Ratings and Provisionals

Running Chess Tournaments
1

I really don’t know how to explain this to someone, but wouldn’t mind if someone can explain this to me:

Say I’m a 1600 player. If I lose / draw / win against a provisional, say they’re 1200 based on 5 games, how do I figure out my rating which is established. Of course, it’s easy to figure out the provisional with the ±400 point thing.

I assume you can’t have swissys figure this out for you since the player is either unrated or has a rating in the Rating1 field.

This question comes up frequently at tournaments when players ask about their rating. With the bonus and so forth, I just let swisssys do the talking, but somewhat impossible with provisionals.

Thanks in Advance.

2

The debate during the old school of ratings, the ‘unknowed and provisional’ rated players are rated first, then the established players. With the change in the rating tables, the old tables have changed to a system that the ‘common player’ would have little clue what the ratings would be. Not sure if the unknowed or provisional players (if they are still rated first) are rated first, as it has been the goal of the executive board not the rating department to tweak the ratings into rating inflation.

3

Huh?

4

If it is a game between a 1600 established player and a 1200 provisional rating, if the 1600 player does win you should win zero or one point. Experts never become masters if they play people with 400 points below themselves, then again experts can become class A player if they always play players 400 points below themselves.

5

Yea, that’s basic knowledge but has nothing to do with my problem.

6

See page 262 rule 6 in the 5th edition of the rulebook.

There is no short answer to this question. Given that the ratings are calculated on the info in the USCF data base, and not what is on the wall chart, the best anyone can do is estimate a ratings change anyhow.

So, keep using SwissSys to estimate those changes, even for a 5 game provisional rated player and his/her opponents. The best you can do is calculate the provisional’s new rating. Put it into the proper field in SwissSys. Then ask the program to estimate the rating changes.

This will all probably be moot in a few months anyhow. Why? Mike Nolan’s on-line tournament submission system will instantly rate any event if the proper hoops are jumped through (rating fees paid, all players current members, …).

Tim Just

7

So Tim,

There’s no magic formula based on a players provisional games would be a factor when figuring out a players rating?

eg:

1200/5 beats a 1600

normally 1200 + 32, 1600-32

but since 5/25 = 20% then

32 * .2 = 7 so 1600-7 = 1593

I was thinking it was something like that.

8

The ratings formulas are available on the USCF website at

math.bu.edu/people/mg/ratings/rs/rs2.html

Reading that paper is not a job for the faint-hearted or the mathematically challlenged.

The first stage of the process is to compute an interim rating for any unrated players. There are several applicable rules for that process.

However, what it all boils down to is that you don’t gain much by beating an unrated unless he does very well against his other opponents. If that was the only game both of you played in the event, you would gain one point.

BTW, I don’t see anything in the ratings formula that would keep someone rated 2199 from making Master by beating someone 400 points below him. It may say that somewhere, but I didn’t find it on a quick read-through.

9

thunderchicken:

Do not worry so much about one game. Thinking what you’re rating will be after only one game with a provisional player at 1200. The problem that happens he could been at a number of tournaments that have let to be rated. With a 1200/5, he could have a rating between 100 to 2000 as the player is still a provisional player.

Being in a tournament you should have been in more then one single game. The person that you lost to could have been in a number of tournaments that need to be rated, it is the same also with yourself. There are so many factors the final result is just let the tournament be rated.

10

That’s

math.bu.edu/people/mg/ratings/rs/rs2.html

Regards,
AJG

11

I do believe that thunderchicken was wanting us to “assume” that the provisional player’s rating was current. Furthermore, I am certain that thunderchicken is aware that more than one game is played in a tournament.

Your reply doesn’t address the premise of his question.

12

It has nothing to do with my rating. At my tournaments, I use the MSA ratings to get a more accurate rating. Sometimes I’ll get kids who want to know about this and I really don’t know how to answer it.

So if someone is 1250 based on 13 games, I’ll still put them as 1250, which makes them eligible for U1400 prizes. But if a established 1200 beats them, since they are provisional, they wont get the full amount.

I’ll take a look at that read.

Thanks Guys

13

This is an ok try for an estimate. It is no better or worse than many others I have run across; however, it is not the exact formula used by USCF.

Tim

14

So that’s somewhat how it works? You can take a percentage based on the number of games and relate it to the established rating system?

eg: 5/25 = 20% of normal points
15/25 = 60% of normal points

I guess this would make sense, just as an estimate since all that derivatives and calculus crap I’ll leave to USCF’s computer to figure out :slight_smile:

15

thunderchicken:

Was thinking you were talking about one of you’re games. There are two ways you can do this at you’re tournaments, but if you do use the second one you have to inform the players that it is “web”.

Can use the official rating of the player, like the August 2004 supplement (used from 8/1/2004 - 9/30/2004) October 2004 supplement (used from 10/1/2004 - 11/30/2004). What the official ratings are during that time would be used for the pairing.The second is this, you could use the web for the most current ratings. During the tournament you must inform the players and should make this posting on the flyiers.

As a persons official rating is 1599 and the web rating is 1605, the tournament has a under 1600 prize and a under 1600 section. If the ‘tournament director’ use the ‘official rating of 1599’, the player could win the under 1600 prize or be in the under 1600 section; if the ‘tournament director’ use the ‘web rating of 1605’: would make the player unable to win the under 1600 section or be in the under 1600 section.

As you do have scholastic players in the tournaments, they would have more provisional ratings. The second point that they only earn trophies then cas prize funds. For the scholastic tournaments would use the web ratings, as some would not have any official rating only the web rating. Would still need to make in the flyiers and before the tournament that you are having the web rating for the tournament.

If a tournament director use a web rating for pairings, the tournament director would have too use the web rating for the prize reward. If a tournament director use the official rating for pairings, the tournament director would have too use the official rating for the prize reward.

16

Do you actually read what anyone ever writes?

17

There are other people that want the same answers as yourself thunderchicken. The fine points with what is different with web ratings and official ratings are not always the same. If a tournament director use one then must also use that rating during the prize reward.

18

I post in my TLA’s that I use the Web Rating. This has already been discussed before.

19

First of all, you should replace the 32’s in the above with whatever the player’s K-factor is. K-factor is generally higher for lower-rated players, and lower for higher-rated players. Also, it’s generally higher for provisional players than for established players.

Second, the rating change enjoyed (or suffered) by the established player (1600 in your example) does not depend on whether the opponent is established or provisional. The 1600 would lose (or gain) the same number of points from a 1200/5 as from a 1200 established.

The 1200, however, would undergo a greater rating change if he is provisional than he would if established, because his K-factor would be higher.

There’s one wrinkle, however. An established player’s rating change will depend more on the provisional opponent’s post-event rating than on his pre-event rating.

All of the above is an approximation (see the websites suggested by others for the exact, complicated, mathematical truth). Any error caused by approximating, however, will probably be less than the error caused by using everybody’s last-published ratings rather than current ratings.

You might want to take a look at the 5th edition rulebook, chapter 8, “The USCF Rating System”. The discussion there is quite a bit simpler (albeit slightly less precise) than the formulas embedded in the USCF computer. Back in the olden days, when all K-factors were either 16, 24, or 32, only three short tables (see pages 266-268) were necessary. I used to distribute these tables on three separate business-card-size documents which players could carry in their wallets. (Somebody told me these cards became known as “Smythe cards”.) Now, it would take more cards, as well as a “master card” to determine your K-factor, like the table on page 265.

When a newbie asks how his rating is calculated, I think it’s OK to just shove the appropriate table (from pages 266-268) under his nose and let it go at that. It’ll satisfy him, and be close enough for practical purposes.

Bill Smythe

20

I think putting that in their face might be unfriendly since I never did get a math minor in college :slight_smile:

I think the information provided can give a logical explination as I would put it:

“When playing a provisional player, you can take a percentage of how many games they’ve played versus the number required for an established rating, and use that as a factor as you would with any established player”