It appears there’s no real time limit for making corrections to posted tournaments. How long do most TD’s keep paper records (pairing sheets, game results, etc.) and/or data files (WinTD, SwissSys or whatever) in order to verify or refute any player claims that an error was made in the posted results or to help settle some other dispute?
Thanks for your input!
as far as swiss sys files. I keep them forever (: When I was doing paper, I kept them until they were lost when i moved.
Hope this helps .
Maybe another way to put this is, how far back has anyone ever had to go in their records, and were glad they still had them? I’m a recovering pack rat and would like to try to get rid of unnecessary stuff sooner rather than later (but please don’t come inspect my home or office for proof of my success!).
The USCF gets reports of errors in someone’s tournament history as shown on MSA virtually every day. We have had reports going back as far as our online records go, which is late 1991.
USCF policy is that TDs are responsible for the accuracy of their rating reports, so if these reports come from players they have to be referred to the TD for confirmation. Moreover, the players may not have enough information to make a correction. For example, a report of an incorrect ID would require the TD to check his or her records to see what the correct ID should be.
As we are still able to rerate events from late 2003, it seems to me that TDs should probably keep that data for at least three years.
As a newcomer to this, I’ve been surprised (shocked, actually), that there’s not a cutoff time for submissions and corrections.
Practically speaking, I guess it’s not that big a deal these days to scan and store records electronically forever (or at least until the storage CD oxidizes). I certainly agree we should be held responsible for the accuracy of our records, and I suppose I could live with 3 years even though that strikes me as rather long. Players should bear some responsibility to report errors within a reasonable period of time, certainly in less than 3 years.
Without a rule to follow, I asked what the typical practice out there was, so I could try to follow the norm. But from another perspective, without a defined rule to be held accountable to, what recourse does anyone have if a TD shreds their records after 3 months? 1 year? 2 weeks?
Is there any data (or reasonable guesses) as to the time period in which x% of corrections are actually being made? Eg, 97% of corrections are made within x months of the event end?
In the corrections process, does the USCF automatically accept/post changes based on the TD’s word, or is there some standard of proof required? Does that standard increase as time elapses? If not, the longer it is before changes are submitted, the more I have to wonder about their accuracy/validity. Even more so when an entire event isn’t submitted for some length of time – I read in another thread about some being turned in more than a year later??
“If it weren’t for the last minute, a lot of things wouldn’t get done.” Seems to me having an official deadline (and therefore a standard to hold everyone to) for corrections and submissions would benefit everyone – players, TDs, the USCF office, and the ratings system itself.
With the MSA players have a tendancy to check on their results as soon as they’re up. For tournaments submitted online this possibly is as early as the next day. For mailed results or large events the lag may be greater, but still the players tend to frequently check to see when it’s up. Players are pretty quick to notice mistakes.
The mistake that may not get noticed is a wrong ID number that causes some player from another state to end out in some tournament that he never played in. For events rated online if the TD is being diligent in checking the validation reports this mistake should be caught before it even gets rated.
However if the tournament was not submitted online or is an older event then these mistakes do come up. I can recall getting an email 2 years after an event from the office saying so and so says he didnt play in your tournament, but is listed as having played. I was able to go back to my computer find out who the actual player was and get the office to fix the ratings of the people involved. Since I do all my reports online, including my chess club’s 4 player quad this no longer is an issue for me.
I keep all my computer records, but the hard copies don’t stay around too long. Believe me when running scholastic tournaments any mistakes are usually pointed out before the TD can leave the building.
Yikes – you have just provoked yet another Mike Nolan 7-column-by-50-line statistical table. 
Bill Smythe
Bill, does that mean I should also post the 40 page demographic study that I did for the Delegates in 2002 and just sent to someone, or the updated version of it I’m hoping to prepare this weekend? Actually, I probably will put it on the website somewhere, perhaps in the ‘sominex’ directory? 
The NY ratings staff was sitting on (aka ignoring) corrections for quite a while, because they did not have an easy tool to make them and they were 3-4 weeks behind on new reports. I don’t know what happened to those correction reports when they moved the ratings department to TN, so I suspect they never got made.
That may be one reason why the TN office has had a steady stream of (possibly re-reported) corrections to make.
I’m hoping that the number of corrections from events in 2004 are starting to slow down. However, last night we got an e-mail wondering why some match from over a year ago doesn’t appear to have been rated yet. :sigh:
It seems to me that the tendency is to allow corrections for mistakes in tournament at any point in the future.
I don’t think this is a good idea. I think there needs to be an official “statute of limitations” on these corrections. The problem is that the TD and the opponent may not have the documentation after several years to prove that the original result was correct if some player requests a bogus “correction”. I also don’t see why the USCF should spend ANY resources to correct a problem that’s so insignificant that the players didn’t notice it for a couple of YEARS. I realize the desire is to have the ratings be CORRECT, but I think after a certain period of time you reach a point of diminishing returns and I think (rather than being subject to the whim of whoever receives a complaint in the USCF office) that there should be an official cut-off.
I think 3 years is too long, but I could live with it (better than no limit at all). I think half that would be better. (After re-reading Mike’s latest post, perhaps the statute could be extended one time for corrections that were requested in a timely manner and then ignored by the office for some reason.)
My feeling is that if the historical record is wrong, it should always be corrected.
However, there should probably be a statute of limitations on whether such a correction impacts someone’s current rating, with an allowance for special situations.
Currently we do rerates for all events that were rated on or after 1/1/2004. At some point we may want to advance the rerate window by a year, to 1/1/2005.
However, when I discussed this with Ken Sloan and other members of the ratings committee a year ago, we were not sure when (if ever) that change should take place, and that discussion led to the block rerating process we use now.
Under that process we could just lower the priority of ‘old’ changes so that those old blocks are rerated less frequently, such as once every other month. (Obviously, if there are no changes, there is also no reason to rerate that far back. The rerate that is taking place this week only went back to June 2005.)
From having worked with the rerate process, I think that even a significant error (say, 100 points), will be negligible after about 100 rated games.
Until Mike’s re-write of the rating system, there was no way to make a correction to an event. Witness this event: uschess.org/msa/XtblMain.php?200012165570 where player #1 should have been IM Ben Finegold (not his son twice 1 & 4). This error was reported shortly after it was rated, but there was nothing USCF could do at the time.
Pete has basically retired from directing and I don’t if he would still have the records from the event. I just know that I didn’t play Spencer twice. I still have the scoresheets from the event if that is proof enough.
Corrections were possible under the old ratings programming, and some were even made, but it could take HOURS to make some corrections. Moreover, some of the corrections that were made were only partially made, which created quite a few problems when we were converting over to the new ratings programmings, because the data wasn’t fully consistent.
Regarding the correction you gave above, I let Walter Brown know about it, he got a good chuckle out of it and will correct the ID. It’s probably a bit too late to make any ratings changes, though.