This, I believe, is Mr. Smith’s point, or at least includes Mr. Smith’s point. Until recently it was possible to generate an ID number for a non-member for just this reason. Unrated players show up to tournaments all the time, and some sections are made up entirely of unrated players. There must be, and in fact is, a method of assigning a rating to players about whom little is known. I’m told the method we use to assign initial ratings to unrateds is actually pretty accurate.
A thought experiment: someone with even more money than Rex Sinquefield bids and wins next year’s U.S. Championship. Prizes are ten times what they were in 2019, but the organizer wants a wild card for his four-year-old grandson who is cute as a button, but has never played a tournament game. The grandson goes 0-11. What should his initial rating be?
This is what I (as a TD) always do with house players, particularly when the house player has a relatively high rating. I once had the misfortune to be the odd man out in the last round of a 4-round tournament in which I had gone 0-3 in the first three rounds. So they gave me the option of playing a 2100-rated “house player”. Why would I want to do that? The whole methodology of the Swiss system is geared toward progressively more competitive pairings, so that if you’re overmatched in general, you get someone closer to your own strength in the later rounds. If I can’t make any progress vs. Class B players, why would I want to play an Expert? The method Alex describes gives the odd man the same general level of pairing he would expect to get with an even number of players, and that seems desirable to me.
That is a no-win scenario. I believe the nominal answer is “400 points lower than the rating of his lowest-rated opponent” – but in a US championship, this would typically give the newbie a +2000 rating, and this would almost certainly be grossly inflated. But doing anything else would be completely arbitrary. In reality, you have no idea what the newbie’s rating should be (other than “probably a good deal less than 2000”).
I can’t believe how many TDs don’t seem to understand how much better this is. Please, don’t just pair the house player against the player who would otherwise receive the full-point bye. Pair him normally into the tournament, based on his score and rating. For pairing purposes, assign the house player a “bye score” – full-point bye, half-point bye, or zero-point bye – for each round where he did not actually play a game. Assign the “bye scores” by feel, so that the house player has a cumulative score similar to what he would be expected to have achieved if he had played every round. Ideally, a house player’s bye scores should be adjusted up or down so that he ends up being in the bottom half – but not all the way at the bottom end – of whatever score group he lands in.
The house player should never be the highest-rated player in his section. A 1700 would be a good house player in an under-2000 section, but not in an under-1600 section. And a TD should not use himself as a house player in a section where he would be higher-rated than anybody else in the section.
You are correct. It is in the 7th edition also in the same location. I have the Kindle edition and it comes right after the rating classifications, but I missed it because I didn’t turn the page. “All games played in US Chess-rated events are” then it finishes on the next page “rated, …”
Let’s get back on track here. What is being done, if anything, to address the issue that started this thread (which was back in June) and is still an issue now?
The office staff does not have time to routinely monitor the forums trolling for questions that someone wants answered. Contact them by email, the project manager is Boyd Reed, boyd.reed@uschess.org. He does have a few other things on his plate, though.
Well, not every organizer can supply a “permanent house player”, much less one who is appropriately rated. If Mr. Kosterman was playing in an U1800 section (claims he was getting beaten up by B players) it would be terribly unfair to pair the 2100 “normally”. I learned the hard way not to treat an ordinary house player like a “permanent house player” when I paired the same player as a bye twice in the same tournament.
Regarding giving an 0-3 player in an U1800 section a “house game” against a 2100 player, I think there are many B players who would like the chance. I think an Expert spectator who is willing to play a house game against a much weaker player is far more likely than average to be willing to do a post mortem. Sadly, I think there are many more players whose parents would be thrilled to get the chance to beat a 2100 player, the fact that they had just lost to three B-players notwithstanding.
It was actually an open section with a number of experts and a few masters, so the existence of a 2100 house player is not as unfair as it might have seemed from my initial account. However, in my opinion, he should be paired as Bill describes, and definitely not paired against the “forced bye” – which is highly likely to be a relatively weak player. In this context, “relatively weak” perfectly describes me – I was rated 35th out of 36 players in the section, which is why all my opponents were B players. With normal pairings, I’m never going to get any of the experts or masters unless I score an upset or two (or one of them is having a terrible tournament).
My point is that after losing 3 out of 3 games, what I most want in the final game – and what I would be very likely to get with a normal pairing – is a competitive game where I have a fighting chance to win. It’s not impossible for me to beat an expert – I have done so a number of times, although they were all many years ago – but it’s highly unlikely. I suppose there might be some who would relish such a pairing – for the challenge, or the learning opportunity, or simply because they wouldn’t be risking any rating points (that matters to some people – it means nothing to me), but not me, especially not in the last round of an 0-3 tournament. I just want a chance (a realistic chance) to win a game.
The 6th edition of the rulebook, as I recall, recommends that the house player be someone who would have been in the bottom third of the field by rating. I doubt that recommendation has changed lately.
I was thinking that a workaround would be to create unique “e-mail addresses” by prepending the ID nnnnnnnn. to the existing email then stripping as needed. (Of course, the “as needed” might be rather complicated unless there’s a single method for returning the working email from the database).
If someone tries signing up for a new US Chess membership with an email that is already assigned to another member, would it be possible for the programming to not accept the membership and come up with a message saying that the email address is already assigned to another member and you must put in a unique email address?
I don’t know about the current system, but in the previous membership system we tried to keep the membership form from giving out information about other members, because it could be used as a way to mine data from our database.
Saying "there is already a membership that uses the address foo@bar.com is a valid email address that they could send email to.