MatPlus.Net

 Website founded by
Milan Velimirović
in 2006

10:31 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 Album Threshold Mysteries
 
You can only view this page!
Page: [Previous] [Next] 1 2 3
(1) Posted by Administrator [Thursday, Jun 21, 2007 15:10]

Album Threshold Mysteries


This thread is for practical reasons compiled from posts in another thread. The paragraph which initiated the topic was:
 QUOTE 
Hans Gruber (June 16th, 2007):
Some recent postings mentioned a case - the name is not important, I suggest - who deteriorates the problems he is copying from. The most serious "problem" related to such a case is NOT his plagiarism. It is much more serious that editors nevertheless continue to publish such rubbish. And that relates to a threat (it's a different forum, I know) for the FIDE-Albums. Having to judge thousands of H#2, of which most are of lousy quality, is much more threatening problem chess than the plagiarism of bad work. (Nowadays, with computer use, it is much easier to get a - lousy - H#2 sound! Therefore, many more entries reach editors for publicaton. The rejection rate did not increase, however.)

 
(Read Only)pid=950
(2) Posted by Thomas Maeder [Saturday, Jun 16, 2007 18:37]

The recent PCCC decision to split up the helpmate section into two and to apply the percent rules to each of the new sections separately has done anything but helping this problem. One of the results of the decision is that a minimal number of helpmate twomovers will make it into the Album.

 
 
(Read Only)pid=951
(3) Posted by Harry Fougiaxis [Saturday, Jun 16, 2007 23:55]

 QUOTE 
The recent PCCC decision to split up the helpmate section into two and to apply the percent rules to each of the new sections separately has done anything but helping this problem. One of the results of the decision is that a minimal number of helpmate twomovers will make it into the Album.

I have not understood, the split (and the assignment of two separate directors in the recent album) is necessary for practical reasons, isn't it?

Perhaps I'm missing something here, but how is the number of selected problems affected? In fact, some of the percentages given on the page http://www.phenix-echecs.fr/articles/FIDE/AlbumFIDE_9800/AF1998-2000.html seem strange to me. For h#s is 9.32% and for retros 27.73%. The instructions that I received for FA 2001-03 (dated August 2001) mention that

Normally, all problems with a total score of 8 points or more go into the Album. If less than 12% of the entries in a section are selected for the Album, the limit is changed so that all problems with 7.5 points or more go into the Album.

Is this not applicable any more? Furthermore, I had the impression that if the percentage is greater than 20%, then the limit goes up to 8.5 points. This is not mentioned in the instructions, but I can clearly remember it was discussed (decided?) in some early meeting, long time ago when the evaluation system changed. I searched in the PCCC site, but I couldn't find anything in the available minutes.

 
 
(Read Only)pid=952
(4) Posted by Hans Gruber [Sunday, Jun 17, 2007 01:17]

Hi all,

the rule to increase the level to 8.5 points, if more than 20% are included, was abandoned some albums ago. It usually applied only to the retro sections - probably because it is much more difficult in this section to compose rubbish (and send it to the Album tourney). The main reason might be that computers cannot easily be used for "composing".

Hans Gruber
 
 
(Read Only)pid=953
(5) Posted by Thomas Maeder [Wednesday, Jun 20, 2007 22:10]

 QUOTE 
I have not understood, the split (and the assignment of two separate directors in the recent album) is necessary for practical reasons, isn't it?


As you may know, I would have prefered the solution discussed last year: that problems would be rated by the judges in the order of preference indicate by their authors, with an immediate stop if a certain number of problems by that author are below some mark (say 1.5).

But you are right, the split is an alternative measure to reduce the amount of work for helpmate judges, and it will obviously work.


What I meant is the following: These days, the average h#2 is anticipated (I'm exaggerating a bit here, but really just a bit) or not very interesting; that includes the average h#2 submitted for the Album. As a consequence, the average mark of a h#2 is lower than that of a h#>2.

Before the split, 12% of all submitted helpmates may or may not get >=8 points; it's not clear whether the problems (in particular, the h#2) with 7.5 points make it into the Album.

After the split, 12% of the h#2 will be much less likely to get >=8 points because of the lower average quality of the h#2; the 12% rule will be applied more probably to the h#2 than it was applied to all helpmates before the split. As a result, the h#2 with 7.5 points are more likely to make it into the Album after the split than they were before the split.

 
   
(Read Only)pid=954
(6) Posted by Harry Fougiaxis [Wednesday, Jun 20, 2007 23:15]

Now, I understand. I am also certain that the limit for h#2 will be at 7.5 points, but personally I am not particularly worried, these will be good problems anyway, only marginally lower than the "true" album quality. It is inevitable, in a sense, for the split into two sections is mandatory nowadays; you had the dreadful experience of having to evaluate 2570 problems for 1998-2000, didn't you?

The total number was almost the same for 2001-03 (1150 h#2 and 1400 h#>2). Considering that 179 h#n were selected for 2001-03, i.e. a percentage of 12.8%, it is quite logical to assume that even with the limit lowered to 7.5, the selected h#2s will be eventually less than 12%. This is quite normal, I should say.

What is still a mystery to me is the percentage shown for the 1998-2000 album. How come it is only 9.32%?!
 
 
(Read Only)pid=955
(7) Posted by Uri Avner [Thursday, Jun 21, 2007 02:08]

I'm planning a revision in the calculation of the cut-off point for Album admission in any given section.
The idea is to calculate the % needed, taking into account only those entries that got the mark of 5.5 and above from all judges. A preliminary statistical analysis is needed to arrive at the right %.
This would most likely solve the inherent problems in the H#2 section, where most of the rubbish is sent, thus artificially making 7.5 the needed score for the present system.
And thanks, Thomas, for supporting my proposal from Wageningen. A version of this might be proposed again in Rhodes which I hope will get more recognition.
(But what all this is doing in the "Plagiators" section??)
 
 
(Read Only)pid=956
(8) Posted by Hans Gruber [Thursday, Jun 21, 2007 17:07]

Hi Thomas & all,

one of the current H# judges mentioned that he found THOUSANDS (really, THOUSANDS) of anticipations concerning H# entries to the Album. Quite discouraging, and I am afraid that you were not at all exaggerating ...

Hans Gruber
 
   
(Read Only)pid=957
(9) Posted by Thomas Maeder [Thursday, Jun 21, 2007 18:49]; edited by Thomas Maeder [07-06-23]

 QUOTE 
HF: you had the dreadful experience of having to evaluate 2570 problems for 1998-2000, didn't you?

Yes. Well, I volunteered for the job, and I knew in advance that there would be many. I didn't expect quit as a many as there turned out to be, of course.


 QUOTE 
HF: What is still a mystery to me is the percentage shown for the 1998-2000 album. How come it is only 9.32%?!

All three judges were very severe; accumulating 2570 problems for judging is a very reliable method to get severe judges. And I was the most severe of us, so feel free to blame me. :-)

 QUOTE 
UA: I'm planning a revision in the calculation of the cut-off point for Album admission in any given section.

My favorite revision would be to abolish the 12% rule entirely, because it establishes an aritifical connection between the number of problems submitted and the number that should make it into the Album.

 
   
(Read Only)pid=958
(10) Posted by Harry Fougiaxis [Thursday, Jun 21, 2007 21:29]; edited by Harry Fougiaxis [07-06-25]

 QUOTE 
All three judges were very severe; accumulating 2570 problems for judging is a very reliable method to get severe judges. And I was the most severe of us, so feel free to blame me. :-)

Well, no offense intended of course, but frankly speaking such a percentage even after allowing the 7.5-graded seems too low to me.
 
 
(Read Only)pid=959
(11) Posted by Uri Avner [Friday, Jun 22, 2007 19:54]

A quote from you, Thomas:
"My favorite revision would be to abolish the 12% rule entirely, because it establishes an aritifical connection between the number of problems submitted and the number that should make it into the Album."
Please read my note again and see that exactly this problem is expected to be solved by the planned revision, without, on the other hand, ignoring the possible bias that may easily be the sole creation of the judges and not just the amount of rubbish.

 
   
(Read Only)pid=965
(12) Posted by Yochanan Afek [Saturday, Jun 23, 2007 01:15]

I proposed once that a problem selected for the Annex Albums (the "almost" ones) will receive half an Album point. Any supporters?
Beyond this really facsinating discussion there is just one little matter that really bothers me: Where is the Album? It's June 2007 for god's sake!
 
   
(Read Only)pid=966
(13) Posted by Thomas Maeder [Saturday, Jun 23, 2007 09:13]

 QUOTE 
I proposed once that a problem selected for the Annex Albums (the "almost" ones) will receive half an Album point. Any supporters?

If I remember correctly, your motivation is to allow titles to be gained more quickly by those who deserve them.

The problem is that your proposal would also make it easier to gain titles for those who (by the current rules) don't deserve them. Titles will thus lose some of their value, which I think should be avoided.

I would welcome solutions which speed up the title gaining process while not devaluating titles; I am not sure that they exist, though.

 
   
(Read Only)pid=967
(14) Posted by Yochanan Afek [Saturday, Jun 23, 2007 10:43]; edited by Yochanan Afek [07-06-23]

Indeed, my proposal is aimed at slightly accelerating the process of awarding titles for those who do fully deserve them, which perhaps might enable them to enjoy this recognition a bit earlier before quitting this world.
Mind you,In all other domains(Bridge, Draughts, over the board chess and ...problem solving!)this title is awarded when the player is still at the peak of his form,rather than as a retroactive "Nobel prize" for lifetime achievements.
 
   
(Read Only)pid=968
(15) Posted by Sarah Hornecker [Saturday, Jun 23, 2007 14:58]; edited by Sarah Hornecker [07-06-23]

@Yochanan Afek:

Then my opinion should come in handy.

 QUOTE 
I think, titles should be awarded for certain successes. There should be titles for every form of composition so that no grandmaster of studies can be considered grandmaster of fairy chess. Next, titles should be awarded by the importance of a person. It is very hard to understand why for example the Platov brothers should not gain a title.


By the importance thing it is clear this can turn out at most composers only when they produced enough. There should be rules for importance but it's clear T.R.Dawson or the Platov Brothers were important enough. Why not honoring them post-mortem after all?

And furthermore...

 QUOTE 
I believe there should stay three titles. These should be National Master, International Master and Grandmaster. For National Master it should be enough to win either five tourneys or reach one of the first three ranks of a WCCT. For International Master it should be enough to win fifteen tourneys or reach one of the first two ranks of three WCCT. For grandmaster it should be enough to win twenty-five tourneys or win three WCCT. There should be points for ranks to make things equal. So there may be awarded ten points for first prize in a tourney, six points for second prize etc. while WCCT should count ten times those points. There must be tables worked out etc.

Or we can entirely skip this stuff with titles. There were many great composers who never gained a title. I see no need to award titles at the moment.


I still agree with myself. Of course the numbers are only an example (but winning 300 points with this system may be easier or harder than with FIDE titles as they are now. After all the FIDE is not a reputable organization so composition shouldn't get named with it. They started to destroy OTB chess. Should we wait until they destroy composition as well?

And yes, I'd like to say more about the FIDE but I don't want to express it all in public.


PS: Of course, grandmaster should not be too easy to win. It should hold it's value.
 
   
(Read Only)pid=969
(16) Posted by Juraj Lörinc [Saturday, Jun 23, 2007 15:23]

I concur to Thomas' opinion that making titles easier to achieve would cause them lose their value. Rather it is better have them as a rare lifetime achievement with some value than to have them as achievement at the peek of career without value.
 
   
(Read Only)pid=970
(17) Posted by Uri Avner [Saturday, Jun 23, 2007 18:02]

The "original sin" lies in our compulsive desire to mimic the otb chess players and have in our field the same titles they do. This has political reasons as well.

But in reality, Art and Titles do not mix well together. Is it just a coincidence that similar titles are not officially granted to creators in any major field of art (literature, painting, music, etc.)?

No wonder that in our somewhat enigmatic domain of chess composition it is the source of so many unsolvable absurd situations. At least it helps to verify our status as Art (as opposed to Sport) where achievements cannot be simply expressed by numbers.

The trouble is that once these titles are introduced there's no apparent way to force them back into the bottle...
 
   
(Read Only)pid=973
(18) Posted by Yochanan Afek [Sunday, Jun 24, 2007 02:31]; edited by Yochanan Afek [07-06-24]

Are the titles the only difference? Does any other art organize so many tourneys and championships? who is the world champion for operas? or aquarels? what is this comparison? Every congress is composed of numerous competitive events. So why won't we simply admit that we are not just pure artists but rather competitive characters who love titles and prizes and other awards.The more the better and never enough! There are only very few exceptions to that, I believe.
 
   
(Read Only)pid=975
(19) Posted by Hans Gruber [Sunday, Jun 24, 2007 18:52]

A nice idea by Uri, to force the spirit of titles back into the bottle ... are there other reasons than vanity why this seems to be impossible? What is the use of title in chess composing? (Are there still countries in which titleholders can improve the quality of their life? If yes, then it is worth to maintain titles and even to make the gaining of titles MUCH easier!) Certainly, titles are not required in order to know that somehow is able to compose excellent studies or problems. To recognise that Rudenko, Rehm, Caillaud, etc. are great composers is easy if you look at their compositions (if you do not intend to do so, you are somehow wrong in our small world). Irrespective of their titles. It is not difficult either to recognise that composers like Feather or Millour are great composers, if you look at their compositions. They are not among the group of titleholders (I believe).

Probably I miss one (or more) reasons why titles are important. I would like to learn more about this.

Hans Gruber
 
   
(Read Only)pid=984
(20) Posted by Yochanan Afek [Monday, Jun 25, 2007 00:32]; edited by Yochanan Afek [07-06-25]

Dear Hans,

Likewise you may wonder what do we need composing tourneys and prizes for? Isn't it enough to look at a problem in order to recognize its greatness? However the average human creature likes a formal kind of recognition too. Just like in the academy. Titles are not important? I don't hear many title holders sharing this view.
 
   
(Read Only)pid=988

Read more...
Page: [Previous] [Next] 1 2 3

MatPlus.Net Forum General Album Threshold Mysteries