Are rating floors based only on a supplement rating? Or are they instead based on any end of tournament rating (or even end of game rating)? I see junior players whose unofficial supplement rating is 2-300 points above their supplement rating and is often moving up or down rapidly. If a player has a supplement rating of 1300, peaks out at 1620, but drops back down to 1599 by the next supplement, would the floor for that player be 1400, or would it be 1300?
Peak rating based ratings floors are based on post-event established ratings. Players who do not yet have established ratings (ie, 26 or more games) do not have peak rating based floors.
The lowest such floor is currently 1200, which is the floor someone who has a peak rating of 1400-1499 would get.
However, post-event ratings CAN change as a result of a rerate, and that could change someone’s peak event floor as well.
There are some anomalies in floor records, but for anyone whose peak rating was achieved in 2004 or later, the peak rating based floor should be consistent with the policy outlined above.
And that means a person with a recent 1620 peak rating should have a floor of 1400. Updated floors may not get posted to MSA for a day or longer, but should be effective immediately. (But the odds of someone going from a peak rating sufficient to earn a new peak rating based floor to that player’s floor in a few events are small.)
Note that games are no longer rated, but events, that is sections are.
Alex Relyea
That could be misinterpreted. We do rate games (what else is there to rate?), but we rate all of the games a player has in a section at the same time, even if played under more than one pairing #, rather than rate them individually.
I think what Alex and Mike mean is that tournament sections are rated as a whole, as opposed to game-by-game. So, no “statistical events” (such as a peak rating) can occur mid-tournament.
Got it, that’s helpful. My understanding is that FIDE rates game by game, which matters if you are trying to achieve a minimum rating to earn a title, i.e. 2500 for your GM title once you have your norms. In that case, you can earn the title by exceeding 2500 mid tourney, even if you lose your last several games. I assumed the USCF was the same, but I understand now that that assumption was incorrect.
That may be true for FIDE norms, but I don’t think that’s the case for FIDE ratings.
No Dayman’s right,
from the FIDE handbook for Titles 1.53(a)
“Such a rating need not be published. It can be obtained in the middle of a rating period, or even in the middle of a tournament. The player may then disregard subsequent results for the purpose of their title application. However, the burden of proof then rests with the federation of the title applicant. It is recommended that players receive a certificate from the Chief Arbiter where they achieve the rating level during a tournament. Such a certificate should include a note of the date each game was played. Title applications based on unpublished ratings shall only be accepted by FIDE after agreement with the Rating Administrator and the QC. Ratings in the middle of a period can be confirmed only after all tournaments for that period have been received and rated by FIDE.”
It used to be for the FM/WFM titles that the rating had to be published, but looking at 1.3 and 1.53 that appears to no longer be the case.
Technically, FIDE rates events in their entirety rather than game by game. However, the calculation is similar to the calculation for a player with a USCF established rating. For each opponent, there is a win probability p based on the difference between the players’ ratings. The rating change for that game is the difference between the actual score and p multiplied by the K factor. The rating change for the event is the sum of the individual changes.
In fact, to be even more precise, FIDE rates all events for a month at the same time. That is, if a player competes in more than one event, the sum of K*(W-We) is computed over all the events, and only then is the result rounded. Also, there is no concept of using “post-event ratings.” The rating used for the player is the rating published on the FRL for the month in which the event started.
Having said that, yes, for players who need to reach a certain minimum rating for the GM, IM, WGM, or WIM title, it is possible to demonstrate having achieved that rating after an intermediate round of an event. If a player is counting on this, the player is advised to ask the chief arbiter for a certificate of his results against the opponents demonstrating that the required minimum rating has been achieved. The QC (Qualifications Commision) will verify the requirement has been met.
Nolan’s correct that I was really talking more about FIDE norms/titles rather; and the point is at best only tangentially related to my original question about USCF rating floors (thanks for answering that one, Nolan).
Interesting side note, Kayden Troff achieved his GM title but getting his rating to 2499.5 after the penultimate round of the Chicago Open, but then lost the last round and lost a few rating points. FIDE eventually confirmed that they would round the 2499.5 up to 2500 and awarded the title. Young Mr. Troff did very well at a tourney in St Louis the next month and got his rating well over 2500, so the only question in his case is when he became a GM (not if), but still an interesting application of the FIDE process.