FIDE, CFC and USCF ratings

[Victor Plotkin]

In another thread (started by Larry Bevand) I had a discussion with some chelltalkers (mostly from QC) about the gap between FIDE, CFC and USCF ratings. I decided to make an additional attempt to convince these players, that CFC rating is significantly higher than FIDE and USCF.

1. Start with easy one: the gap between USCF and FIDE.

Thanks to Kevin Pacey, we know the formula, calculated by USCF about the gap between USCF and FIDE rating.

180 + 0.94 x FIDE if FIDE <= 2000
20 + 1.02 x FIDE if FIDE > 2000

If translate these numbers to normal language, we get:

1900 player - the gap is 66.
2000 player - the gap is 60
2100 player - the gap is 62
2200 player - the gap is 64
2300 player - the gap is 66
2400 player - the gap is 68

We have no reasons to argue with USCF about this data. So, the average gap must be around 64 points. Not too far from my estimate of 60 points.

Actually, we have no argument here, everyone agreed that USCF is slightly higher than FIDE. The argument was about CFC rating.

2. The number of 130, which I believe is the average CFC-FIDE gap, actually came from some random calculation. A few years ago (I recall, it was Canadian Closed 2011 in Guelph) I calculated the gap for all participants. The average gap was 129.

As a mathematician, I know that data of just 30+ players is not exactly a reliable one. However, it should be close to the real number.

I checked the gap for top-10 players from current Canadian Junior and Canadian Closed-2017.

Starting with the Canadian Junior, we have:

Chen: 2324 - 2198 = 126
Hua: 2317 - 2218 = 99
Cai: 2297 - 2130 = 167
T.Song: 2250 - 2174 = 76
Vettese: 2249 - 2102 = 147
Dukic: 2238 - 1952 = 286
Liang: 2231 - 2143 = 88
Nasir: 2221 - 2068 = 153
Zhao: 2195 - 1928 = 267
Yie: 2146 - 1900 = 246.

The average number is 165.5. I believe, this number is really important. Nowadays, junior are the most active part of Canadian chess community. In some week-enders, I played all 5 games with juniors and was the only competitive adult player in the Open section.

In Canadian Closed-2017 I checked the actual (that time) rating and not the last one. So we have:

Sambuev: 2609 - 2513 = 96
Noritsyn: 2598 - 2473 = 125
LeSiege: 2572 - 2525 = 47
Hambleton: 2547 - 2471 = 76
Thavandiran: 2474 - 2334 = 140
Yu Zong: 2466 - 2393 = 73
Sohal: 2383 - 2319 = 64
Plotkin: 2380 - 2265 = 115
Zhang: 2357 - 2248 = 109
Kleinman: 2356 - 2366 = -10

The average gap is 83.5 points. This number is very low. I can find 2 reason to explain this:

a) Many players from QC (5), where the gap is much lower.
b) 2 players (Hambleton, Kleinman) played mostly overseas in recent years. Their improvement in FIDE rating was not yet translated into CFC rating. Actually, 3 players from this top-10 played in Reykjavik few months ago and gained more than 100 points combined (Hambleton, Kleinman, Plotkin)

I calculated my personal gap for last 7 years (end of which year 2011-2016) and the current one.

So we have:

2011: 2374 - 2253 = 119
2012: 2350 - 2214 = 136
2013: 2426 - 2228 = 198
2014: 2465 - 2306 = 159
2015: 2422 - 2295 = 127
2016: 2408 - 2245 = 163
2017: 2375 - 2255 = 120

The average gap is 146. While it's little bit tricky to analyze the numbers of just 1 player, I think that my personal numbers should be really reliable in this case.
a. All this years I was pretty active and played a lot of games.
b. Most of this games were played in Canada
c. The rating was relatively stable
d. My k-factor was 20 for FIDE rating (not 40, like for many juniors).

So, we have a few numbers. Some of them are more reliable, some of them are less reliable, but we should be able to see the general picture.

Canadian Closed - 2011: 129
Canadian Juniors-2017: 166
Canadian Closed-2017: 83
My personal rating in 7 years: 146.

Funny, but the average for these 4 numbers is... 131!! Without my numbers it's 126 exactly.

3. So we have, more or less:

USCF - FIDE = 60
CFC - FIDE = 130

That means, that CFC - USCF = 70.

The only 1 issue I have here is the question: is the FIDE the same for Canadian and American players with the same strength?

While I do believe, that our FIDE rating is deflated compared to Europe, I don't think we have a big gap with US with our FIDE rating. Like in Canada, there are many underrated juniors in American tournaments, which makes it difficult to gain FIDE rating in USA.

[Hugh Brodie]

Re: FIDE, CFC and USCF ratings

There can be a problem taking junior player's ratings. Normally their ratings tend to rise quickly. When did each of their last US events take place, and are their US ratings permanent? Best would be to take people (adults) who play regularly in both Canadian and US events (e.g. Mike Dougherty - CFC: 2244; US: 2257).

[Victor Plotkin]

Re: FIDE, CFC and USCF ratings

Your example with Mike uses his current (I hope temporary) drop with CFC rating. His average CFC in last few years was around 2280-2290.

[Paul Leblanc]

Re: FIDE, CFC and USCF ratings

When I did the comparison that I put in the other thread I noticed that many of the juniors had USCF ratings that lagged their CFC ratings not because there was a difference in the rating systems but because their CFC ratings were rising fast and they didn't play frequently enough in the US for their USCF ratings to keep up.
I suppose if you want to do a direct comparison you'd have to examine only players that play frequently enough in both systems for both their USCF and CFC ratings to be current.
Your use of FIDE vs CFC to imply CFC vs USCF is a clever way around that, provided of course that you choose only players with current CFC and FIDE ratings.

[Victor Plotkin]

Re: FIDE, CFC and USCF ratings

When I did the comparison that I put in the other thread I noticed that many of the juniors had USCF ratings that lagged their CFC ratings not because there was a difference in the rating systems but because their CFC ratings were rising fast and they didn't play frequently enough in the US for their USCF ratings to keep up.
I suppose if you want to do a direct comparison you'd have to examine only players that play frequently enough in both systems for both their USCF and CFC ratings to be current.
Your use of FIDE vs CFC to imply CFC vs USCF is a clever way around that, provided of course that you choose only players with current CFC and FIDE ratings.

Paul,

There are very few players with enough CFC and USCF games at the same time. Most of these players are juniors, their ratings and a real strength change too fast. Thus, any research based on these players is not too reliable.

[Victor Plotkin]

Re: FIDE, CFC and USCF ratings

I understand, why some old players still believe that CFC rating is similar to USCF or even lower. For many years, before 2008 it was indeed the case. I started playing in 2004 and the average CFC-FIDE gap was around 50-70 points that time. Really, very similar to USCF-FIDE one.

In 2008, for some "funny" reasons CFC decided to organized "artificial" inflation. CFC gave huge bonuses and participation points. As a result, many active players gain 100 CFC points or even more during 2 years. At the same time, their FIDE (and a real strength) remained the same.

We all know that big inflation is a serious problem for the real economy. The same applies to the rating system. "Artificial" inflation damaged credibility of CFC rating system. Only 2 or 3 years later, CFC decided to stop this experiment.

Last many years, we don't have any serious inflation. Our rating auditor, Paul Leblanc periodically checks different average rating and finds them fairly stable. However, the damage was done.

In my opinion, the main problem is not the CFC-FIDE gap. The main problem, this gap is very different in different parts of our big country.
Ontario, as the most active (chesswise) province, "enjoyed" the period of inflation more than any other province and much more than QC.

I estimate the current gap is around 150 points in Ontario (yes, included juniors, they are big and the most active part of the chess community), 100 or less in other provinces and around 60 points in QC. Sooner or later, we will have smaller differences in CFC-FIDE gap, but it will take a long time. In case of QC, the problem will exist almost forever, because very few players from QC are regularly playing in tournaments outside of their province. Definitely, Bator is one of them.

However, even for him, the gap was 150 points or even more, while we lived in Toronto. I remember, his CFC was above 2700 level and it's around 2600 right now. At the same time, his FIDE was fairly stable, around 2520-2550 level. So, he dropped 100 CFC points not because he suddenly became a weaker player (he is too young for this) but because he plays mostly in QC now. His average gap in recent years was about 80 points.

I have to say, I have no simple solution for this problem (how to cut difference between CFC-FIDE gap in Ontario and QC). At the same time, some obvious measures could and should be done by CFC and chess organizers to create a better rating system.

1. Every foreign chess player, who plays for the first time in Canada, should get starting rating not FIDE, but FIDE + X. We can discuss the exact number of X. In my opinion, it should be 130. Sure, I understand, that other players may have different opinion. However, it does not make any sense to have this number less than 100.

The same rule should apply for player who did not play in Canada for a very long time and his old CFC rating doesn't reflect his real strength.

2. Organizers should use exactly the same rule, while deciding about class-prize eligibility, section's cut-off or even Swiss pairings (if they use CFC rating for pairings).

[Kevin Pacey]

Re: FIDE, CFC and USCF ratings

...
1. Start with easy one: the gap between USCF and FIDE.

Thanks to Kevin Pacey, we know the formula, calculated by USCF about the gap between USCF and FIDE rating.

180 + 0.94 x FIDE if FIDE <= 2000
20 + 1.02 x FIDE if FIDE > 2000
...

Actually, in the other chesstalk thread you referred to, I was merely all but quoting verbatim part of a page on a USCF pdf that Tom O'Donnell provided (and mentioned said page number). Tom deserves all of the credit, at least for finding said pdf.

[Egidijus Zeromskis]

Re: FIDE, CFC and USCF ratings

1. Start with easy one: the gap between USCF and FIDE.

Thanks to Kevin Pacey, we know the formula, calculated by USCF about the gap between USCF and FIDE rating.

180 + 0.94 x FIDE if FIDE <= 2000
20 + 1.02 x FIDE if FIDE > 2000

If translate these numbers to normal language, we get:

1900 player - the gap is 66.
2000 player - the gap is 60
2100 player - the gap is 62
2200 player - the gap is 64
2300 player - the gap is 66
2400 player - the gap is 68

We have no reasons to argue with USCF about this data. So, the average gap must be around 64 points. Not too far from my estimate of 60 points.

It's interesting that the 6th USCF handbook gives simpler formulas:
USCF = FIDE + 50 (average conversation)
USCF = 0.895 * FIDE + 367
USCF = FIDE +100

It seems that the TD can choose one of those formulas for placing players into class sections. "... thus providing a degree of protection for the players with the established USCF ratings..."

The same handbook:
for Canada - no adjustments needed.
Quebec (FQE): add 100 points.

[Victor Plotkin]

Re: FIDE, CFC and USCF ratings

It's interesting that the 6th USCF handbook gives simpler formulas:
USCF = FIDE + 50 (average conversation)
USCF = 0.895 * FIDE + 367
USCF = FIDE +100

It seems that the TD can choose one of those formulas for placing players into class sections. "... thus providing a degree of protection for the players with the established USCF ratings..."

The same handbook:
for Canada - no adjustments needed.
Quebec (FQE): add 100 points.

I love this wording ... " degree of protection". Really, a very modern politically correct language. In my language, I call it discrimination of foreign players.

I played many tournaments in USA. Every time TD added 100 points to FIDE rating. They know well, that we average gap is around 60, but add 100 points - extra 40 points.

Actually, Canadian players suffer double-discrimination. American TD's not only add 100 points to our FIDE rating, but also take CFC rating as is. Usually, the take the maximum for (FIDE + 100) and CFC.

I can provide one example from my previous experience. A few years ago, I played in one of the major American tournaments. My ratings was something like 2260 FIDE and 2410 CFC. They have a class prize for U2400. Before the tournament, I asked the TD about my eligibility for the prize. The answer was negative because my CFC was too high.

So, a random 2330 FIDE (around 2390 USCF) American player is eligible for this prize, but a 2260 FIDE Canadian is too strong to compete for it. Really amazing.

The funny thing is, a few days later I played in Canadian Open. They also had class prize for U2400 (this time CFC, not USCF). Obviously, with my CFC of 2410 I was not eligible for the class prize. However, a random 2390 FIDE player, who never played in Canada (from Europe, Russia, China, India...) is eligible for this prize. Again, 2260 FIDE is too strong, but 2390 FIDE is OK.

Canada is really a very friendly country...

I have to say, that last Canadian Open TD used FIDE rating only for the Open section. Thanks to Andrew Peredun.

[Kevin Pacey]

Re: FIDE, CFC and USCF ratings

Am I correct that once a player (US or non-US citizen) at least has an established USCF rating (based on playing 25+ games!?), US organizers of events played on US soil must use that player's established USCF rating for pairing & prize elibigibility purposes (unless FIDE ratings are previously announced as to be the ones used, in case of a FIDE rated event/section)?

[Paul Leblanc]

Re: FIDE, CFC and USCF ratings

If we feel that it would be beneficial to bring CFC ratings down a bit (this is a big "if"), there is a solution; shut down the bonus point system for a few months and see what happens.
Most players I know want their ratings to go up. I never get complaints from players saying their rating is too high.

[Bob Gillanders]

Re: FIDE, CFC and USCF ratings

If we feel that it would be beneficial to bring CFC ratings down a bit (this is a big "if"), there is a solution; shut down the bonus point system for a few months and see what happens.
Most players I know want their ratings to go up. I never get complaints from players saying their rating is too high.

Paul, instead of shutting down the bonus system, we need to super charge it.
My sense is that a lot of juniors are getting trapped in the 800 - 1200 range. Winning most of their games but not seeing big rating increases because they are playing mostly junior tournament and playing even lower rated kids. Believe it or not, their ratings are being depressed just like the adults in the 1200-2000 range who have to play under rated juniors.

The bonus system adds 20 points (B2) for those achieving new lifetime highs. This is should be at least 100 points for those rated under 1200. Maybe you can get that on the floor for the upcoming meeting.

[Vlad Drkulec]

Re: FIDE, CFC and USCF ratings

Paul, instead of shutting down the bonus system, we need to super charge it.
My sense is that a lot of juniors are getting trapped in the 800 - 1200 range. Winning most of their games but not seeing big rating increases because they are playing mostly junior tournament and playing even lower rated kids. Believe it or not, their ratings are being depressed just like the adults in the 1200-2000 range who have to play under rated juniors.

The bonus system adds 20 points (B2) for those achieving new lifetime highs. This is should be at least 100 points for those rated under 1200. Maybe you can get that on the floor for the upcoming meeting.

We can get anything on the floor as a discussion item through new business. Motions from the executive - which all motions are these days because of the ridiculously long notice required on member motions under the NFP act- need three to seven weeks notice to comply with NFP act regulations.

[Kerry Liles]

Re: FIDE, CFC and USCF ratings

under the NFP act- need three to seven weeks notice to comply with NFP act regulations.

I'm not even sure there is normally 3-7 weeks notice before a meeting is held!

[Vlad Drkulec]

Re: FIDE, CFC and USCF ratings

Am I correct that once a player (US or non-US citizen) at least has an established USCF rating (based on playing 25+ games!?), US organizers of events played on US soil must use that player's established USCF rating for pairing & prize elibigibility purposes (unless FIDE ratings are previously announced as to be the ones used, in case of a FIDE rated event/section)?

I don't believe that is correct. It certainly is not the Continental Chess Association procedure.