How to answer Tim Just's 1300 choice!

Running Chess Tournaments
1

The question in the supplement, as the answer d is correct. The director can use a) 1200 b) unrated or c) 1300. The problem with the supplement question, it does not take into effect what will or shall happen at any or all events.

If the director use the web rating (unofficial rating of 1200), it will change the roster line up if there is more then one unrated player in the event. It will change the prize, if there is more then one unrated player with a unrated prize. Can understand the drive to use the web rating if the player does not have a published rating. Even with a web rating without a published rating, the player should still have very few rated games. Players can have wild swings in ratings with less then 10 rated games, even more between the first event to the second event. The problem I have with the web rating, if you use the web rating for one person you should use the web rating for all the players. If one player has no published rating with a web rating of 1200/03, then someone with a published rating of 1390 with a web rating of 1405 – then it should be UNR and 1390 or 1200/03 and 1405.

With the answer to the question of ‘b’ and ‘c’ can be the same or different. It is just how the director wants to look at it. Even unrated players will get a assignment or better idea a base line rating. Not sure what the base line rating for unrated players is for Swiss Sys 6.0. If I’m right the base line rating for unrated players is 1000. Having the base line rating can be different with each event, as a group of scholastic player can go as low as 700 to adults as 1300. With adults, the rating department has give the adults the base line rating of 1300. If you play a adult without any past games, the rating change or gain for you would be on a player at 1300.

uschess.org/msa/XtblMain.php?200507183171

The above is a crosstable with a player that never had any rated games before. She lost both games, since she lost to myself with a much higher rating, her base line rating (regular) did not change. On the other hand, my rating did change as I was paired with someone with a base line rating of 1300.

The director can and should use the base line rating of 1300 for the adults. As the rating department is going to use the first base line rating of 1300. This could be a problem with the roster, if the unrated players have a high rating to start out. Can understand to have the base line rating be 1000 for the unrated players. If the tournament is going to have a large number of players between the ratings of 1000 - 1299, would have the base line rating for the unrated players be much lower. This could be a problem with the players under 1000, as a few adult tournaments will have players under 1000. In that case, the players with ratings of 999 or lower have proven they are at that rating level. There is no rational reason to push the base line rating of adult players under 1000. If the tournament does not have published players under 1300, it does not matter what the director use for the base line rating. As the unrated players are going to be at the bottom of the roster.

What I do have as a problem, if the tournament has more then one unrated player with a web rating. If the tournament only has one and only one unrated player. Would not mind to use the web rating, as the web rating of 1200 is between the given rating range of base line of adults (1000 - 1300). Even then, if there is only one and only one unrated player with a web rating, would and should use that rating as the base line rating. What I have as a problem, if the tournament has more then one unrated player with a unrated prize. If I give the player a assignment of a rating, not as a assignment of a rating as a base line rating but as a class prize rating. Then I can be taking the player out of the change to win the unrated prize money. If I give the unrated players different base line ratings, but to keep them in a group of unrated players. The base line rating can change the roster to player stronger or weaker players.

2

This is the other problem I have with the web rating, as the base line rating of the players. The director can give each new player a assignment of a rating, with different base line ratings of what the unofficial ratings state; or the assignment of a rating for the first event, of whatever the director feels is just. In each case, the players still have the right to win the UNR prize.

This will have different base line ratings for the group of UNR players. If the director gives the base line rating of say 1300, for all the first time players. Then use the web rating or the unofficial ratings of the players, this can be any rating. If the tournament has 3 UNR with the unoffical rating of 750, 1425, 1850. Then have 3 players all given the assignment of 1300, as they have never been to any USCF rated event. This is a problem, as the rating range is now 900 points between the top player and the bottom player. The base line rating of 1850 is going to be in the top boards, making the player have a harder time to win any games. As the base line rating of 750, will be in the bottom of the boards making for weaker players. The players in the 1300 base line rating, will be in the same group so they should be in the same board range during the start of the event. In fact, the 1300 base line players should be in the same line up of board numbers during the first round (like board 20, 21, 22).

The 750, 1300, 1300, 1300, 1425, 1850, they are all looking for the UNR prize. Since they have different ratings, except the group of 1300 players, the 1850 will always have a higher board number then the 750 player: if they are always in the same score group. When as the group of 1300 players will change board numbers if they stay in the same score group. The reason to get a higher board number, would be for the support or need of color history. In theory, the higher the board number in the final rounds or last round, would be the stronger players. This would make the 1850 base line player have a harder time to win the UNR event, as in theory would have to play harder players some time during the event.

If on the other hand, if they were all in the same rating base line. It would be the need to take care of the color history, that would place one player higher on the board number then the rest. Even if they stay in the same score group, they should not be off a few board numbers during the start or the end of the event. If they are given large difference in assignment of the base line rating, the board history will show the players having huge difference in the board rating history. Making some players have a harder time to win the same prize, when others have a much better chance to pick up a point or two with the lower boards.

3

The process by which an initial estimate is chosen for an unrated player is well-documented in the technical description of the ratings system, there’s a link for it on the ratings page of the website but here’s the URL:

math.bu.edu/people/mg/ratings/rating.system.pdf

Two change were made to the formula during the development of the new ratings programming, and the technical description was clarified in several places so that the programming is consistent with the formula.

The two changes were to use an initial estimated rating of 1000 for someone in an adult membership category if we have no other basis (such as age) for assigning an estimate and to acknowledge that the USCF office can use a foreign rating other than FIDE of CFC (Canadian) as an initial estimated rating for someone (based on 0 games.)

The 1000 estimate was Bill Goichberg’s idea. The committee objected somewhat to the arbitrary choice of 1000 but ultimately felt that Bill’s suggestion was as least as accurate an estimate than the other (equally arbitrary) choice of 750 when we have no information upon which to base an initial estimated rating.

I don’t believe that Bill’s recommended change significantly altered the post-event rating of any player for whom it has been used.

4

That was not the question I was asking – thanks anyway. Was asking the base line assignment the director will give the UNR players. Myself, like the SwissSys likes to give the base line rating of 1000. Do not have the problem if the director gives the assignment of 1300, as it is common if age is understood of the rating program to start the player at 1300.

Tim Just questions of a) 1200 b) UNR c) 1300 or d) all of the above. Just to state the reason to pick the rating of 1300, if the player will get a base line rating of 1300 as a adult: then the player should get a base line rating of 1300 to take care of the roster. If you have a open event with one section, the UNR players can have a higher roster line up then the scholastic players, or the adults with very low ratings.

The reason like myself and others like the 1000 base line rating, as the idea of age will not make a issue. Having a adult for the first time event having the base line rating of 1300 is fine; it would be a error in judgement to have a scholastic player start out at 1300. Could have two different base line ratings in the same event, one for the adults and one for the scholastic players. The problem with that, they are both after the same UNR prize.

Sure the 1000 rating can be problem as the base line, if the person is a adult it could be deflationary; if for a scholastic player, it could be inflationary. Even with the 1300 base line for the UNR players, it would make the players in the 1000 to 1299 rating group feel a kick in the teath to be lower in the roster then the UNR’s. There will be adults under 1000, then again you cannot have adult UNR’s have such a low base line rating. If I am right, if you have two adults and they never did play in a rated game before, if they play each other and get a draw then withdraw – their rating should be 1300/01. If that is the case, if the director give the base line rating of 1000, the rating assignment of the base line rating is off by 300 points.

If the director sets the base line at 1000, the players will be in the much lower board numbers. If the base line is set at 1300, the players will have higher board number with much stronger players. If the base line is set at a unrational rating of say 2200, the players would be in the top of the boards. The higher the base line rating of the UNR’s, the greater risk of major board upsets given rise to stranger board and color history. The more boards in the event with higher UNR’s base line ratings, the stranger the out-come of the early rounds.

My problem with the at will assignement of the UNR’s as any rating that looks fine. This would place the players in much different board historys. If we can say the assignment of the rating of 2600 for a 6 year old scholastic player in a roster of 100 people is not smart – as the scholastic player should not be able to win one game – but have a large over infationary post rating. If the base line was 1000 not 2600, the player will have a better chance to win one game – but have a much weaker post rating then above.

This is the reason to have all the UNR’s have the same base line rating. The goal of the director is to give the players a even chance to win the prize, not give the players a change to get over inflation as a rating without a win.

5

Douglas, your statements tend to be so long and poorly written that it is often difficult to determine what it is you’re talking about.

As far as I’m concerned, absent of VERY GOOD EVIDENCE to the contrary, players without a published rating should be UNRATED. Period!

Mike Nolan

6

Mike Nolan:

We are in agreement Mike. The question for you Mike, if you have unrated players: where do you place them in the roster. Most directors would place the UNR players at 1000. Or do you place the UNR’s under the worst published rating? In that case you could have a scholastic player with a rating of 250. You could place the UNR’s, even the adult UNR’s under the base line rating of 250. If the event has 100 players, you got to find the base line rating for the UNR’s.

If the worst rating is 1400, there is little problem to have the UNR’s placed at the bottom of the roster. It has little importance if you want to think of the players as 1300 or 1000 or whatever. As it is clear the UNR’s are under the rating of 1400. It would be strange to have 10 adults as UNR’s, with the worst rating of 100. Having 30 scholastic players in the range of 100 to 600. With the 70 adult players with the rating range of 950 to 1850. Sure you can place the adults as UNR’s below the scholastic player of 100. It would be strange pairings for the UNR’s, oh well.

7

Aren’t Unrated players placed at the bottom of the score group? Not sure what you’re asking.

Terry

8

Terry:

Sure you can place them at the bottom of the score group, but some of the computer programs will not place the UNR’s that low. If the bottom of the score group is 100, is that a rational place to place them?

If you look back at the tread about the assignment of the ratings. As Mike did not want the answer to the question C. I know it is hard to follow as it is in two threads not one, it was not my choice but Mikes wish.

It was the idea to give the player the rating of 1300. That is a base line assignment of a UNR rating. There can be players with lower ratings in the score group, like scholastic players. As you did say the answer is D, as the director can use a) 1200 b) UNR c) 1300 d)all of the above. This is to answer the question c) 1300. The director can give all the UNR’s a assignment of the rating 1300.

Sure, in very small events with all ratings over 1000, the UNR’s are placed at the bottom of the score group. This is very much the standard, as it is not common in adult tournaments to have so many players under 1000. If on the other hand it was a mix, with adults and scholastic players. If the bottom of the score group was the rating 100: with 20 players from 100 - 999. If you have 20 adults as UNR’s, if you place the players in the bottom of the score group you would in fact have their base line rating under 100, as the lowist score group in the event is 100.

If the director does place all the UNR’s below the lowist score group. The rating of the players in that group could be scholastic players. Some could be as low as the rating of 100. That would give all the UNR’s a base line rating below 100. That would be sandbagging all ratings of all the UNR’s, in fact they could have a final score higher then the class prize.

The problem with the assignment of the rating 1300 for the base line rating, it could be a inflationary rating for all or most of the players. It would be a nice idea, if there is going to be a large pool of UNR’s in the event. If the event is going to be a large prize, having a higher then the norm rating will weed out the much weaker players.

Its’ true Terry, the director can give any assignment of rating. It becomes a question for the assignement or better the base line rating. If the bottom of the score group is very low, it would only be sandbagging the ratings of the UNR’s. If there are a number of players with very low ratings in the score group, it will give all the UNR’s a much lower final rating after the event.

9

Douglas, I am not certain what you are getting at.

I think that you are trying to determine a pairing-purposes-only-rating to use for the unrateds that does not pair them too low in a score group. I’m guessing that you are concerned that if unrateds are paired too low then their real ratings will be based on lower rated opponents than they would have had otherwise. I couldn’t quite follow your syntax, but I think you do realize that a pairing-purposes-only-rating is not used when the USCF does the actual rating calculation.

In this example, assigning a player a rating of 1300 will have an effect on the pairings, but the potential change in the opponents the player gets paired against is the only affect that the 1300 assignment has on the actual rating calculation. The calculation is still based solely on the actual results and does not get informed of the pairing-rating that was assigned.

Also, Tim did state that assigning a 1300 rating is something he would be unlikely to do (and I agree with him there), but it is a legal option for a TD that feels there is enough reason to do so.

10

For what it’s worth, the old Harkness Blue Book recommended revising the ratings of new players during the course of a tournament, as more evidence of their playing strength became available. Of course, this was in the days before bloated class prizes, when the only function of the rating system was to generate equitable pairings.

11

Terry,

Please see rule 28A, p. 108, second paragraph and the TD TIP that follows. It explains why you might want to place the unrated players higher than the bottom of the scoregroup.

Regards,
Tom

12

In an adult tournament, I prefer to put unrateds at 1399. For one thing, it makes it easier to avoid accidentally giving a full-point bye to an unrated, or pairing an unrated (rather than the lowest-rated) down into the next score group when there is an odd player.

I don’t write 1399 on the wall chart, though. I just write UNR.

Bill Smythe

13

Let me try it with only one topic then a number of the same topics. When building a score group, it can be cut into a number of sections. Number of directors will base the section on 8 to 10 per-section. If one section has the rating scale of 1190 - 1470, the bottom section the rating scale of 520 - 840. If the UNR player is a adult, it would be better for the player to be in the section of players from 1190 - 1470. The director could place the UNR in the section of players from 520 - 840.

If the UNR is in the section of players in the rating scale of 1190 - 1470, the UNR still has a chance to win the section. As any UNR’s are unclear how they will do in the first event. The post rating of the UNR can have a large swing in theory between 800’s - 1600’s, in practice the rating should be between 1100’s - 1400’s.

If the UNR is in the section of players in the rating scale of 520 - 840, the adult UNR has a more then average chance to win the section. There are a number of adults under the rating of 1000. The post rating of the UNR in this section will have a large swing in theory between 100 - 1200: in practice the rating should be between 700’s - 1000’s.

Since it is still in the early performance rating of the UNR. The member would have a much lower rating, being in a section of low rated players. It is the goal to win the section, it is the long term goal to get the rating at the right and correct rating level.

If the UNR is in a low rating section, the post-rating can be a very under-rated player. The drop in practical post-rating of 1100’s - 1400’s into 700’s - 1000’s, would have a much lower rating going into the next event. The next event can increase the players post-rating with a great deal of inflation in rating. The people that will be paired up with this person, will have a small rating deflation.

The difference between the two sections, the bottom of the top section is 1190, the top of the bottom section is 840. This is a base line rating range of 841 - 1189. If the director wants to give all the UNR’s a base line rating of 1000, this will be within the base line rating range of 841 - 1189. True, if the UNR is in the top section the bottom of the rating scale is 1190. If the UNR is in the bottom section, the rating at the bottom is 520. This will give a base line rating range of 100 - 519.

The question becomes, do you feel the rating range of a adult UNR is between 841 - 1189 as a base line rating, or 100 - 519 as a base line rating?

14

Douglas, the multiple section example you gave (10:41 on August 4) is one reason TDs have the discretion to assign ratings to be used for a tournament. If you feel strongly that one unrated should be assigned 1300 and another unrated should be assigned 500 then you are allowed to do so. There are some things to realize.

  1. Whenever you assign a tournament-purposes rating to an unrated you run the risk of having that assignment questioned, sometimes quite vocally. Be ready to explain your reasoning. For rated quads at my club I’ve assigned a tournament-purposes rating to unrateds to get them in a decent section, but I’ve had the advantage of having already seen results of the players against other club members in our serious but non-rated G/90 ladder games. If you don’t have the advantage of having prior information about the player’s skill level then it is harder to successfully explain your justification for assigning a tournament-purposes rating.

  2. If a player is assigned to a section with a noticeably lower strength than the player’s strength, then it is quite possible the player will end up with a perfect score. The first provisional rating of an unrated with a perfect score is not 400 points above the average rating of the opponents, but rather 400 points above the highest rated opponent (in a multi-week tournament earlier this year an unrated missed the first couple of rounds and had to skip a later round - he ended up with a perfect score for the games played, including a somewhat lucky win over a 1600+ and ended up with a provisional expert rating even though I’d guesstimate his true strength to be around 1400).

The post-event rating might be higher when placed in the lower section than when placed in the higher section. In your example, a perfect score unrated (1300 strength) in the 520-840 section might end up with a rating of 1240 while that same player in the 1190-1470 section may go 1.5-2.5 and end up with a rating of 1230 (assuming the average opponent was 1330) or even under 1200 if the average rating of the opponents was less than the average of the section.

  1. If the unrated is a parent of a scholastic player it is not unusual to find that the parent is not as strong as many of the unrated scholastic players.
15

I’m not understanding. The question Douglas posed was “where to place unrated players in the roster”. To simplify things, assume a 40 player tournament. Prior to the 1st round, players are given pairing numbers based on their ratings. Unrated players are placed below the lowest rated player, and are assigned arbitrary pairing numbers. This is what I was referring to. Maybe I’m not seeing the whole picture?

Terry