Correspondence ratings vs over-the-board ratings!

Running Chess Tournaments
1

There are a number of over-the-board players and correspondence players with established ratings. The players with both ratings have been asking themselves why their correspondence rating is greater then their over-the-board rating. The ratings for the first time correspondence player could be over rated in the first place, even if they pick the bottom rating class (class D).

At one time or the other have been in correspondence play, checking the other players over-the-board rating, finding most but not all the other players in the event without any over-the-board rating of any type (regular or quick). Only someone in the office, could give a break down of the correspondence players with established over-the-board (regular) rating, with established correspondence rating, against, only having a established correspondence rating with no over-the-board (regular) rating.

When someone starts to play correspondence chess, without any over-the-board rating, the worst rating class they can ask for, would be as a class D player; if they do have an over-the-board rating, they could have their rating as a class E or below, only having their rating class placed as a class D correspondence rating. The correspondence department will let a player pick the rating class (class A, class B, class C, and class D) if they have no over-the-board rating.

USCF regular rating distribution chart as of November 1, 2004

Rating range 100 - 199, percentile 4.83% of non-scholastic members regular rating.
Rating range 100 - 299, percentile 7.86% of non-scholastic members regular rating.

Rating range 100 - 399, percentile 11.41% of non-scholastic members regular rating.
Rating range 100 - 499, percentile 15.43% of non-scholastic members regular rating.
Rating range 100 - 599, percentile 19.63% of non-scholastic members regular rating.
Rating range 100 - 699, percentile 24.16% of non-scholastic members regular rating.
Rating range 100 - 799, percentile 28.60% of non-scholastic members regular rating.
Rating range 100 - 899, percentile 32.87% of non-scholastic members regular rating.
Rating range 100 - 999, percentile 37.18% of non-scholastic members regular rating.

Rating range 100 - 1099, percentile 41.55% of non-scholastic members regular rating.
Rating range 100 - 1199, percentile 46.24% of non-scholastic members regular rating.

If someone started to play correspondence chess, even if they did pick the rating class of class D (1200 - 1399), 46.24% (non-scholastic members) would be over rated to their over-the-board regular rating. If a scholastic player did play correspondence chess, if they did pick the rating class of class D (1200 - 1399), 92.21% would be over rated to their over-the-board regular rating. Having players with over rated ratings, would only make the other players have in time over rated ratings – when they play and win.

When someone has an over-the-board rating with a correspondence rating, with them having a established over-the-board rating at or higher then a class D. The players they do meet up and win against in correspondence chess, they should be winning against players with over rated correspondence ratings from the start.

The only way the correspondence ratings could get closer to the over-the-board ratings, would be asking the correspondence department to accept players asking for a rating class of class E (1000 - 1199), and class F (800 - 999). Even with this, the percentile would (non-scholastic members regular rating) be 28.60% below a class F, or (scholastic members regular rating) 66.41% for the scholastic players. When over-the-board players are asking why their correspondence ratings are inflated, as they are being paired up with players with over inflated ratings from the start.

For all members the average regular rating is 1068 (class E), or, for all members the quick rating is 977 (class F) – how can the correspondence department only start players with their worst rating class as class D (1200 - 1399)? For only the (adult members) the average regular rating is 1198 (class E), or, adult members the quick rating is 1097 (class E); the scholastic members regular rating is 668 (class G), or, scholastic members quick rating is 675 (class G) – how can the correspondence department only start members with their worst rating class as class D (1200 - 1399)? If someone has a average rating or lower then the average rating, the player would be asking the correspondence department to place them into a higher rating class in the first place.

When the established over-the-board player is asking whey their correspondence rating is 300 or 500 points higher then their over-the-board rating – the evidence points to the correspondence rating being inflated from the very first game.

2

There are only around 750 USCF members with both an active OTB rating and an active correspondence rating since 2002.

However, in looking at those players, it appears that for players under 1200 OTB their correspondence rating tends to be about 300 points higher than their OTB rating.

For 1600 OTB players, their average correspondence rating is only about 200 points higher, by the time you get to 2000 OTB players their correspondence rating is about equal to their OTB rating, and for masters and above their OTB rating is higher than their correspondence rating.

While it is questionable that OTB and correspondence ratings measure the same skill set, a reasonable conversion function from correspondence to OTB would appear to be:

OTB = CORR * .75 + 600
3

With the regular rating (2004 Annual Rating List, the USCF regular rating distribution chart as of November 1, 2004), the total active members are 65,455. The active over-the-board (regular rating) member with being an active correspondence member, only being 1.15% of the total active members in over-the-board (regular ratings) activity for the 2004 Annual Rating List. Since being data of 750 correspondence members since 2002, the percentile of active over-the-board (regular rating) with being active correspondence members, should be equal or lower then 1.15% during the calander year of 2004.

4

Over the board = Correspondence
(OTB * 0.75) + 600 = correspondence

(100 * 0.75) + 600 = 675
(200 * 0.75) + 600 = 750
(300 * 0.75) + 600 = 825
(400 * 0.75) + 600 = 900
(500 * 0.75) + 600 = 975

(600 * 0.75) + 600 = 1050
(700 * 0.75) + 600 = 1125
(800 * 0.75) + 600 = 1200
(900 * 0.75) + 600 = 1275

(1000 * 0.75) + 600 = 1350
(1100 * 0.75) + 600 = 1425
(1200 * 0.75) + 600 = 1500
(1300 * 0.75) + 600 = 1575
(1400 * 0.75) + 600 = 1650
(1500 * 0.75) + 600 = 1725
(1600 * 0.75) + 600 = 1800
(1700 * 0.75) + 600 = 1875
(1800 * 0.75) + 600 = 1950
(1900 * 0.75) + 600 = 2025
(2000 * 0.75) + 600 = 2100
(2100 * 0.75) + 600 = 2175
(2200 * 0.75) + 600 = 2250
(2300 * 0.75) + 600 = 2325
(2400 * 0.75) + 600 = 2400

5

If someone does have an over-the-board rating, for finding the mode of their over-the-board (regular rating) to their correspondence rating: if they are going to join correspondence chess for the first time.

Over the board = correspondence

Over the board regular rating 800 - 1066 = correspondence class D

Over the board regular rating 1067 - 1333 = correspondence class C

Over the board regular rating 1334 - 1599 = correspondence class B

Over the board regular rating 1600 - 1866 = correspondence class A

Correspondence class D
(800 * 0.75) + 600 = 1200
(1066 * 0.75) + 600 = 1399.5

Correspondence class C
(1067 * 0.75) + 600 = 1400.25
(1333 * 0.75) + 600 = 1599.75

Correspondence class B
(1334 * 0.75) + 600 = 1600.5
(1559 * 0.75) + 600 = 1799.25

Correspondence class A
(1600 * 0.75) + 600 = 1800
(1866 * 0.75) + 600 = 1999.5

The conversion should be:

Class D 800 - 1066
Class C 1067 - 1333
Class B 1334 - 1559
Class A 1600 - 1866

6

correspondence = over the board (regular rating)

Correspondence class J: 675 - 799 = OTB 100 - 266
Correspondence class F: 800 - 999 = OTB 267 - 533
Correspondence class E: 1000 - 1199 = OTB 534 - 799
Correspondence class D: 1200 - 1399 = OTB 800 - 1066
Correspondence class C: 1400 - 1599 = OTB 1067 - 1333
Correspondence class B: 1600 - 1799 = OTB 1334 - 1599
Correspondence class A: 1800 - 1999 = OTB 1600 - 1866
Correspondence expert: 2000 - 2199 = OTB 1867 - 2133
Correspondence master: 2200 - 2399 = OTB 2134 - 2399

As the OTB rating floor is 100, the correspondence rating floor should be 675.
(100 * 0.75) + 600 = 675

7

If there was a 1 to 1 relationship between OTB and correspondence ratings, I might agree with you. But there isn’t.

There’s closer to a bell curve relationship. There are OTB players whose
correspondence rating is significantly higher than the conversion
formula I gave, and those whose correspondence rating is significantly
lower, too, but I think the majority will be within 50 points or so based on the conversion formula I posted.

The same thing is true when comparing FIDE and USCF ratings.

With only around 15 active correspondence players rated below 600, the numbers may not be large enough to support further analysis.

8

Will agree, there is or never shall be any relationship between the two ratings. When leaving correspondence chess my rating was over 1700, after leaving with over 40 active boards going, my rating did drop to little over 1000. Asked for my correspondence rating, if it could be removed to unknowed – as there is no designs to be in any correspondence chess events during the rest of my life.

They (correspondence department) just changed my correspondence rating from a low 1000 rating to a flat 1500 rating. Very sure Nancy (Nancy Evans) would not increase my OTB rating by the same amount because of a phone call. With one department willing to change the rating (as being the departments idea) because of a phone call, and the other would find it as a taboo. Then the two ratings can and only be as a bell curve relationship.

9

There have been a couple of references in this thread to a November 2004 ratings distribution chart. I’ve poked around the USCF site, and haven’t been able to find it. Is this chart online somewhere?

10

uschess.org/ratings/ratedist.php

11

I have both ratings, although my correspondence rating is with ASPCC. My OTB hovers around 1800 and my correspondence rating around 2080. One difference for me is that I’ve never faced those darn underrated scholastic players in correspondence. I’ve contributed numerous rating points to them OTB. I believe my OTB rating is deflated because of the youngsters in the rating pool, and my correspondence rating is inflated for unclear reasons. Most players in our club tend to get low initial ratings. I can’t say whether that is true in USCF.

12

When I did play correspondence chess, only recall of one time being paired with a scholastic player (he, only sent one or two postcards). Do not know of any scholastic player, that plays correspondence chess. If a parent ask me if their child should play correspondence chess, would tell them it would not be a good idea.

Its’ not the factor with correspondence chess, with scholastic players inflating the ratings. Feel its the factor of someone picking a higher rating class from the start, in the argument they would be paired with stronger players.