ο»Ώ Chess ratings discussions

   		Free Internet Chess Games Server

Install FICGS apps
play chess online


Game result  (chess)


B. Ozen, 2269
N. Wilson, 2121

1/2-1/2

See game 152323




 Hot news
 Discussions
 Files search
 Social network



American Go Association

                                          
Forum


Back to forum


Thibault de Vassal    (2006-07-01 17:44:41)
Format For Championship

Hello John.

After all, if it wasn't unusual, the interest would be lower for sure... :)

So you noticed, the 8 players with the highest established correspondence chess ratings play a pure knockout tournament.

I thought about this format a long time ago (and a long time). Combining a knockout tournament (more "spectacular") and a round-robin cycle (everyone can play, no more than 5 cycles) gather together the advantages of both. It is one of the reasons I made FICGS... I think pure knockout or pure round-robin wch cycle is not efficient enough for chess championships.

The other thing you'll notice in the rules : "The special rule is that in case of equality (4-4), the winner is the player with the strongest tournament entry rating if all games are draw, the player with the lowest tournament entry rating if not all games are draw. The winner is qualified for the next stage."

This rule (in case of equality in the round-robin tournaments, the player with the strongest TER is qualified too) is another way to avoid short draws... It may sound strange at a first sight, but I really think it's fair enough and a good way to find most probably the really strongest players in the last stages. Anyway, it's amazing for sure :)


The FICGS chess wch rules :

http://www.ficgs.com/membership.html#tournament


Thibault de Vassal    (2007-01-21 14:31:00)
Chess ratings

Hello to all.

I have been submitted this case :

"Rating system: Look at this situation: Current rating ot the player: 2200, 10 games, all players with TER 2200, result 9=, 1+ Case 1: Finished games: 01.01-28.02: 9= --> New rating 01.03: 2200 01.03., no other finished games in this period: --> New rating 01.05: 2228 Case 2: Finished games 01.01-28.02: 9=, 1+ --> New rating 01.03: 2200 --> New rating 01.03. 2214 In case 1 the value of the won game was 28, in case two it was 14. IMO the value of a result should not depend from the number of games you finish in a period. The value of a result only should depend on the rating of both players that they have at the start (preferable) or at the finish date of the game and the result. And the rating formula should be like NewRating = LastRating + SumOfAllValuesOfFinishedGamesInThisPeriod. That's the way (idea) IECG is computing the ratings."


This result looks quite normal to me as a player's level may increase as time passes. Rating calculations are done after periods long enough to avoid big differences... I don't know if you're right about IECG rating calculation, FICGS ratings seem to evolve quicker and I think it's best. By the way this system is used by several chess federations.

Best wishes.


Heinz-Georg Lehnhoff    (2007-01-21 15:16:26)
Chess ratings

I have formatted my message (and partly corrected) to be able to read the text better.

Rating system:
Look at this situation: Current rating ot the player: 2200,
10 games, all players with TER 2200, result 9=, 1+

Case 1: Finished games:
01.01-28.02: 9= New rating 01.03: 2200
01.03. 1+, no other finished games in this period: --> New rating 01.05: 2228

Case 2:
Finished games 01.01-28.02: 9=, 1+ --> New rating 01.03: 2214

In case 1 the value of the won game was 28, in case two it was 14.

IMO the value of a result should not depend from the number of games you have finished in a period. The value of a result only should depend on the rating of both players that they have at the start (preferable) or at the finish date of the game and the result. And the rating formula should be like
NewRating = LastRating + SumOfAllValuesOfFinishedGamesInThisPeriod.

That's the way (idea) IECG is computing the ratings.


Thibault de Vassal    (2007-03-19 18:30:34)
Tournament performance

Perf in crosstable is the tournament performance. It is calculated exactly the same way chess ratings are calculated every 2 months. It uses your TER (tournament entry rating) and TER of your opponents.


Jason Repa    (2007-07-15 21:18:20)
IQ is intelligence. Skills are Skills

You're resorting to telling blatant lies again eh burrows? I guess I shouldn't be surprised. This is what you were doing before when you painted yourself into a corner.

I in in way, shape, or form said that "chess ratings prove nothing". You need to re-read my post and stop teling lies. That's very pathetic behavoir.

As I said, it's clear now why you're an under 100 BCF player.


Jason Repa    (2008-05-13 15:25:34)
Elo rating system

I trust I didn't use any "biggie" words you were unable to comprehend Stephenson. Hopefully you now understand your error and realize that the elo system is not only peculiar to FIDE chess ratings, but indeed used by many chess organizations, as well as other games.


Jason Repa    (2008-06-04 10:02:42)
Poker

Groves, it's pretty sad if you aren't capable of understanding what you did. But in your case I suppose I shouldn't be surprised.

And I don't need to meet you in person to know what you're all about. Your words speak volumes. And FYI chess ratings aren't obtained by flipping coins. You're a mere 1600 and change player. It's not the result of "bad luck".

You equate the stating of facts with "insulting". The problem is when I'm dealing with a very modest individual like you, every objective fact I state is interpreted as an insult.

As I said, read a poker book, or get someone to read one to you so that you can learn the basics of the game. Perhaps then you'll understand that there is more to the game than simply managing your money and knowing when to bluff.


Thibault de Vassal    (2008-09-18 20:48:47)
Big Chess Championship

Well, I have to find some time to make changes first. Big Chess ratings are probably too confusing :/


William Taylor    (2008-09-20 17:55:27)
Too confusing?

Why are Big Chess ratings more confusing than any other kind of rating? τΏτ


Thibault de Vassal    (2008-09-22 17:07:19)
Big Chess ratings

They are because "provisional big chess ratings" are correspondence (classical) chess ratings minus 300 points. I estimated this was a good idea because Big Chess was quite different but I'm not sure anymore... What do you think ?


William Taylor    (2008-09-22 18:33:42)
Big Chess ratings

Hmm. As you say, the game is quite different from ordinary chess, and as we can't use engines and very little research has been done into Big Chess we might all be 300 points or so weaker at it than at real chess. It might be quite nice to have the ratings over a similar range to ordinary chess ratings, but that doesn't really matter. Presumably once you have determined the provisional ratings you then use the same algorithm as you do for normal chess ratings? It's hard to suggest a format without an estimate of how many people would play, but we don't necessarily have to have an accurate and established Big Chess rating system to hold a championship.


Thibault de Vassal    (2008-11-22 22:43:23)
Ratings

Chess ratings increase quite quickly here, once you reach 2000-2100, every player most probably use chess engines, but there must be such rules (max. provisional rating) to guarantee the quality of the games at a higher level. It shouldn't take more than 4 or 6 months to reach a 2200 level. Correspondence chess is a game of patience :)


Heinz-Georg Lehnhoff    (2009-03-10 15:22:31)
Big Chess World Championship

Hello Big Chess fans!

I have finished about 120 Big Chess games. Only 5 last more than a year (all with Thibault :) ).

I propose that the WCH should be a 2 stage round-robin tournament starting once per year. As usual the groups of the first stage should be filled according the Big Chess ratings.

The winner of each group is qualified for the second stage. If more players are needed to build a final group of 7, 9 or 11 players Thibault can invite the best of the second placed of the first stage groups.

If all players of the final group are known the second stage should be started within a month.

I hope this WCH will be a success :)


Scott Nichols    (2009-04-05 18:50:07)
Programming

Hi Thibault. Most of us have no clue about what it takes to program. Something we think looks like an easy add-on may be a programming nightmare. It is VERY impressive all of the things you do on this great site. Just one question, is making the chess ratings happen instantly like the poker does one of those programming nightmares?


Thibault de Vassal    (2009-04-05 21:23:03)
Rating calculation : real time or not

Hi Scott, it would be not a nightmare for sure. But the way correspondence chess ratings are calculated is more a question of history IMO, just like FIDE WCH. Many players would probably think this is just a nonsense, while the others may find good reasons for this change. It seems to me that correspondence chess ratings have always been calculated every 2 or 6 months according to the organization. I thought about this question already, there are advantages in both solutions, so I'm not against the idea to open this (big) debate.


Thibault de Vassal    (2009-06-22 16:41:00)
Discussion at Rybkachess

That's an interesting discussion... Once more, the confusion reigns between Freestyle chess (commonly played at classical & blitz time controls) and Correspondence Chess, particularly for centaur players who did not experience correspondence chess at a 2500+ level.

IMO (in brief) on several points :

1) All these made-for-engines books have no other interest than to "manipulate" chess engines & other made-for-engines books, actually this has almost nothing to do with correspondence chess (where they are completely useless at a high level, let's say 2300+) or even chess.

2) Many players do not realize the multitude of factors that appear to be more important that the basic strength of centaurs once the correspondence chess 2400 mark is reached and that still increases at 2500 and 2600... The higher the level, the more "opening books" depend on the recent games played by the opponent (and his level), the number of current games played, the score to reach in 8 games matches, the importance of rating, the goal in life, even the month/season for a few players and many other things according to the persons... Actually these "openings books" just live the time to use it one time, so a better term is preparation, actually opening books do not exist anymore in correspondence chess at a very high level, at most it may be useful against weaker players.

3) The previous point is enough to explain the rating changes of most 2400+ players ! In example...

- GM Farit Balabaev is a very experienced player who constantly has(had) more than 100 running correspondence chess games at several places for years, he's also a fast player, it is quite logical to me that he looks for quiet games and fast draws (or lose sometimes to very strong players who want to win more)

- Wolfgang Utesch, FICGS WCH finalist, like many players at one time in their life, decided that other things were more important and that correspondence chess was too time consuming, particularly once the 2500 mark has been reached...

- Eros Riccio obviously decided to win every correspondence chess competition at FICGS while playing a high number of games at several places AFTER having topped the FICGS rating list with the highest rating so far (which he did), so it is natural to look for a few quick draws in matches if 8 draws mean a victory for him (and a few rating points lost, that is quite inhuman anyway :))

- Michael Aigner tops most FICGS rating lists by playing only games at 40 days + 40 days / 10 moves time control, which is an enormous performance as obviously the longer the time control, the higher the rate of draws. I do not know how many current games he's playing at ICCF or IECG and other organisations but I suspect he plays a quite reasonable number of games.

- Xavier Pichelin may top the FICGS rating list this year as he's an incredibly dangerous player with White and Black and with a reasonable number of running games.

Many strong players also choose to play some tournaments for "fun" or to experiment openings and may lose some points while their real strength is over 2500 or more... so it is quite hard to make the difference between the real strength and correspondence chess ratings. So many parameters... It is likely that we'll see one day a 12 games match between Eros and Xavier (Michael do not play fast correspondence chess time control, yet I hope), we all wonder what rating could achieve Vasik Rajlich (Rybka's creator) and other very strong freestyle players but it is very hard to predict only by knowing their results in freestyle tournaments. Correspondence chess is a mirror of real life.


Thibault de Vassal    (2010-03-13 21:43:04)
Rating calculation

You have to make the difference between ratings updated in real time (like advanced chess ratings) & the FICGS correspondence chess rating calculation, the idea of those ratings updated every 2 months is to avoid peaks, consequently when you win, draw or lose three games after the last rating calculation, your future rating does not take account of the first result THEN the second one THEN the third one, it is actually completely recalculated by taking account of all results at the same time, so you DO NOT win or lose points AFTER EACH result, your performance is recalculated according to the formula explained in the rules & that looks like the french FIDE rating calculation. One thing that explains "strange" variations after 2 or 3 results only is that the rating calculation is just more accurate when you have many results & particularly when your score is near 50%.

That should answer to the discussions I had with Kamesh & Ralph, but maybe my explanations are not so clear, sorry about that again.


Thibault de Vassal    (2010-04-05 00:58:06)
Advanced chess ratings calculation

For some reasons that I'll explain below, I updated the advanced chess (bullet, lightning, blitz, freestyle) rating calculation rules to the following :

"Performance = Opponent Current Rating if the game is drawn, + 350 if the game is won, -350 if the game is lost.

The following bonus / malus applied to White and to Black makes ratings fair, as it is not possible to force a player to take White or Black before a game :

(White) Performance = Performance - 50
(Black) Performance = Performance + 50

If there's a winner and if his rating is below 2400, his new rating his :

New Rating = ((8 x Current Rating) + (2 x Performance)) / 10

Otherwise :

New Rating = ((9 x Current Rating) + (1 x Performance)) / 10

The rating calculation does not take account of wins obtained by a stronger player when the Elo difference is greater than 350 points, the same with losses by a weaker player.

In case of a draw or loss against a player rated more than 200 points less, the opponent's rating considered in calculation is : Current Rating - 200. A player who wins a game cannot lose Elo points, a player who loses a game cannot win Elo points."

More details :

http://www.ficgs.com/membership.html#rating_advanced_chess


The rule that just changed is "If there's a winner and if his rating is below 2400, his new rating his : New Rating = ((8 x Current Rating) + (2 x Performance)) / 10".

This rule will probably be updated again in a few months with a rating limit of 2200 instead of 2400, when advanced chess ratings will be more coherent with correspondence chess ratings.

The reasons are :

1) Advanced/freestyle chess is often neglected partly because players will likely lose some rating points (many strong players using Rybka 3-like engines still have a rating of 1800 or 2000, there are several reasons to this), the main point is probably the interface but I'm fixing it (e.g. the new touch-move option - see Preferences).

2) Chess engines are just stronger and stronger while the ratings do not increase with the previous rules, as a consequence players who just tried advanced chess once years ago shouldn't still top the rating list. It is of course a way for players to find their place quicker in the rating list & to incitate players to play more games as well.


Thibault de Vassal    (2010-04-06 15:41:33)
Ratings calculation

Hi Lazaro, correspondence chess ratings are not calculated in real time, it is calculated every 2 months, see the complete rules :

http://www.ficgs.com/membership.html#rating_chess

You may see an estimation of your future rating (taking account of the games played after the last rating calculation) at :

http://www.ficgs.com/players/munoz_lazaro/history.html

A 1885 rating should look like more logical to you :)


Thibault de Vassal    (2010-04-13 22:30:46)
Big chess ratings

Hi Lazaro!

"Big chess ratings are first estimated from current correspondence chess ratings (current rating -300 points, with at least 1400), then adjusted in real time after each result (...)"

http://www.ficgs.com/membership.html#rating_big_chess

When you entered your first big chess tournament, the TER was not specified but your current correspondence chess rating was used (minus 300) to calculate your first big chess rating with your first result, so a rating of 1609 seems ok taking account of your opponents ratings. The first results may look quite arbitrary but some rules prevent to lose too many points when losing against a strong opponent with a low rating. Anyway there should be more class categories to get more chances to improve ratings (to be continued), we needed more players but maybe we can do it now.


Thibault de Vassal    (2010-04-19 16:41:06)
Advanced chess ratings calculation

As advanced chess (not correspondence chess) ratings move too fast, I just updated the calculation rules :

If there's a winner and if his rating is below 2400, his new rating his :

New Rating = ((18 x Current Rating) + (2 x Performance)) / 20

Otherwise :

New Rating = ((19 x Current Rating) + (1 x Performance)) / 20


Thibault de Vassal    (2010-04-25 02:52:37)
FICGS

I understand you Wayne, I don't want to make such mistakes and that's why we talk so much about these changes in this forum :)

On entry fees for a higher class tournament, I agree on the main point of course, but some advantages had to be discussed. The success of this site is also money and money prizes in the future IMO so I prefer to discuss such ideas than to do nothing.

That was the first point. Then there are some other points that remain to be discussed IMO : 1) Maybe correspondence chess ratings should increase (in average) as engines become stronger. 2) Titles calculation rules should probably be harder as a consequence, maybe it should have been changed already.

Correspondence/Advanced chess is constantly evolving, our marks move fast, so rules may have to change. I don't think that FICGS can turn into a kind of Yahoo chess (I did not ever play there btw), the most important thing is the atmosphere and I know that if I make a mistake, someone will let me know very quickly as it happened once a few months ago. We all make that success in that way!


Thibault de Vassal    (2011-04-24 12:03:10)
Active rating lists

Hello Harshil.

It is mathematical, actually. Your rating cannot reach 2000 if you play only class D tournaments (players rated 1400-1600), which is your case. By the way you played only 3 chess rated tournaments, that is still quite few...

High correspondence chess ratings take always 1 full year to be reached, every strong player knows that.


Thibault de Vassal    (2012-01-10 23:07:31)
Ratings UpDates

Correspondence chess ratings (that are taken in account when a new tournament starts) do not change after every game, but yes you can see your provisional "future rating" as George says.

It is different for Go to allow strong players to climb faster the enormous ladder (2500 points for Go at most). For Poker the difference is less obvious but the game is less serious than chess and it is quite exciting to see this rating list evolving each day :) At the end I wouldn't change anything now.


Thibault de Vassal    (2012-02-27 02:22:02)
Folding in Poker

Well, mathematically poker ratings below about 1900 mean something of a different nature (level of course but also the number of games played) than ratings over 1900 because it is much easier to win points below 2000 (see rules) and it was even easier before february 2011, so it would be much easier after the change asked by Scott. A player who starts at 1600 will need to play more games than a player who starts at 1800 to reach 2000, but not necessarily to make more efforts. In addition there are ways to manage ratings to enter certain waiting lists more quickly. Also, considering the slow inflation that exists the contrary of what you say is true in a certain measure as well, new players will have to play more games than you to reach the top, actually the whole thing is really complex.

But... anyway I'll try not to change the rules again/too many times to avoid such (logical) reactions and that's why I take time to think about this one again.

I think that this change would make the poker ratings more attractive but less realistic and accurate so...... any other opinions? :)

Also, new players DO NOT get a free ride to the B class tournaments, many still start with 1600 according to the level they pretend when registering. So the difference is not so much, actually it may help you to climb the scale faster if you can beat a 1800 player easily... Really complex as I said but anyway I think that ratings are more accurate when players can start at different levels, because more players in the different categories mean more games in each one (players will find their rating faster) and because everybody do not lie every time. Everybody will not agree with this but I have a certain experience with the chess ratings now and I'm quite certain that most changes were good ones, so probably for poker.


Thibault de Vassal    (2012-07-28 00:10:50)
FICGS IS BACK !!!!!!!!!

First of all, I've added 15 days to all players in running games because of the delay and the fact that many of us may have no access to internet during the summer vacation (this time is added to the 46 days, 13 hours and 20 minutes since the crash for players expected to play), this issue was discussed at Rybkaforum, of course it may be unfair to few players in certain games where their opponent had few time but I did not find a better balanced solution, sorry about that :(

Among other consequences, the current championships cycle will last 10 months instead of 8, and july correspondence chess ratings will be updated very soon.

Well, how to start... fortunately such an event is rare but possible, and following the Murphy's law, it happened (first time for me), the server's hard disk crashed and the least I can say is I've not been lucky, even if I obviously did some things wrong.

Of course I had enough data at home to rebuild all games until a few hours before the crash but I thought it was worth it to pause the server during a few days/weeks to recover more moves, and if possible ALL moves. I really hoped that it would work and at the end it did, but not completely... for unknown reasons. I had also other data to recover from the server, including some FICGS data that were not backuped correctly (my bad), because I did not think far enough 6 years ago when I coded the first FICGS scripts... That will be fixed very soon.

So, because the DDrescue process did not work -unlucky- just after the crash, my server provider (OVH in France) had to send me the hard drive and it took sooooo much time already :/

Then I tried to recover some files and the databases by myself and I learnt much on how to save a hard drive but each process was really long, it took several days again...

Finally none process completely succeeded, few sectors of the hard drive remained unreadable and unfortunately the FICGS database is divided into very numerous parts written everywhere on the disk.

At the end, I brought the disk to the very best professionals able to save it... the process was quite long again and it did not completely worked as well, for an unknown reason the current database was still not readable but they did much better than me at the end.

Finally the whole process was worth it, but I did not expect it could take so much time.... 46 days, 13 hours, 20 minutes. And that's a shame :(


Of course, I could have used a RAID 10 server, I was not favourable to this choice because it is not 100% safe as well, I don't know it enough and it's much more expensive. I'll reconsider it though.

But the other things I did wrong are clear anyway, I lacked of experience in such a situation and most important, I'll do now better backups also on another server every hour. Next time (if any), we'll lose at most 1 hour of moves but the server will be able to restart within 1 day.

One thing is sure, internet was really empty for me without FICGS during this long month and a half and I missed our tournaments too much so that happen again! Have no doubt, FICGS would not have stopped in all cases but once again I'm really sorry about that and all consequences... I can only hope that you'll enjoy your games as before.

Thanks for your understanding.

Best regards,
Thibault


Thibault de Vassal    (2018-03-27 02:19:00)
Ratings inflation period

Dear chessfriends, in order to make FICGS correspondence chess ratings somewhat more coherent with other ones & with the real chess level (which is quite important in many points of view), we start an "inflation period".

During this time (that will last several months or maybe a few years according to the effects), the rating calculation will differ as explained in the rules: "During an inflation period, 10 points per game are added to the bonus, then any negative bonus is divided by 2. [Edited]"

As a side-effect, fortunately this may encourage even more games and more wins :)

Let's just wait and see the evolution in the next rating lists, then there may be adjustments.


Thibault de Vassal    (2018-11-01 02:24:14)
Ratings inflation period

Dear chessfriends, the inflation period continues for correspondence chess ratings, now with slower effects:

The rating calculation differs as explained in the rules: "During an inflation period, any negative bonus is divided by 2. [Edited]"


Thibault de Vassal    (2021-04-28 00:56:13)
Players ratings

Hello Misha,

"My messages" page specifies : Dear chessfriends, the correspondence chess ratings and FICGS chess database have been updated on March 02, 2021 (next update around May 01, 2021).

Advanced chess ratings, big chess ratings, Go ratings & poker ratings are updated after each game.


Garvin Gray    (2024-09-19 18:46:08)
Rating System.....

Correspondence chess ratings next update on November 1st, 2024. FICGS chess database has been updated on September 16, 2024.








 		 
 
Support to all people under attack    




Social network : create your photo albums, discuss with your friends...
Hot news & buzz : discover the latest news and buzz on the internet...
Discussions : questions and answers, forums on almost everything...
Seo forums : search engines optimisation forums, web directory...


Play the strongest international correspondence chess players !


Ortiz, Clodomiro     (DOM)        [member # 9683]

Correspondence chess : 2438       FSM

Ranked  #  61   in the rating list.

Rating history :  No change this year.



Advanced chess :   2350      




This member has no profile in the social network.




FICGS Go server, weiqi baduk banner facebook      
Correspondence chess

World championship

Play chess games

Go (weiqi, baduk)

Advanced chess

Play big chess

Chess trainer apk

Rated tournaments

Poker texas hold'em

Fischer random chess

       FICGS correspondence chess banner facebook