I would like to see a "Title Supplement"

All Things Chess
1

I really like the concept behind the relatively new “titles” that USCF publishes. However, I find that many players do not even know that they exist. I would like to give them a little bit of emphasis at my tournaments by giving out some sort of trinket/medallion/recognition to anyone earning a new norm at one of my events.

However, to do that, I need to know what titles they have already achieved, and that information is not found in the ratings supplement.

So, I would like to that information added to the ratings supplement, or, since doing so might break WinTD, I would like to see a separate “title supplement” published. The title supplement would include ranks achieved and the number of norms achieved toward the next rank. That way, a tournament organizer could know if a player’s performance would earn a new norm, or a new title, and that could be recognized at the post-event awards ceremony or publications. Of course, there would be a few week lag, and slight adjustments to players’ ratings between supplement publication and the actual tournament could create situations where a director/organizer might show a norm being earned when, in reality, it wasn’t, but that would be a small price to pay, in my opinion.

[i]System Administrator’s Note: The number of norms one has will have no bearing on whether or not a specific event earns a norm at any given level, that’s based strictly upon how the player does relative to how a player at that norm level is expected to do against those players. To cut the math short, the player must have a score MORE THAN A POINT HIGHER than the expected score for a player at that level. In other words, if a 2000 player would be expected to score 3.25 points against players with those current pre-tournament ratings, then the player must score more than 4.25 points in order to earn a Candidate Master norm from that event.

Using the pre-tournament ratings available on MSA and the spreadsheet that is available at glicko.net/ratings/normcalc.xls will give a fairly accurate indication as to what norms would be earned before norms are computed and posted, usually on Wednesdays after the weekly rerate. Using current published ratings will be slightly less accurate. [/i]

As an aside, WinTD seems ubiquitous, but it also seems to be getting a bit dated. The instructions refer to using floppy disks. Isn’t it about time to change the format of downloadable ratings supplements to be a little bit more friendly to modern computer systems? I know I’m in the process of creating my own tournament software, and reading in the supplements has some interesting quirks that just seem to be artifacts of old floppy disk ways.

This comment should probably be sent to Tom Doan at Estima, the USCF has no control over WinTD documentation.

2

As far as the documentation for rating supplements goes, I just load the golden database for the month that I need and then point the USCF directory to it. I stopped doing the monthly update a long time back, losing the provisional number of games but gaining any membership expiration date changes for players that had not yet played games hitting a supplement since those changes.

As far as titles go, I can see where it might be nice to recognize a new (and significant) norm during the awards ceremony. That would require knowing how many norms have already been attained (so you don’t give an expert norm recognition to somebody that already has at least five - though a not-yet-rated event may affect that number) and what score is needed to get a norm (only an estimate even if you are calculating using the latest ratings instead of the supplement ratings - again due to some not-yet-rated events that may affect the calculation).
If you want to ignore the potential effect of not-yet-rated events, then you could simlpy use the rating estimator on potential candidates and see whether or not they scored a point better than expected. For example, if a 1750 player played a 1675, 1700, 1725, 1750, 1775, 1800 and 1825, then going 3.5-3.5 would gain zero points and scoring 4.5-2.5 would be that point better needed for a norm. Since you really need to look at the class breaks though, that calculation is irrelevant. You instead look to see that an 1800 playing that particular field and scoring 3.991-3.009 would break even and thus a class A norm would require that 1750 scoring at least 4.991-2.009 (or 5-2). If you don’t want to unnecessarily play with fractional numbers then you see that an 1800 with those opponents would lose rating points going 3.5-3.5 and gain (1) rating point going 4-3, so 4.5-2.5 is not enough for a class A norm while 5-2 would be enough.
If you have a small tournament then you may be able to identify norm candidates, check their MSA for norms, and do those calculations, but it would be tougher for a large event.
Also, a starting player (or foreign player) will not have any norms to start with and even a poor result will give a new norm.

System Administrator’s Note: A player’s current rating is irrelevant when computing norms, norms are computed against the rating for that title level, so a Category 1 norm is computed as if the player was rated 1800, a Candidate Master norm as if the player was rated 2000, etc. We compute norms for all players in regular rated events, and those are shown on the crosstable for the event, but once a player has earned a title, only additional norms towards titles the player has not yet earned will show on the player’s norms detail page.

3

I’m willing to ignore the effects of not yet rated events, and I know how to do the calculation, but you hit the nail on the head as far as the missing information goes. I need to know if someone has earned five category 3 norms before I congratulate them on getting a category 3 norm.

4

When I had previously read the norm information at glicko.net/ratings/titles-0910.pdf, I had taken the term “Y -rated player” to indicate a player of rating Y, not as a hypothetical player of Class X, rated at the bottom range of that class.

Reading it again in hindsight, I see that was the case! This was an error on my part. Thus for a player “below” a class, getting a norm was more difficult than I had anticipated, while for a player rated within the class, getting a norm was easier than I had anticipated (because in each case the rating used for the calculation was either higher or lower than I had anticipated.)

I would suggest that we update that piece to clarify that term a little more. (Now that I see it, it makes sense!)

Feel free to make suggestions to Mark Glickman, mg@math.bu.edu.

5

The WinTD documentation is fine. it’s the rating supplement documentation that needs a little help. Here’s what it says:

October 2011 USCF Rating Supplement

To use this supplement, extract the files tadskcnt.dat, taratsup.cvr and
tarsfle1.dbf to one floppy disk.

Insert the floppy disk into the computer and use the import feature of your
tournament administrator program to import the data to your system.

…

What I would like to see is a description of what is actually in it, and no reference to floppy disks. (I have no idea what is in the .dat or .cvr files, but I don’t seem to need them.) My tournament administrator software won’t be using floppy disks. Right now, I’m having to reverse engineer the contents so I can import them into something I can use. (And for you computer geeks out there, that means I figured out what connection string to use, and imported everything into an ado.net dataset, and then I can manipulate it to my heart’s content.)

6

The FAQ on TD/A has file layouts for the two rating supplement DBF files and a link to the documentation for the 3 DBF upload files, though we hope to have a new upload format later this year. (Assuming it will be an XML format, it will be interesting to see how big the upload file for an event like the USATE or World Open would be with color information included.)

7

Thanks. That heps.