MSA Ratings

Running Chess Tournaments
1

The MSA database is a big advance over the old online ratings pages. However, I think there needs to be some clear indication of which rating is official, as there is on the old pages. The October supplement has been posted on MSA for a week or two and is listed as “Current” and “Last Published.” I have gotten several questions from scholastic players and parents wondering why their tournament rating is not current.

Some questions are inevitable no matter what, but in this case the (currently official) August rating is not visible anywhere without clicking on the Rating Supplement tab, and even then there is no mention of “official” or effective dates.

Bart Mebane
Virginia Scholastic Chess Association

2

I believe you have to officially go by the printed supplement. I run tournaments which from what I understand, I’m required to use WEB in my TLA if I choose to use MSA for my tournaments.

We have our state championship coming up this weekend, which is a few days before the October supplement, yet, do you use the July or the October supplement?

What about updating SwissSys with the latest Membership information, but without the new ratings?

Some players who register for this tournament could be out of a rating specific tournament if you use MSA or October.

Good question though, here’s more information:
uschess.org/forums/viewtopic.php?t=77

3

Our tournaments are too large to enter ratings manually, and I always use the current downloaded supplement in Swiss Sys. I do update new ID’s and expirations from MSA (but not ratings). There is no confusion on my part (at least about ratings :slight_smile:. The problem I was referring to is that players and their parents also look at MSA. They want to know why the rating I use at the tournament is not the one they see on the web site. I can explain it to them, of course, but if the current official rating were clearly labeled on MSA, there would be much less confusion.

Bart Mebane
Virginia Scholastic Chess Association

4

There has been talk about having a more frequent download than what’s on MSA, which I believe is a work in progress.

I think how I’ve explained it was that it wasn’t a published rating and is not official until the official supplements are released.

You might also point out that temporary ratings have been on 64.com for years.

5

I think you hit the nail on the head. The October rating is now listed on MSA as the Last Published rating. If they would list October as Current, but keep the August rating as Last Published (until Oct 1), then it would be much easier to explain to people. Thanks for the replies.

Bart Mebane
Virginia Scholastic Chess Association

6

After you go to MSA and click on the name, you get a summary which includes “Current” and “Last Published.” “Last Published” is the last published Supplement (or Annual Report, in the case of December) and that is the Official rating. But…as it is the case now, the October Supplement ratings are already in the “Last Published” column, although they are not effective until, as you know, 1 October. “Current,” in the case of those who have had activity since the Supplement rating, is the unofficial, although at times, if you click on “History” you might see something even more recent. But, again, the official rating is the Supplement rating – every two months.

Obviously, for those of us running big tournaments, it is logistically impossible to check the latest unofficial rating at the tny site, so we can’t. Even with small tnys, we may not have the internet capability on site. So…that is why, I tell parents or players, we must be consistent and go by Official ratings. I also tell them, if I have the time, that the Official rating used is for pairing purposes only, but when the tournament is rated the USCF computer will use the latest rating on file, and apply it to the appropriate formulas.

As far as your State Tournament, Thunderchicken, since it is a few days before 1 October, you must use the July Supplement. Ours starts on 1 October, so I will use the October Supplement.

Yes, you have to put “WEB” on your TLA, if you want to use the latest Unofficial ratings, but be careful and be consistent.

7

I think that’s what we’re going to do. Terry (evansvillecc) pointed it out the other day to me about it.

Would be nice to be able to use the latest if the player wanted it, but that opens up a lot of worms, especially sandbag ones.

8

Here is the other problem if you allow players to use the Unofficial rating – In a tournament with Class prizes, a player could be a, say, 1601 in his last Official, but a 1590-some in the latest Unofficial. He was a Class B player, but now wants to compete for the C prizes. Or it could be the other way around. No telling. So…keep it simple and keep it fair.

Now, going back to scholastic tournaments, and I have run many with as many as 700+ players, as well as being part of the TD staff at National tournaments, I suppose you could, for Unrated players with Unofficial ratings, use the latest web ratings, if so advertised. Actually, a TD has a certain amount of latitude in assigning a rating for pairing purposes. In adult tournaments, with big prize funds, you are likely to find, as I have, players with FIDE ratings or ratings from other countries. Now you have more decisions to make and assign ratings using the appropriate conversion. The botton line, again, is that you need to be consistent and fair. The last Official rating rule makes it both fair and simple, and would apply to most players.

9

I have a mickey-mouse ratings-lookup program (which I wrote in DOS-based C) which allows either of the two most recent ratings to be looked up. For example, if the October supplement is available but it is not yet October 1, I can look up either August or October ratings. This allows me to use August ratings most of the time, yet I can still use October ratings if the player would otherwise be unrated, or if the player specifically requests the later rating (I would do the latter, however, only if his rating has increased, not decreased, since the August version).

Bill Smythe

10

I think it’s best to use not-yet-official ratings (e.g. October supplement ratings before October 1) only if the player would otherwise be unrated, or if he specifically requests it. In the latter case, furthermore, it should be done only if using the later version would increase his rating, not decrease it. And in the case of unrateds, use some discretion – sandbagging is possible here, too.

Bill Smythe

11

A draft version of a new standard for ratings supplement files is under discussion with the pairing software developers and the MIS committee.

It combines the bi-monthly supplement format with the Gold Master format. This would enable us to provide updates for new and renewing members even if their USCF rating has not changed.

Whether we would go to something more frequent than bi-monthly ‘official’ ratings is probably more of a Rules/TDCC issue.

I think the concept of a published rating is becoming archaic, though I do agree that using only the current rating could lead to other problems, especially those dealing with advance planning for events, such as the national scholastics. A school needs to know well in advance of the nationals which players can play in which sections.

Changes to MSA are also being worked on, but there are probably going to be some database structural changes made in conjunction with switching over to the new ratings system, so I don’t really expect any changes to MSA to be released ahead of then.

I think for active players the suggestion of showing the two more recent published ratings is a good one, but it probably doesn’t make much sense for inactive players.

I’m also considering adding the peak rating over some time frame (such as the last 12 months) to the data file. I have suggested out on rec.games.chess.politics that big money events, like the upcoming $500K tournament that Maurice Ashley is behind, might want to use a peak rating rather than the current or most recent published rating, as that would virtually eliminate sandbagging by established players.

However, Bill Goichberg believes that most claims of sandbagging are for rapidly improving young players or for unrated players, often foreign players. The data tends to support this.

12

The ‘tournament director’ should use the supplement History, not the web rating unless stated in the tournament information. It is true that scholastic players and the parents of the scholastic players: will be asking questions on (ratings, ect.) a number of topics during the tournament. Having a scholastic tournament, the players are on a different level then a tournament full of seasoned established players.

At this time, the MSA does have the October supplement history, the players general information being the (2004-10) last published. Anyone that has been around tournaments, would know they are and will only be official from October 1, 2004 till November 31, 2004. Directors should use the official rating from the time they are official – not before or after. It could help in some way, place the time and era they are official as the official rating in the MSA.

13

From the August Rating Supplement:

Which Supplement to Use
“Unless otherwise announced, ratings from this list should be used between August 1, 2004 and September 30, 2004. Beginning October 1, 2004, the October supplement should be used”.

I think I was just saying that you wouldn’t want to force a player to use the rating from the October supplement.

If they have a published rating which is higher than that of the August list, and THEY wish to use that rating, then it would be your judgement (or Gary’s) on whether or not to allow it. It would be okay if you did. Just don’t allow them to use a lower rating, would be my suggestion.

14

Hallelujah! – especially since I suggested this idea (or, more likely, was among several who suggested it) in the first place. (Blush.)

Hallelujah, again (and the same comments following Hallelujah, too).

Ratings software may want to implement this idea by showing either ALL the latest ratings or ALL the second-latest, and allowing an instant toggle between the two. This could keep screens from becoming cluttered and confusing. (That’s what I do in my mickey-mouse ratings-lookup program – Blush again.)

Another great idea (this time not mine).

Bill Goichberg may have a vested interested in believing this. In any case, SOME sandbagging complaints, even if the minority, are legitimate, and peak ratings should be a good way to fight the problem. For some reason Bill G. has always been more enamored of the once-tried “norm” ratings than he is of published peak ratings.

Bill Smythe

15

Here’s an insane thought that would virtually abolish all sandbagging ever. I’m not arguing for this thought’s popularity, just that it should exist in the dark corner of people’s minds if cheating ever grows out of hand (doubtful).

In a nugget: don’t used advertised rating classes for the purposes of prizes. Everyone should show up because they like chess, not because their xx99 rating guarantees them a great shot at a prize.

A basic computer program could be implemented after the end of the first round witnessed by all participants who care. This program, as an example, would average the ratings of the top let’s say, 10% of all participants. Then a random factor of ±, let’s say 50 points, could be added to that number (for total fairness sake) and then “classes” are determined by blocking people off by a predetermined interval chosen by the organizer (I’m assuming the USCF class default of 200 rating points would be the most likely choice).

So if the top 10% at a small tournament are rated 2150, 2100, 2000, and 1950, the average would be 2050. Then, if you wanted this option, add ±50 to make totally random. Let’s pretend we just kept the 2050 for this exercise. Then classes would emerge at 1850-2049, 1650-1849, 1450-1649, 1250-1449, etc.

If a player ends up being the only one in his class, it could be pointed out that it could have easily happened had we used the traditional USCF rating class boundaries as well; he just had some odd luck.

Obviously, this idea would be unpopular because people don’t like randomness. However, the mathematical conception of the idea has a lot of merit. The details can certainly be improved upon by someone more intelligent than myself. Selling its fairness though would be hard.

If you have any critiques please ask yourself beforehand if they don’t also apply to the current system. In any case, I hope you enjoyed the little contribution. Ben Bentrup

16

This is one of those ideas that is so good, that it never stands a chance of becoming reality.

Play in a tournament because you like chess?? Sacrilege! Not knowing in advance which section you’ll be in, so you can’t sandbag?? Heresy!

Actually, the whole idea could be simplified a bit, by just putting the top N players in the 1st section, the next N in the second section, etc. N could vary from, perhaps, 50 for a very large tournament, down to 4 (quads) for a small club tournament.

You’ll never know whether you’ll be the top player, the bottom player, or somewhere in between, so sandbagging will be pointless.

But that spoils the whole idea, doesn’t it? An xx99 player has the God-given right to win his section, right? :cry:

Bill Smythe

17

Thanks for the kind words…I’d be somewhat careful about making any decisions before the first round though because that might influence some of the standby players on their decision on whether or not to join the tournament.

 1832 player: "Hmm, looks like the cut-off for the top section will be 1834, I guess I will join after all, tops in my section.
        [***A pre-registered expert rated player calls in sick.***]
 1832 player:  Wow, lucky me...I almost joined up there.  I would have become the lowest one in the top section.
        [***A 1900 shows up at the last minute.***]
 1832 player:  Whew, that would push me back to the 51st seed again.

Such sectionmongering even if at the level of educated guessing should not be allowed.

So a random process, visible to all who care, after the first round seems best. The random process, as you suggest, could be simply picking for a large tournament a number between 40-60 and blocking players that way. Again, better minds than mine can say.

Obviously, late players in the top half of their new “class” might be denied class prize eligibility or charged a double entry fee or moved up or something. Another rule is that players ratings would not be allowed to challenge their rating once that random drawing should have occurred.

Again, the benefits of predictability are many, and thus I don’t expect much or encourage much. Still, beating sandbaggers is a worthy cause!

Ben Bentrup

18

Or this example:

1832 player: Hmm, registration just closed. Wall charts going up, what I’m the bottom player in my class? Hey, what is that? My cell phone ringing on the other line with my long-lost cousin wanting me to go babysit his chihuahua for the day? Oops, organizer, gotta go…refund please?

:confused: Again, determining classes is best done after the first round and when all money is permanently collected.

Ben Bentrup

P.S. 1831 player: What? You want me to move up a section? Oh wait, I suddenly got this toothache…

19

I see nothing wrong with a person who is playing in the UXX00 section when he is rated XX99. One of the reasons a person goes to a chess tournament that offers prizes is to win the prize (hence the need for a prize). Everyone in the tournament wants to win the prize. If the XX99 player gets beat good. If the XX99 player wins good. The other players are doing the same thing. The goal of all the players in the section is to go after the prize.

Why join a tournament if your goal is to lose? Sure the XX99 player could probably play up and even win a prize in that section, but let the XX99 player choose that for himself.

I think the decision the XX99 player makes depends on his confidence and leadership abilities. A leader, of course, would play in the higher section. Whereas if he had no confidence and lacked leadership he would play in the UXX00 section.

Not everyone is meant to be a leader and if they are, they will make mistakes along the way. So let them make the mistake of playing in a lower section so that they can learn. If you force the decision on them they won’t learn anything.

Chess teaches more than just how to play chess. Just like everything else in life, what you do outwardly will show who you are inwardly. Chess can expose the different skills you have or it can show what skills you lack.

James Riker
chess@jriker.com