimes' in past, but now I'm reformed - so here's my take for what it's worth:

I've noticed that nowadays many composers spend maybe 10% of their time at the board and 90% at the computer working on their problems.
But these people really are doing themselves a disservice, because it really is more satisfying - and more fun! - to try to get it right before presenting it to the silicon.

Of course the current predilection for helpmates is fueling this trend, as are modern fairy forms involving complex transformations which are virtually impossible to make without the computer.
But in the latter case composers would be advised to stick to short problems without extra fairy elements to give themselves a chance of getting them as correct as possible using their own brain-power.

Is it honest for a composer to ask solvers make the effort to solve, without a computer, a problem which he couldn't make the effort to compose without a computer?

(I'm not talking about computer generated/tablebase problems here, but just 'normal' composing)
 
(Read Only)pid=11008
(2) Posted by Steven Dowd [Wednesday, Oct 2, 2013 18:03]

Is this related to the discussion about - I don't have the issues readily available so I hope memory holds - "Peter Harris problems" in a recent issue of feenschach? Problems that used too many fairy elements or conditions and were essentially unsolvable by the human mind?

I personally stay away from much fairy chess because it simply is too complicated for me to properly understand. I might look at a fairy with one or two conditions or pieces every two or three months just to see if I can follow the logic and beauty of the solution. Solving it is of no importance to me.

It strikes me as a fruitless endeavor to bemoan the use of new tools. What was once postal chess will never be the same. In the 1980s I played postal in Gambit thematic tournaments, and these were not your usual gambits, but the kind of speculative nonsense like 1. d4 Nf6 2. g4? or my personal favorite of all time, the Elephant 1. e4 e5 2. Nf3 d5?? Nowadays local club tyros know how to refute these things.

You had people in those days argue that using opening books in postal games was an unfair advantage; today it is computers.

The last book I saw on the Blackmar Diemer was pitiful, with the author trying to avoid crushing lines that just refute White's strategy. You can feed the position into silicon and clearly see White is just a pawn behind.

But one of my original points was that why mention what a computer brings to a problem, at all (in a way that can only be interpreted negatively in my opinion), when first place is taken by an admittedly beautiful database problem? How the problem was generated is entirely irrelevant unless it is an issue for both problems. And yes, my obvious beef is that I was singled out.

What I find amusing is that a problem I did with Mirko was excluded from the award (and again, it may have been excluded with good reason - I have my own reasons for not liking the final product). Without Gustav, we would have taken five years to get it sound, and I don't think Gustav made it "easy." The cooks that popped up were beyond human comprehension and I kept notes on all of them and tested a new version each day based on what I thought would get rid of the cooks. This took several weeks. In the end, we might have just left the idea on the scrap heap.

The old method, where we would have sent the problem off and waited for a human to cook it, and then us try to fix - I agree this was a great way to make problems and increased the connection between composer and solver, but like my days of speculative gambits in postal chess, long gone.

Thanks to Neal for starting this new thread.
 
 
(Read Only)pid=11009
(3) Posted by Kevin Begley [Wednesday, Oct 2, 2013 23:26]; edited by Kevin Begley [13-10-03]

The first Babson #4 was made with the help of a computer.
And, more to the point, it was not made (for decades) without such help!
Same story with the first sound K vs K (1+1) proofgame.

The tools used to raise a building have no significant impact upon the artistic quality of the product -- better tools only hasten compliant construction.

For now, composing chess problems with computers remains no less an artistic endeavor (for humans).
The real difference today -- which nobody addresses -- is that the best tools are no longer universally obtained.
Some are not even commercially available -- they are specifically developed to confer some individual with advantage!
Maybe for the first time in history, a problem composer must work to develop their own tools, or suffer some (not insignificant) disadvantage.

I had hoped Michel Caillaud (in his recent online lecture covering this very subject) would share some of his wisdom on this matter.
But, as always in chess, it's too easy to shy away from the underlying controversy.

A determined composer (like Steven), with hard work, can still overcome this advantage (thanks to help from the available tools, like Gustav).
But, by comparison, this is entirely a low gear climb.

I would never dismiss such a climb as merely "something doable."
But, an objective judge can neither marvel at his traction and tenacity.
Their job is to consider only the beauty of the peak that he has reached (it matters not whether on the hoof, the horse, or the helicopter).

But, a helicopter is of no use without a pilot -- and, I would argue that pilot is always the human imagination.
Unfortunately, today's judges will discard virtually all consideration of the imaginative element of composition, so as to focus entirely upon the dull construction of some unoriginal alphabetic pattern.
It should be carefully noted: we require neither original problems, nor artistic beauty, if we aim to competitively measure constructional skill.

Either way, everybody wants to have a helicopter (fueled and ready).
And, I often wonder how soon fairy composers will begin developing their own set of tables, for specific pieces/conditions.
Maybe it has already started -- maybe developers simply refuse to divulge such tools (such revelations tend only to undermine the developer's best interests, anyway)!
If you think specialization is bad now, just imagine the future ... or wait for it to impact you.

For studies, there is almost universal access to the EGTB.
For proofgames, everybody has access to Natch.
But, not everybody has access to a selfmate table.
Not everybody has access to a series-helpmate table, or a madrasi table, or whatever table suits your preferred tastes.

Well hey, everybody knows how to make such a tablebase (if not, there are plenty of articles available on this).
It just takes a great deal of memory, a few months of cpu-time, and some hard work (especially verification/debugging).
But, it can not be ignored that the key to quality composition has shifted (I would suggest dramatically), toward computer programming.

I can understand why this transition doesn't sit well with some -- notably, for those who are foolish enough to dream of preserving the traditional means.
I submit that this "philosophy" is hostile to the process of creating enduring works.

The transition is inevitable (like it, or not).
Either curse Heraclitus for his constant flux, or accept his truth -- endlessly bemoaning an unpleasant reality is fruitless.
And worse, there are judges who will insist on awarding prizes based upon their desire to deny this inescapable reality.
This does significant harm to this art.
They don't offer any alternative aesthetic philosophy -- they only hide behind a false rationalization for systematic favoritism.

I submit that chess problem judging became the worst system in practice, intended to subjectively measure the quality of any art, the moment it buckled under the strain of computers.
Until problemists settle on some shared philosophy governing this new paradigm shift, our judgements will grow increasingly more controversial (if not flatly wrong).

For truly lasting problems, I see no alternative but to take the long view -- the sooner we embrace computer aided composition, the better it is for our art.
I consider modern composers to be no more than the first human to find beauty in a diagramed problem.
Constructional considerations are rare today -- when they do exist, they are often rendered beyond esoteric (read: a matter beyond any harsh consideration).
Ultimately, it is the human imagination which must triumph (over the formalistic realization of the unoriginal idea).

Unfortunately, modern judges are completely unprepared for such a paradigm shift.
They cling to the tradition of an obsolete, orthodox model.
And, this systematically denies today's composers of any hope for fair judgement.
In this environment, the more lasting your problem, the less likely it is to receive due recognition.
Such an atmosphere is not only likely to devalue the judgements of the day, it forebodes an enduring devaluation of the art.
 
   
(Read Only)pid=11010
(4) Posted by Neal Turner [Thursday, Oct 3, 2013 10:25]; edited by Neal Turner [13-10-03]

I can see now that we need to clarify more precisely what we're talking about.
At the risk of oversimplifying we can maybe recognise three types of composing methodology:

1) The human does the composing and uses the computer merely check for correctness.
2) The human uses the computer as an active partner in the composing process.
3) The human sets up the computer to generate positions/solutions which can then be selected according to his own aesthetic criteria. (Apologies to Torsten if this is too simplistic)

The points I was trying to make in my first post were:
- The best method is 1) - call me old-fashioned if you like!
- Many people are using 2) not out of necessity, but out of laziness.
- Modern fairy forms are so complex that it's impossible to use 1)

But the most important point was that we mustn't forget the solvers!
Our picture of solvers is coloured by a small number of super-solvers who compete in the solving ladders.
But I don't think I'm far out if I guess that 90% of solvers don't even attempt anything longer than 2 moves!
If we lose the solvers then we're dead in the water.
 
   
(Read Only)pid=11012
(5) Posted by Kevin Begley [Thursday, Oct 3, 2013 21:22]; edited by Kevin Begley [13-10-03]

@Neal,

You forgot the most important category:

0) The human does the composing, and no computer can provide any practical help (whatsoever)!

This still occurs frequently today -- even in the fairy forms, which are not nearly as complex as you would like to suggest.
Try some fairy retros/Proofgames, sometime.

Solvers would do much better to focus on problems beyond the computers.
Not only is this where the cooks are, it's where solvers are more desperately needed.
Furthermore, it's the only way to discourage dishonorable solvers.
And, the problems with computer-immunity offer the best insight into the real talents of a given composer (you will not discover such information buried in the album-points).

This frontier marks the true separation -- and if ever there were a legitimate reason to award an inequitable quanity of album points, per problem, this is the only criteria worthy of consideration.

And, I see no evidence how a diminishing pool of solvers (e.g., of long selfmates) can leave this art, in any way, "dead in the water."
I will concede that judges would do well to consider the problem's impact on the human (solver) audience.
This art is, after all, intended to be at least considered, by a human solver (composers who employ excess fairy elements have not only failed in the realm of economy)!
But, we need not solve every problem encountered -- it's possible to empathize with the solver's challenge.

ps: All of your categories (1, 2, and 3) are effectively identical.
The only difference is, the savvy composers practicing your later options will less frequently need to ask the computer for verification (it's baked into the process).
The only real distinction is problems which can not be verified -- anything that could be made by computer, has lost all boast of a low-tech realization.
Anybody who tells you otherwise is lying to protect a fragile ego.
 
 
(Read Only)pid=11013

No more posts


MatPlus.Net Forum General Composing & Computers