Background: When I was working on the ratings distribution chart, I noted that we have a very large number of scholastic members with ratings of 100, most of them provisional ratings. I pointed this out to Mark Glickman, chair of the ratings committee, here is the committee’s response:
This approach seems reasonable to me, and after some discussion with Mark on the details I don’t think it would be all that difficult to implement.
Mark does not see the need to make this retroactive, but we would need to pick some effective date, either based on the date the event was held or the date the event was rated. If the Board endorses this concept, we should probably try to have this implemented before the next Ratings List goes out. That will be the February list, which will be created in early January.
I’ve told Mark I would forward his recommendation to the Board, and am sending it to Mike Nietman so that the Scholastic Council/Committee can comment on it as well.
Isn’t there a danger that this will create a “permanent underclass” of unrated players? It also seems like it would be a hard sell to the players – “Play in a USCF-rated tournament and get a rating, unless we decide not to give you one.”
Instead of replacing ratings < 100 with "Unrated," another possibility might be to replace ratings < 100 with the actual 2 (1?)-digit rating? Is it bad to have such ratings floored at 100, or would it be better to lower the lowest floor to 0? Is there a growing epidemic of 100-players in the USCF?
Maybe we should print the player's coach's name next to any rating < 100.
Everyone understands there are a number of scholastic players at the rating floor of 100. In my judgment, someone at the rating floor of 100 has a low renewal of their USCF membership. Is this rating policy change designed to stimulate players at the rating of 100 to renew or rejoin the USCF because their 100 rating was expunged just to get a first tournament UNR? There will be more scholastic players and parents willing to renew or rejoin the USCF with a first tournament UNR then an official rating of 100.
My problem with this idea is this. If a scholastic coach has a student in the 100 to 300 point range, the coach could tell the student to sandbag the rating down to 100. Then come back in a year or two with an expunged rating. With an expunged rating, the player could start over again without the draw back of a rating of 100.
If a scholastic player is at the rating floor of 100, I really do not see the scholastic player really going to renew or rejoin the USCF. Even if the USCF expunge the rating, I really do not see that is going to make a difference with the majority of the players at the rating of 100. What I do see will happen if the 100 rating is expunged, it will increase the average rating of the scholastic players.
Doug, read more closely. It says “remains unrated” meaning you only remain unrated if you were already unrated and you didn’t manage to get above 100. The proposal doesn’t say anything about keeping people at their proper floor if they fall from a provisional or established back to 100. So no expunging or clean slate.
Before it is floored at 100, does the rating floor of 100 become official with a web rating or with the first official rating?
Crankwire, read more closely, is says remains unrated. Are you defining being unrated with a web rating or are you defining unrated with the first official rating – what is it? Players could play in a tournament every single day. An unrated player could play in there first tournament on Saturday, there second tournament in Sunday, then there third tournament the next Saturday. Then come to my tournament on Sunday, so this will make it the fourth tournament. I really do not care if the other three tournaments have been rated or not, even if they are rated they are web ratings. I’m still going too accepted then as being UNR’s, you are free to accept there web rating. So Crankwire there are two different views what is an unrated player.
At the present time if you have one, two or three rated games, the player does not get an official rating. Say for example if you are a new player with one rated game, the USCF accepts that player as unrated till the player has four or more rated games. True, the player has a web rating of one game, but it does not say that rating will be official in fact it will say it is a UNR.
Since Mike Nolan has not made it clear if the rating committee is going to define it with unofficial ratings before it becomes official, or the first event rating, my judgment is it being unofficial ratings before the first official rating. I’m not going to say your wrong Crankwire; it is just that it has not been defined at this time.
If it is just the first event with a rating of 100, it is not that simple to get a USCF rating of 100 with the first tournament. It could happen, and it does happen from time to time. It is just that the scholastic tournaments with three rated games are not official posted. The scholastic player with three or less games will be unrated. It can be some work to deal with the ratings of 100 with three rated games or less. That will be some work for Mike Nolan to get the rated games of three games or less to show up as official. It just looks like a lot of work to make ratings of 100 with three rated games or less to become UNR when they are UNR in the first place.
That could be the most redundant computer software upgrade Mike Nolan would ever have to perform. But, Crankwire if you are correct Mike Nolan would have to write up a computer program to say a rating of 100/03 would be UNR when it already shows the player as being UNR on the MSA. Now that I look at it, I hope you are correct as I would love to see Mike Nolan doing redundant computer work.
[hr]
Mike Nolan:
Can you define when the rating of 100 would be classified as UNR?
What if the rating of 100 is a rating with three or less rated games, as it stands at this time the player would be an UNR player?
Would the rating of 100 be classified as UNR with web ratings only, as anyone with a web rating without having their first official rating will be a UNR player?
If the rating of 100 be classified as UNR when the player should have an official rating of 100?
If I play in a FIDE tournament without having a performance rating to get a FIDE rating, FIDE does not give me a rating. Say someone did play in a FIDE tournament say twenty years ago, would FIDE check the tournament of twenty years ago with the next FIDE tournament to make sure my performance ratings of both tournaments would grant me a FIDE tournament? Or, would FIDE just skip the tournament of twenty years in the past just to handle what the performance rating will be with this one tournament only? As this begs the question Mike, if a scholastic player has one tournament with a 100/03 (100 rating with three games). If that player is just like FIDE, then the USCF would look at the rating of 100/03 when the scholastic player plays in the second tournament or not?
This is what I think the USCF has as the goal Mike. It the player has a rating of 100/03 or 100/10 or just an established rating of 100; the player would be classified as 100. If that is the case Mike, then it would make the average rating of the scholastic players have a higher average rating. As UNR’s are not counted as the average rating of the players. But, if the player still gets the credit of each and every game, even if the player is at the rating of 100: it does not change anything for that player. If that is the case Mike, then it just a cute way for the USCF to inflate the average ratings of the scholastic players. As the 100 ratings would not be a factor as they will be reclassified as UNR.
Douglas you seem to have a problem with about any and all ideas. It seems to me that the majority of them are problems with small possibilities and generally negative. Yet at the same time I give you part of the credit for influencing the rules committee to change the write first move rule to move then write.
No solutions are 100 percent. There will always be loopholes. The way the rules exist now have loopholes. That is why they evolve and change occasionally. And true sometimes a fix will introduce new problems, sometimes worse than the old.
Here is an example of a loophole that currently exists. Say you don’t want to keep score and you are willing to sacrifice some time for it. Say you are on your floor or you just don’t care. You really want to play rated speed chess anyway. But you are at a rated tournament where the time control is Game 80 with a 5 second delay. Maybe you are much higher rated than your opponent or maybe the opposite.
You simply set there and wait 75 minutes thinking about your first move and the follow ups until you hit the 5 minute mark. Now you aren’t required to keep score.
True you are going to irk your opponent but hey it is legal.
And it is a rateable game.
Now what happens if your opponent decides to respond it like matter?
He or she sets there for 75 minutes before responding.
Then you have a legally rated 5 minute game in essence.
The ratings process is a serial process. If you have four tournaments, that have not been rated, they will each be rated one at a time. The rule applies after the rating of each. This rule has nothing to do with what the TD uses for pairing purposes. TDs are supposed to use published official ratings, so if the events have not been rated and published, it’s as if they had never played yet.
Currently, a player needs four rated games in order for the rating to be published officially.
The proposed rule is as follows (in my understanding):
After processing an event in which a player has played his/her fourth game, or a later event if they are still UNR as of their preceding rated event, if the rating calculated is 100 or less, the player does not earn an official rating.
Provisional and established calculations remain the same. The game count increases as usual. The only thing that changes is that the player will show up in the official published ratings as still unrated.
If they manage to finish above 100 after serially calculating any single one of their events, and the game count is more than four at the end of that event, they will finally get a published rating instead of UNR. This could happen on the fourth game, the fifteenth, the fiftieth, it doesn’t matter.
This would I think take care of a lot of kids that only play once or twice and then never play again. It is kind of pointless to have them in the rating pool as official. Their rating will improve with experience, but if they are not interested in getting the experience, then why include them in the reports as rated?
That could be what the USCF rating committee is thinking. If the player is at the rating floor of 100, just to have the player always being unrated or UNR, it just gives rating inflation with the average rating of all the scholastic players. As you have just removed all the players at the rating of 100 off the list. In my view, it is just a fancy way just to cook the books. Maybe the USCF should make anyone under the rating of 2000 as unrated also. That would make the average rating of the USCF membership at an expert level or better.
Maybe what the USCF should do is make the rating floor be at 500. If you are at the floor of 500, you will be an unrated player. So anyone with a rating of 501 will be used to find the average rating. That would make the average rating be inflated, but is not the rating of 100 being unrated or UNR just be inflated too?
Do not want to waste time to repost the whole text, but there is not a loophole in the rule. There is rule 18G1 (page 75), as it says “… is present but shows little interest in considering the position.”
The problem with allowing ratings to go into double digits is that also means they are likely to go negative. (Otherwise we’ll repeat this discussion about the number zero at some point.)
BTW, I believe that at the time it was being debated at least one member of the ratings committee expressed a preference for 400 as the minimum rating rather than 100. (I"m only the liaison to the ratings committee, not a member of it, but this is what I’ve heard from that member.)
The number of players with a published rating of 100 has been increasing.
A year ago we had 3634 players with a published rating of 100 in the last two years (eg, published in 2004 or 2005).
Looking at those who have had a published rating in 2005 or 2006, that number has climbed to 4857.
Anyone rated at 100 is most likely having trouble understanding the rules of chess, let alone playing a serious game.
Is the purpose of this to ensure that if they play again in the future, and they are playing at, say, a 600-1000 level because they have learned the game better, that they will immediately establish a provisional rating in that level that will track in to their real strength, as opposed to adding 30 points a game to their 100 point floor and having their rating slowly (and without accurately indicating their strength) move up?
A player rated 100 has a K of 107.45, so he or she can gain considerably more than 30 points from a win against a higher rated player.
Also most of these players have provisional ratings, and provisional ratings use a different formula that makes their ratings subject to more fluctuation.
For example, a player with a rating of 100/4 who wins a game against a player rated 200 has a new post-event rating of 200/5. By contrast, an unrated player who wins a single game against a player rated 200 has a post-event rating of 600, though that rating would not be publishable since it is based on less than 4 games.
What this change would do is get those players to their true strength quicker, by ignoring those initial results.
Keep in mind the goal of the ratings system is to predict how these players are likely to do in the future. What this change says (in effect) is that a sub-100 result is of no value in predicting that future, especially when compared with the result in that player’s next event. I tend to agree with that logic.
Mike, whatever bottom rating the USCF has, someone is going to hit that floor. If there are at present 4,689 members at the rating of 100, that is your rating. If it is the goal to change these players from a published rating of 100 to a UNR, it is not going to change the facts with these players. It looks as this change is designed to make renewal of there USCF membership look better. How would it make it more marketable to rejoin the USCF, if the 4,689 are reassigned as being UNR?
Mike I would like to understand, if someone has an established USCF rating of 100: how would reclassified as UNR going to alter the players established rating? As the player could after time gain rating points from the floor, or, have the rating drop down to the rating floor. Will the player that has an established rating of 100 get a second change with provisional ratings again? If that is the case, you could see coaches telling there players to sandbag just to get a second change with a provisional rating.
Mike, what would be the average rating of all players with 4,689 members with a rating of 100; then, a second average rating of all the players except the 4,689 members with a rating of 100? My bet that the average rating of the average USCF member will increase with 4,689 members with 100 ratings removed from the list. It looks like it’s a nice way to cook the books Mike.
The RC’s proposal has NOTHING to do with established ratings, or even with anyone who currently has a provisional rating of 100, regardless of how they got it, since retroactivity is not being proposed.
It would only apply upon implementation to someone who was unrated at the beginning of an event and had a rating BELOW 100 from that event. Rather than round that rating up to 100, the player would remain unrated.
If such a player had a second sub-100 performance, that player would still remain unrated. But as soon as that player has a result of 100 or greater, that player would no longer be unrated. (However, depending on how many games the player had in that event, that player may not have a publishable rating yet.)
This is similar to the provision in FIDE ratings that a player who does not yet have a FIDE rating must have at least one point in an event (and at least 3 games) for it to be counted towards a FIDE rating. FIDE also requires that a player have 9 ratable games before they publish a rating for that player, whereas the USCF will publish a rating based on as few as 4 games.
Then Mike, if a player ends with a rating of 100/04 the player would be a UNR, if the player ends with a second event with a rating of 100/08, the player would still be UNR? Does the player four games or eight games count towards the established rating?
If it is not going to be retroactive, how do I define the two standards with the parents with players with 100 ratings from two different but equal ratings? That would affect the pairings if one scholastic player has a 100/04 rating before the change and a 100/04 rating that is labeled as being UNR after the change. As all UNR’s go under the lowest rated player unless the director gives a rating assignment. That could be upsetting to a parent with a scholastic player at the official rating of 100. If I am forced to give a bye, I’m not really allowed to give UNR’s a forced bye in a tournament. But I could have to give a forced bye to a 100 player during the tournament. Oh, that’s going to be fun, thanks.
As I understand the proposal, a sub-100 result would not count towards the games it takes to get an established rating.
This is consistent with how FIDE handles getting the 9 games it takes to get a published FIDE rating. Players without FIDE ratings who score less than a point or who have fewer than 3 FIDE ratable games in an event do not have that result counted towards their 9 games.
I am concerned about players rated 100, whose rating changes to UNR that want to go to the scholastic nationals. Unrateds have to enter the championship sections. The players rated 100 should not be in that section. They wont even attend the nationals.
I know some will say that players rated 100 shouldnt think about going to a nationals…they have a point. But, maybe the player is gaining strength quickly and would like to try it.
