Difference in quality between CORR games and Practical chess

Discussion of anything and everything relating to chess playing software and machines.

Moderator: Ras

budfit

Difference in quality between CORR games and Practical chess

Post by budfit »

To all chess community:
Recently we had a very long discussion on chess database logics and what quality means for different kind of chess players. The post started innocently with posting 52,000 games for free on A00 - Irregular Openings and informing chess community where to download it. (you can still find at http://www.openingmaster.com )

One of the last topics raised by James Constance was talking about deduplication process but also the inclusion of ICCF / Correspondence games into main stream database.

http://www.talkchess.com/forum/viewtopi ... 88&t=23640

Our answer and also basis for perhaps another discussion is below. We would welcome anybody who has his/her opinion on this subject to post the answer here. Appreciated.

A)
We start with simplification of two bodies (and again this is just simplification we respect very much other chess associations as well)
1) Practical chess games under FIDE
2) Correspondence chess games under ICCF

The last ''conflict'' between FIDE and ICCF was when group of top players from ICCF wanted to play on FIDE chess Olympics. The answer was NO, you have your own Olympics. Let's see why. There are players like Ulf Anderson who are good Corr. chess players but also practical players. There was no will for confrontation. Secondly, practical chess as we all know is played under totally different conditions. The players who played both forms can best understand this. I used to play some time ago the practical chess myself for about 5 years, for todays FIDE rating I would say something between 2000-2200 ELO. Then I entered the Corr chess world due to several reasons.

What is the difference between Corr chess (ICCF) and Practical Chess (FIDE)?

1) It's all about which players play the game. Implicitly it goes hand in hand with quality of the game. If we can compare ELO of players who play both types of competitions , the ICCF players have 100-200 lower score compared to those from FIDE.
2) Then it is about the external factors influencing the practical chess such as stress or other indispositions. The corr chess player unlike his colleague from practical chess, can use all the tools available - books, databases, computers engines etc... Therefore the results of ICCF players are lower in reality.

And why am I talking about this?
The ChessBase first started to divide players into two groups. It was due to underestimation of corr.chess games from the past. This schizophrenia is kept until now. However today it doesn't have any more meaning as play under the stress is one thing (including making stress mistakes) and deep analysis of chess moves is another (when we talk about corr chess players we mean those with higher ELO)
The future speaks in favor to correspondence chess only because of distance and costs associated to moving between point A and B. This is also a silent point of conflict.

To conclude : Virtual Chess or Practical F2F chess. Each has it's pros and cons. We cannot accept the dogma of the past that only practical chess games are real games where many officials these days try to present it this way.

B)
Answer to James:
Each chess database builder considers what to insert into his/her chess database under certain common rules. We think, correspondence chess games should be included because they don't talk about two chesses but one. Actually in the practical chess games one may find more mistakes than in corr chess which is perfectly ok, but nobody discuss about it. Therefore the ratio of infiltration of correspondence chess games into the main OM database is higher. Mainly we talk about the quality games or insufficiently present openings such as A00. Therefore we say now, yes, we have more correspondence chess games included in the ''main'' database but this is mainly due to increasing level of the quality and not about some irrelevant criteria set up by ''officials''.
Therefore if a chess player wants to order the database which should support his plays he or she can better decide based on his own criteria.

Best Regards,
Alexander Horvath, SIM ICCF
http://www.openingmaster.com
User avatar
Rolf
Posts: 6081
Joined: Fri Mar 10, 2006 11:14 pm
Location: Munster, Nuremberg, Princeton

Re: Difference in quality between CORR games and Practical c

Post by Rolf »

There is a nice paradoxon, that although corr games have fewer "mistakes" their value and quality is much less than
human chess played games. Also because corr games are partially
created with computer help... Nevertheless also for corr databases it's necessary to give a good header with Elo and all informations. It's true that Ulf Andersson is also a good OTB player, but the other way round it has no value. Normal corr players have no strength OTB.
-Popper and Lakatos are good but I'm stuck on Leibowitz
BubbaTough
Posts: 1154
Joined: Fri Jun 23, 2006 5:18 am

Re: Difference in quality between CORR games and Practical c

Post by BubbaTough »

Howdy,

I for one love the idea of a chess database with some degree of rating regularity that includes recent chess, older (pre-rating) chess, and correspondence chess. Whether the ratings in these three groups are perfectly weighted or not is not terribly important, as long as I can do a search criteria (such as >2500) that ensures only high quality games. When I occasionally play chess I like to be able to look for games of a certain quality (and decrease it if I don't find enough games) and combining these three groups lets me do that in a more comprehensive way. The site I use currently: http://www.chesslab.com/PositionSearch.html does not allow that (on the other hand it is free, which is nice).

I look forward to trying out the free portion of your site (A00) if/when I get the chance...even if I have to ignore a few quasi-duplicate games :wink: .

-Sam
James Constance
Posts: 358
Joined: Wed Mar 08, 2006 8:36 pm
Location: UK

Re: Difference in quality between CORR games and Practical c

Post by James Constance »

Alexander

I have nothing against including correspondence games in the main database for quality reasons. I had incorrectly assumed that you follow the practise of chessbase and chess assistant where the main database doesn't include the correspondence games.

Best

James
budfit

Re: Difference in quality between CORR games and Practical c

Post by budfit »

Hi James,

I responded to your original post in the other section. Follow : http://www.talkchess.com/forum/viewtopi ... 31&t=23640 to see the response.
Alexander
I have nothing against including correspondence games in the main database for quality reasons. I had incorrectly assumed that you follow the practise of chessbase and chess assistant where the main database doesn't include the correspondence games.
Best
James
Just to correct your statement. It is not right that CB and CA do not include the correspondence chess plays in their main database. They do. They also offer a separate database called Correspondence Database but this is also offered under the Opening Master as separate.

Best Regards,
Alexander Horvath
http://www.openingmaster.com
User avatar
Rolf
Posts: 6081
Joined: Fri Mar 10, 2006 11:14 pm
Location: Munster, Nuremberg, Princeton

Re: Difference in quality between CORR games and Practical c

Post by Rolf »

budfit wrote:Hi James,

I responded to your original post in the other section. Follow : http://www.talkchess.com/forum/viewtopi ... 31&t=23640 to see the response.
Alexander
I have nothing against including correspondence games in the main database for quality reasons. I had incorrectly assumed that you follow the practise of chessbase and chess assistant where the main database doesn't include the correspondence games.
Best
James
Just to correct your statement. It is not right that CB and CA do not include the correspondence chess plays in their main database. They do. They also offer a separate database called Correspondence Database but this is also offered under the Opening Master as separate.

Best Regards,
Alexander Horvath
http://www.openingmaster.com
If I didnt know that Schweijk would come from Prague, I would call you now the big Schweijk of the Universe of all databases. Rather late I begin to realise with how much humor and irony you tell us why we should tgake all the two million doubles in your OM as a present of the internet and not as disturbing. Also after you destroyed the bad data core so that the tables of the tournaments can no longer be decrypted, I begin to realise how much surplus freedom I win if I use your OM as a practical player. Now I can again play my master opponent without having to create always new and endless play records of their career. Finally chess has come to its roots back to coffee house. Also the whole Elo nonsense was always a more disturbing pain in my *** because I could always see how low my own Elo was comparfed to the others. Now everything begins at ZERO again. This is also a true democratic achievement from your side.

Long live Schweijk. The German and Slovakian Schweijk!

I would ask for 100€ as a fair prize in the light of the many democratizations. The ballast from CB was since long a stick in my sane eyes. Now I can see the blue sky again.

Schweijk regards.
-Popper and Lakatos are good but I'm stuck on Leibowitz
dufekj

Re: Difference in quality between CORR games and Practical c

Post by dufekj »

Dear Rolf and Alexander,

maybe biggest difference between both is only about using (OM) database.
As i read, the main problem for Rolf is "... without having to create always new and endless play records of their career.
Also the whole Elo nonsense was always a more disturbing pain in my *** because I could always see how low my own Elo was comparfed to the others" and, lost but not least doubles.

Firstly - questin for Rolf - it's more improtant for you perfect collection games of your opponents (with perfect name, ELO etc) but with less (and very often with much less) games,
or time to time you can little work with names (unifying names, for example), ignoring ELO etc but with huge amount accessible games? I preffer second, because games are reason why i using CB, enhanced informations for example ELO
hasn't any influence on played games. Of course, it will be very nice if games included right information. (today probably nobody use old Weir's cbnormal.exe)

Secondly - war with doubles is never ending and you can never win (only Chuck Norris can draw). With non-commented games is situation easier, because there is not problem with different
comments and you can compare clearly only body: now simple question, are games in followed example doubles (same tournament, and year, but how much changed, if they were played in different tournaments and years)?
game 1 Player A - Player B 1.e4 e5 2.Nf3 Nc6 3.Bb5 a6 4.Ba4 Nf6 5.0-0 Be7 6.Re1 b5 7.Bb3 d6 8.c3 0-0 9.h3 Bb7 10.d4 Re8 11.Ng5 Rf8 12.Nf3 draw
game 2 Player C - Player D 1.e4 e5 2.Nf3 Nc6 3.Bb5 a6 4.Ba4 Nf6 5.0-0 Be7 6.Re1 b5 7.Bb3 d6 8.c3 0-0 9.h3 Bb7 10.d4 Re8 11.Ng5 Rf8 12.Nf3 draw
game 3 Player A - Player C 1.e4 e5 2.Nf3 Nc6 3.Bb5 a6 4.Ba4 Nf6 5.0-0 Be7 6.Re1 b5 7.Bb3 d6 8.c3 0-0 9.h3 Bb7 10.d4 Re8 11.Ng5 Rf8 12.Nf3 Re8 draw
game 4 Player A - Player D 1.e4 e5 2.Nf3 Nc6 3.Bb5 a6 4.Ba4 Nf6 5.0-0 Be7 6.Re1 b5 7.Bb3 d6 8.c3 0-0 9.h3 Bb7 10.d4 Re8 11.Ng5 Rf8 12.Nf3 draw
game 5 Player A - Player B 1.e4 e5 2.Nf3 Nc6 3.Bb5 a6 4.Ba4 Nf6 5.0-0 Be7 6.Re1 b5 7.Bb3 d6 8.c3 0-0 9.h3 Bb7 10.d4 Re8 11.Ng5 Rf8 12.Nf3 Re8 draw

what games from this sample are doubles?
If you use databse for preparing against concrete player and you will delete game 2, then you lost information about this game and you never complete tournament table etc
User avatar
Rolf
Posts: 6081
Joined: Fri Mar 10, 2006 11:14 pm
Location: Munster, Nuremberg, Princeton

Re: Difference in quality between CORR games and Practical c

Post by Rolf »

dufekj wrote: Firstly - questin for Rolf - it's more improtant for you perfect collection games of your opponents (with perfect name, ELO etc) but with less (and very often with much less) games,
or time to time you can little work with names (unifying names, for example), ignoring ELO etc but with huge amount accessible games? I preffer second, because games are reason why i using CB, enhanced informations for example ELO
hasn't any influence on played games. Of course, it will be very nice if games included right information. (today probably nobody use old Weir's cbnormal.exe)

Secondly - war with doubles is never ending and you can never win
Dear Jiri,
now I can smell a different and fresh wind coming into our debate. Let me shortly mention that I wrote the quoted stuff only for irony purposes because in the meantime I have understood Alexander well enough. It was irony but my true opinion is different and exactly the contrary of what I wrote.

Basically you brought now all the relevant aspects in short summary.

Let me answer all this in serious.

- if I were a book author I would wish a database with Elo so that I could justify the typical computerchess decision that I have a high quality book without coffeehouse game scores

- if I were a dedicated player in competitions I would like to have Elo, also no move game scores, byes etc, asll that to know in advance everything possible about the tournament behavior of my opponents

- if I were a journalist of chess and expert player formerly or actually, then I would want o get everything like the dedicated player but also very detailed info about the tournament classes, team event details, exact board numbers and all this, at best some expert commentaries

- if I were a beginner I wished to begin my career with many many details and not just a sort of computer standard like move generator what is possible in whatg position with what evals; that should all be integrated in my chess programs features, like training the different phases of a game, but one thing is clear, I didnt need the low Elo output from people out of third national leagues because as a true player such variations should also come to my mind in my own homework, and these millions of games would rather disturb my attention, but I knew as a talented and smart beginner that in the Himalaya team championships there were some 10 billions games played every weekends but I just must decide to ignore it because already theree million game scores are too many, so that for me the claim from Alexander with even MORE and MORE games would only confuse me

- as a beginner in creating databases on commercial bases I would primarily be interested in the feedback from as many people as possible. I would discard anything beforehand because in special on the internet I cant know for how many long years people might have input their heart blood into the mere collecting aspect of chess gamescores. Of course this is much less important for master players in chess.

- my personal feedback as honest as possible stands still in my several messages to Alexander, but afterwards I got more and more the impression that he failed badly to reinforce the efforts of his correspondents because he had asked good questions in a topic that is seldom discussed. Also his business idea is fairly legal. But if you then cant handle the unanimous criticism of too many doubles, and you begin then to lose attention between accusing the critics and linking the weaknesses in the bases to the fatally bad reality on the internet THEN you lose all seriousness, which is fatal for a businessman. You cant have it all at once from possible the same persons. Some can give you deep critics and you cant ignore this just by insinuating that such a specific critic does only want to nag you. Then you are making a serious psychological mistake. Because the clue is to take frfom everybody tghe best for your own ideas and goals but then to likewise give all the people the impression that you are very thankful especially for their input. However if you are a ittle bit undereducated yourself you might be blindfor the class of some messages or points of your correspondents. I know for sure that my very first example where the base in OM neglected ther fact tht it was a rapid game that this was a crucial mistake. But Alexander was incapable of getting this as a matter of fact and he already showed me that he's not that talented for such exchanges. To say the least. The pontificating speech gave it a nice haut gout, Actually I think, Al is a bit overworked IMO. But his limits were already showing right from the start. Because he behaved against his own and apparent goal to make some friends and possibly new cooperators. Definitely he lacks all the talents Vasik has in his speech - to motivate everybody with his different talents.



Hope this helps.
-Popper and Lakatos are good but I'm stuck on Leibowitz
budfit

Re: Difference in quality between CORR games and Practical c

Post by budfit »

First, thanks Jiri for your post your appearance was just-in-time...

And now to our one and only reporter Prof. Rolf. Somehow you changed rhetorics, and we really wonder why? And we were just getting used to your tone. Perhaps you learnt who Jiri Dufek is before replying or perhaps he gave you enough final facts which opened your eyes. Those same statements which we provided 25 times already but which you haven't been able to see until now.
It was irony but my true opinion is different and exactly the contrary of what I wrote.
Do you mean in all your previous posts?
Also his business idea is fairly legal. But if you then cant handle the unanimous criticism of too many doubles, and you begin then to lose attention between accusing the critics and linking the weaknesses in the bases to the fatally bad reality on the internet THEN you lose all seriousness, which is fatal for a businessman
Our statement is unchanged. We answered every single post with professional manner without accusing somebody of something. I hope you are not linking word ''accusing'' to ''please provide more details'', or "which type of analysis have you used when you get your results'' . These were our only accuses we have made through out all posts. Which is quite funny to say now, seeing your ''big reply'' to Jiri.
You finally understood there are two different types of methods (as well as chess players using data)
1) those who prefer perfect headers
2) those who prefer to find information for his game

we chose 2.

Best regards,
Alexander Horvath, SIM ICCF
http://www.openingmaster.com
James Constance
Posts: 358
Joined: Wed Mar 08, 2006 8:36 pm
Location: UK

Re: Difference in quality between CORR games and Practical c

Post by James Constance »

budfit wrote:Hi James,

I responded to your original post in the other section. Follow : http://www.talkchess.com/forum/viewtopi ... 31&t=23640 to see the response.
Alexander
I have nothing against including correspondence games in the main database for quality reasons. I had incorrectly assumed that you follow the practise of chessbase and chess assistant where the main database doesn't include the correspondence games.
Best
James
Just to correct your statement. It is not right that CB and CA do not include the correspondence chess plays in their main database. They do. They also offer a separate database called Correspondence Database but this is also offered under the Opening Master as separate.

Best Regards,
Alexander Horvath
http://www.openingmaster.com
Alexander

As far as I am aware the chessbase correspondence database is not included in Megabase, and the Chess Assistant correspondence database is not included in Hugebase. They are entirely separate databases. A quick search confirms this.

James