Rook
We could also consider migrating the CFC rating to the Glicko-2 rating system. The Australian Chess Federation has already done that. The system gives both a rating and a measure of the reliability of the rating. If the CFC rating of a given player is really unreliable, this will be shown.
https://en.wikipedia.org/wiki/Glicko_rating_system
http://auschess.org.au/acf/ratings/
CFC President
Originally Posted by Pierre Dénommée
I don’t see the advantage of migrating to someone else’s rating system.
Rook
The advantage is that the system identifies unreliable ratings. The actual system is quite old, it does not take the Standard Deviation into consideration. Sandbaggers who looses to 1500 players and defeat 2400 players would have their rating flagged as unreliable by the new system. Australia considers that only very reliable and reliable ratings should be used on the Australia top players list.
Any formula that gives both a value and a measure of the confidence of that value is preferable to a formula that only gives a value. Doing ratings the Elo way is becoming more and more obsolete as time pass. Instead of being subjective, the reliability of a rating becomes both objective and measurable. FICS has already moved away from Elo ratings and new endeavours, such a Couter-Strike GO computer gaming, have begun rating with the Glicko formula.
When enough member countries will have moved away from Elo, FIDE will likely follow.
CFC Secretary
To Garland, I've never found navigating the FIDE list particularly difficult - the main thing I have to remember is to set the country to Canada. Nearly everything I do on the FIDE site is NOT logged in.
If there are obvious errors Hal is the best way to get it looked after. For instance some years ago FIDE had me as female and Lynn Stringer as male! I e-mailed Hal, they took care of it quickly! On another occasion they forgot to remove Abe Yanofsky from the International Arbiter list - everybody knew Abe was a GM but few knew about his IA...
CFC President
I have not seen many obvious examples of sandbagging recently though I do recall some instances in the 1970s and 1990s. I don’t think it is practical to change systems at this point.
Bishop
Well, let's look at examples. (Apologies i do not have full access to the FIDE database as i am not registered, so in some places I'm making an educated guess. Someone can please correct my calculations).
Going with Top 10 CFC rated players:
Evgeny Bareev: played 26 FIDE rated games over past 36 months, so just FIDE works.
Razvan Preotu: played more than 20 FIDE games this summer, so just FIDE works.
Eric Hansen: played more than 20 FIDE games this year, so just FIDE works.
Nikolay Noritsyn: played more than 20 Fide games over past 18 months, so just FIDE works.
Bator Sambuev: Lots of FIDE games, so just FIDE works.
And yes, no need to do any math for the top 20 Canadian players. They have all played sufficient FIDE games.
Let's move to the women:
Qiyu Zhou - easily has 20 FIDE games this year.
Agnieszka Matras-Clement - also meets the 20 game FIDE number
Alexandra Botez - played only 16(?) FIDE games the past 2 years, but has CFC rated events to make the difference. FIDE rating: 2051 CFC rating:2273 Rating = [16 * 2051 + 4 * 2273]/20 = 2095
Valeria Gansvind - has four rating entries, so i think she meets the 20 game requirement.
Natalia Khoudgarian - played only 15(?) games the past 2 years, all FIDE and CFC rated. Eligibility issue.
Lali Agbabishvili - played 20 FIDE games this year, so FIDE rating.
Jackie Peng: Not enough games to be eligible.
Nava Starr: Last FIDE event in September 2015. I'll assume 5 games. Has more than 15 CFC games. FIDE rating: 2127 CFC rating:2142 Rating = [5 * 2127 + 15 * 2142]/20 = 2131
I don't see how these calculations are more complicated, and for almost all players the FIDE rating is the one that ends up being used, 100%.
I think this is a win-win. You get your FIDE ratings used for the top players, while the others get rated based on their total games played, not just FIDE games, and we don't discriminate based on gender, age, etc.
CFC FIDE Representative
Garland,
In your example we use FIDE rating only for the Woman Team also. Even for active player with FIDE around 2000-2100, CFC still is very reliable, especially for juniors (girls) with the FIDE K-factor of 40.
Another problem in your example: your propose to rely more on CFC rating for the less active player. Usually, CFC is higher than FIDE by 100-150 points. So, less active player gets a big advantage. I do not think we want it.
I don't see any problem if we use different formula for different teams. Their average rating is completely different (2600 and 2100). I see absolutely no gender discrimination here. We can use borderline of 2300. If Woman Team has the average FIDE above 2300, we switch to using FIDE rating only for them as well. Hopefully, the average rating for the National Team remains above 2300 for the foreseeable future.
Last edited by Victor Plotkin; 11-22-2016 at 08:20 AM.
CFC Voting Member
Victor, I prefer the way you phrased it in the last paragraph (average rating vs. gender, there'll always be someone crying "gender foul". I also want to take a moment to say thanks for even having thought of all these angles, and how to bring them together. I respectfully caution any of my peers for overly picking on any single criteria. I have deep confidence Victor is building something much more just than we have ever seen before, and he's broken it down so we can vote on parts of it. However, doesn't it defeat the purpose of the exercise if we overly shoot down or reinvent-the-wheel on any single aspect, especially if the result is out-of-sync with Victor's overall vision? This is just my personal opinion, and of course you don't have to agree. So far, I expect to be voting with confidence for Victor. If it's a close call in my mind, I'm going with Victor's judgment.
CFC Voting Member
Aris, I agree with your post 100%.
Victor, take it easy and thank you for well thought motions. I'll vote Yes to all 6 motions.
Bishop
Folks, in case anyone has the wrong impression, I also recognize and appreciate the thought that Victor has placed in drafting his proposal. I just want us to consider other possible options, hopefully better ones, before committing to a final formula.
I just want to make it clear that my posts are given as constructive criticism, with the goal of giving us the best formula possible. It was never intended to "shoot down" Victor. Far from it. But the voting should be based on whether we believe each portion of the proposed formula is objectively the best option, or if a better one exists. And if we think there is a better option we should voice that option and be prepared to defend it. Isn't that the point of these meetings?
Posting Permissions
