From what I have read and understand, the minimum rating will be 100.
It is the 2 digit, less than 100, ratings that will not be used.
Those with existing ratings that are 2 digit, under 100, will still exist and will not be reverted to UNR.
This action will only affect those just starting, once this is in effect. These new players will need to earn at least a 100 rating in order to get out of the UNR category.
Mike:
I am willing to accept this rating change. But, I would like to see that a tournament director does not get the entrants credit with anyone with a sub - 100 rating. As I do feel there are a number of tournament directors that use scholastic tournaments to pack the tournament to get the norms and entrants numbers to become a candidate for the next level of certification. Anyone that gets a USCF rating of 100 was not ready to play over-the-board tournaments. The organizer and the tournament director when dealing with someone that has never been in a rated should have some contact with the parents and the coaches before the tournament. If the tournament director does not get the entrants credit with the sub-100 players, maybe the tournament director will talk with the scholastic organizer and the coaches to get the scholastic players that are ready for an over-the-board tournament.
plus the NEGATIVE ratings…
I think the proposed cure is approximately as bad as the perceived disease. Not much better, and not much worse.
The real problem is that the 100 floor is too high. (If it weren’t, few players would run into it.)
Dropping the floor to 10, or 1, or 0, would not make much difference. How much better is a 100 player than a 1 player?
The only reason to have the floor at all is to prevent the ratings from going negative. Another way to do this would be simply to raise the scale. Add 1000 points to everybody’s rating, on a one-time basis. That way, the current floor of 100 becomes 1100, which can then be lowered to the new 100 (formerly -900).
Of course, such a proposal would never fly, as people are accustomed to the current scale. So another possibility would be to allow ratings to go negative, say to -900. Trouble is, negative ratings are unattractive, and a rating of exactly 0 might be confused with unrated, depending how the software is written.
So it’s probably not worth it one way or the other. The proposed change is insignificant at best, and could be damaging in some ways, as John Hillery has pointed out. As for any possible inflationary effect from the current 100-point floor, there ought to be better solutions, such as reducing bonus points.
The proposal would create discontinuities, too. A player with an initial performance of 99 who then defeats a 200 would suddenly jump to 600 (based on 1 game), whereas a 100 player in the same boat would jump only to 200 or so. Something is wrong with a proposal which causes a 1-point pre-tournament difference to result in a 400-point post-tournament difference.
Bill Smythe
Well, of the 4531 players in the three Spring Nationals this year, only 46 had a pre-tournament published rating of 100, and about half of those were players who had lost all of their previous games. By comparison, there were 623 unrated players in those events.
When we’re discussing negative ratings we must be at the end of chess history! The “right” way to handle it is to allow ratings to go way negative. That doesn’t look very nice, but here’s one way to make it look better.
We could have “reported ratings” and “actual ratings”. Down to some level (say 100) the actual rating would be reported. Below that we smoothly compress the actual rating up toward the threshold so that the reported rating never goes below zero. Here’s a way to do it:
Let “actual” ratings go negative with no floors. Use them as usual for rating calculations, with no floors. If “negative Bobby Fischer” manages to achieve an actual rating of -3000 by losing 1000 games in a row against people almost as unfortunate as he is, so be it. (But his actual rating of -3000 would actually be reported as zero: see below).
Above the threshold, the actual rating is reported.
Below the threshold, transform actual ratings into “reported” ratings:
reported rating = (100 / e) e^(actual rating / 100)
where e = 2.718…
Replace the 100’s with any other threshold value where you want the compression to start, say 500 or even 1000. Note that this formula matches both value and slope at the threshold, and that as actual rating approaches minus infinity, reported rating approaches zero.
actual rating = 100 [ ln (reported rating) - ln (100) + 1 ]
where the ln is the logarithm in base e (natural log).
This may appear to cause a rating inflation, but that’s only true of reported not actual ratings, and so any effect is entirely contained below the threshold. When someone rated 120 plays someone rated 10 (reported rating), he has a lot at stake because the actual rating of that 10 player is -130. This should also reflect the actual probabilities in the game; the 120 player would beat the 10 player practically all the time, because it’s very hard to get a reported rating as low as 10, i.e. an actual rating of -130.
Actually, I would say that the real problem is some kids playing rated chess before they are ready to play in a tournament. I remember a 2nd grade child calling me over every other move to her game to ask me if it was checkmate. (Even when the king was not in check!!)
When you get people playing who don’t even understand how to play, then the ratings get meaningless. I agree with Mike that if the purpose of the ratings is to have a chance to anticipate future results, then having ratings initially stay ‘unrated’ until there is a meaningful result makes sense.
Who knows, it might have a response that kids will try harder to learn the rules in order to establish a rating.
mrribby - I would be one of the ones who say that a sub 100 player has no business in a national scholastic. However, if their rating is increasing rapidly, then all it would take is one tournament before it with a “ratable performance” and they would not be unrated anymore.
That’s the main feature of the proposal, Bill. The point is that most players rated 0100 will, at some point in time, instantly become 0600 players - we just don’t know when that will happen. When it does happen, we want the rating system to correctly estimate his performance in future events.
Yes, Mr. Getty, but the only way to have your rating be used for the nationals, based on the cutoffs for official ratings, is to have the rated games be played about three or four months before the event…The increased performance may hit during this timespan.
I like Artichoke’s logarithm proposal best. I was about to post a similar suggestion, until I saw his post.
Even better, set the cutoff (below which logarithms take over) at 300 or 500 instead of 100, and offset it by 100 so that the lowest possible “reported” rating is 100 rather than 0. I.e., as the actual rating approaches minus infinity, the reported rating would approach 100.
Bill Smythe
I agree with this, but maybe I’m not understanding why it’s a problem to have a bunch of players on the books who are rated at 100. Someone care to explain this? Or, put another way: why is it a problem to have a lot of kids who get a rating of 100 and then don’t renew their memberships? They tried it, they didn’t like it, they moved on.
It would be more strange to me to have a large bunch of unrated players. As a player, it would be more strange to not earn a rating right away, even if it is a lousy rating.
I agree that holding unrated tournaments for inexperienced kids is a great idea, and we happen to have a pretty good system for that in our community. But not all communities do. While many of you would like to see kids get their experience playing in unrated tournaments, there is no mechanism in USCF for supporting unrated scholastic tournament activity, and no financial feedback loop to encourage USCF to support it.
If you have a player that is UNR, it does not change the average rating of the whole membership or the average rating of an age group. If the USCF reclassify new members at the rating of 100 to UNR, it will inflate the average rating of the whole membership or the average rating of an age group.
I’m starting to like this proposal more and more as I think about it.
How about this? A kid comes in new, goes to a few tournaments, never wins a game, and gets a 100 rating.
Same kid plays in his school club for a while, actually LEARNS how to play, and now is playing at an honest 900-1000 level. (Which is still a fairly weak player, but a reasonable spot to be at as a rapidly learning new player.) This kid gets to go in and get “upset prizes”, play and win in sections that he shouldn’t be playing in, and hurt the ratings of other players who get a loss against a 100 player against their rating.
I haven’t seen a lot of tournaments with “Best under 300” prizes.
My daughter has played in a couple of tournaments with U300 prizes. Maybe you just haven’t seen as many tournaments with K-1 sections?
On our MLK day tournament this year, we will be running an Upper Primary U300 section and an Elementary U500 section. We do this for a couple of reasons. One, in our 2-3 grade sections and 4-5 grade sections, we routinely get 90 to over 100 players. By splitting the sections at U300 and U500 we get about equal sections and hopefully will get a clear winner in each section.
You’re a life member with a 1700 rating and based on your USCF ID you have been a USCF member since at least the 1970’s. You’re not a TD, so you aren’t likely to be at scholastic events. How many tournaments with ANY players rated under 1000 have you been at in the last 25 years? 
Is this better than losing to an unrated player, other than s/he gets to go in and win the “best unrated” prize instead of the “biggest upset” prize?
What hurts a rating more: losing to an unrated kid, or losing to the 100 rated kid?
Right now “unrated” means “unknown,” which is not the same thing as “lousy.” My admittedly unscientific sense is that in scholastics, the 100-rated kids are more predictable opponents than the unrated kids. Under this proposal, you’d be putting some people with known weak strengths into the batch of unknowns.
I still don’t understand what good the proposal does the USCF, or USCF players.
I’m not totally unaware of scholastic events, but ours in GA have been grade based, not rating based. But mnibb’s specific example and tanstaafl’s experience is certainly relevant, so I yield to their superior expertise.
BTW Mike, you cheated me out of 100 rating points, based on my unpublished rating and even my published rating which is closer to 1800. 
And I’ve been to a fair number of events with players under 1000, either in my section or in others.
Sorry, I should have said you’re a player in the 1700’s.
I thought about some kind of logarithmic rescaling to avoid ratings going negative (it would also probably prevent ratings from ever getting to 100), which I would probably start at 300 or 400, but I would defer to Bill Smythe’s math skills. (He was a math major at Northwestern, I was in EE/CS.)
I don’t know if the Ratings Committee looked at that possibility. Probably not, since the chair tells me they consider their proposal a cosmetic one rather than a substantive change in the formulas.