I have pasted Appendix F of the Vega user’s manual below and would like comments/feedback concerning rule # 29D1 and the unrated player. Specifically, does the way Vega implements the rule cause a major problem and who is affected the most, the unrated player or the the player he/she is paired with? Item 2 in the appendix concerns the same color three times in a row and is a clarification that Vega, by default, avoids it but will allow the TD to override.
I was in communication with Luigi Forlano, the author of Vega, and received his permission to post Appendix F for comment.
Appendix F: Notes on the USCF swiss implemented in Vega
Vega implements all the USCF rules (5th edition 2003) with two exceptions listed below that should be
notified to the players before the tornament starts. This last point is very important, that is the player
should be aware of these variations because could prevent him to partecipate the tournament.
The unrated players in Vega are the players with rating equal to 0. In each score group the players are
sorted by rating and alphabetically when the rating is equal. In this way the unrated stays at the bottom of
the list. As far as the BYE assignment is concerned Vega follows the USCF rule 5th edition.
Instead Vega treats differently the unrated player with respect the odd player determination. In this case is
used the following new variation:
29D1. Odd player determination
a.(NEW) In the case of an odd number of players, the lowest-rated player is
ordinarily treated as the odd player and is paired with the highest-rated player he or
she can play in the next lower score group. Care must be taken in doing this that
the odd player can be paired in the next score group, that the remaining members
of both affected score groups can be paired with each other, that the odd player
has not yet played all the members of the next lower group, and the color
consequences are acceptable.
In this way, because the unrated player is even the lowest rated player in a score group, he will
have a higher priority to be the odd player in contrast with original 29D1a.
The colors in a series by default are treated by the variation 29E5f1 that states:
29E5f1: Last-round exception. Except for the last round, when it may be necessary to pair
the tournament or class leaders, players shall not be assigned the same color in three
successive rounds.
This case is the default case and it is handled automatically by Vega. This variation is currently
included in all FIDE swiss systems. If the tournament director needs for some reason to permit
three colors in a row before than last round then he should set ‘on’ the flag “Accept WWW/BBB”
in the pairing page. To remove the variation 29E5f1 the flag “Accept WWW/BBB” should be ‘on’
during all the tournament.\
The problem with the software with the unrated players, because the unrated players would not have a rating. If the player did have a rating, then you would not have this problem. There are a number of schools of thoughts for the assignment of a rating of a unrated player. Someone from the rating department should answer this question. If there is a tournament with all unrated players, what baseline rating does the rating department use. If Norm can answer this question, would use that rating for you’re program. As the baseline the rating department use for rule 29E1 and rule 29D1. Back in the 1980’s, the rule for the baseline of unrated players meeting each other was 1000. What the baseline rating for a tournament of all unknowed players meeting each other for the first time, my data is out dated. If you could place a rating into the program, then you would not have this problem.
The problem isn’t with assigning a 0 value for the unrated player. The problem comes into play when the unrated player is in an odd score group such that a player needs to be dropped to a lower score group. The USCF rule basically retains the unrated player in the score group and drops the lowest rated player to the lower score group. Vega would drop the player having the lowest rating to the next lower group which would be the player with the 0 score, the unrated player.
If the persons ratings is 1000 then there would not be a problem. If you give a assignment to the player a rating of a 1000, how can the computer software not over ride its own problem.
Let’s say I give the unrated player a rating of 1000 for pairing purposes. If that player ends up in a score group having an odd number of players, then the rules state that the lowest rated player should be dropped to the next lower group not an unrated player. So if his artificial rating of 1000 still puts him as the lowest player, he will be dropped to the next lower group, not one of the rated players above him. So given that the software drops the lowest rated, regardless if they are unrated or not, the question then is, if notified beforehand that this will occur, is this a signifiicant change to the pairing rules that would adversely affect the players. Or would the effect to mininmal?
If you break you’re tournament into sections, if you’re tournament does have a odd number of players. Even if you do give the unrated player a assignment for a rating, then you still cannot even give this player a “bye” during the tournament – that would affect the players more then having players under 1000.
Having players in a open-age tournament, will not have that many players with a rating less then 1000. Even if you do give a assignment for a unrated player at 1000, you would have very few players under this rating or will not have any.
If you have a scholastic tournament with a unrated player, then the rating of 1000 would be a little high. When the 2004 Annual Rating List comes out, use the average for the scholastic members. The 2003 Annual Rating List has it at 626 classical and 660 quick, for the scholastic members. Or you could have it for the non-scholastic members, it is 1267 classical and 1262 quick. Would rather use the baseline of what the rating department use when they have a tournament with all unrated players.
The question you have, if you have two sections in a tournament and the unrated player is in section 1 and there is a section 2. Say you have a assignment of 1000 for a unrated player. More or less that rating would place the player in the bottom section, or would be at the bottom of the ranking of the players.
This is the problem I think your talking about. Say you have two sections with 19 players. You breaking the tournament into a section of 10 and a section of 9. You have the unrated player assignment at 1000 and would be ranked number 10. The next section of 9 people, would have a rating less then 1000. Try this, find the rating of player 11. Give a re-assignment to the unknowed player the same as player 11, place that player in the top section and the unknowed player in the bottom section. The unknowed player would be the top ranking player with player less then 1000. As you would be needing to give a bye, the unknowed player cannot get a bye in the tournament. The assignment of the rating of a 1000 or for this problem for a adult person. Could this happen in a tournament yes, could I win the lottery of millions of dollars yes. This problem would be so rare, so uncommon, would not worry about it ever being a issue.
The problem doesn’t arise at the beginning of the tournament, it’s after a few rounds have been played and now we have 5 players with a score of 2.5 and 3 players with a score of 2.0. The unrated player has 2.5 points and since there is an odd number of players with that score, one is moved to the next lower score group, those with 2.0 points. The lowest “rated” player in the 2.5 group is the unrated player but the rules say he should not be moved to the next lower group, rather the lowest rated player in the 2.5 group should be moved. The software is just moving the lowest player and because of that I would like to know if it is truly detrimental to one of the players and if pre-posting that this method of resolving odd score groups is sufficient.
It’s possible that this would have a harmful effect, but not terribly likely. The reason for dropping the lowest rated player is that the Swiss System is trying to converge the scores. You are dropping what is supposed to be the weakest player in the higher score group to the strongest player in the lower one, and it is assumed that he will probably lose. Unrated players are of unknown strength, and are ranked arbitrarily. In my opinion, this pairing variation is slightly inferior, but there’s certainly no reason you can’t use it if you announce it in advance.
Now understand you, thought you were building a score group of people in a section. It would not matter too much if you drop the unrated player into the 2.0 group or someone else into that group. When you drop one person into the 2.0 group you will have 7 people not 6 for pairings. You could drop the unknowed player into that group, or pick the next player in that group of 2.5 points.
Try this, check and see how the colors work with the round with the unrated player and the one that is rated. If the colors work out better for the whole group with one player going down then the other: then take that for the pairings. This is a given, players will forgive the director for making unorthodox pairings, they have long memories if you mess with there color history.
When you were talking about assignment of a rating for a unknowed player. Thought you were talking about building a section in a mini-swiss.
Thanks. It sounds like it is not that big of a deal as to which player actually gets dropped to the lower group. The program handles due color quite nicely so that should not be an issue. You can also turn on “verbose” mode which will print out the rules enacted for each pairing.
You don’t happen to know the reasoning behind not dropping the unrated player do you?
The idea is to drop the WEAKEST player, so that he’ll lose to somebody stronger in the next lower group.
An unrated player is usually not the weakest player in the group. In fact, just being in that group tends to prove that he is about as strong as the AVERAGE player in that group.
If you assign an unrated player a rating of 1000 for pairing purposes, and he still ends up as the “lowest” player in the group after a couple of rounds, that is a good indication that 1000 was too low an estimate to begin with.
It is hard for a director to give a assignment of a rating for these unknowed players. Would need to take in as a factor the age of the person. For a scholastic player, it would not be so hard to pick a rating around 500 to 700 for the K-5 age group; for the K-12 that be between 700 to 1000.
For a adult, it is hard to say as they would have some skills internet servers like yahoo, ect that assign ratings. The problem with that rating, as internet ratings are designed for a player for a computer screen then the over-the-board. Would have to take there word what there internet server rating is.
The assignment of a rating for a unknowed player is not that big of a problem, if the tournament is a open. As the unknowed player or in one case a provisional player with a pre-tournament rating in the 900’s did win the under-2000 prize. When talking to unknowed players, have made it a issue for them to play only in tournaments more designed as a open then a mini-swiss or a tournament designed with sections.
If the director does have a assignement for a unknowed player, and the tournament is designed with sections of players with common ratings. If the true rating of a unknowed player would be as a class B player. Being in a section with players with ratings in the class D/E players. Then the unknowed player would have a better shot to win the section and come out with a lower provisional rating. If assigned in a section with class B/C players, would have a weaker chance to win the section: would come out of the tournament with a higher provisional rating.
It becomes a parodox for the director when the assignment of a rating for a unknowed player. If it comes with a tournament with sections, the director could place a strong player into a weak section. Or, the director could place a weak player into a strong section. It is up to the director, how they would deal with this issue.
The only way a director can take care of this problem of assignment of rating for a unknowed player. Would have a tournament designed as a open. In a open the unknowed player in the last round would have the right to be on board 1 on the last round, or be at the last board during the last round. If the director plans to have a large percentage of unknowed players and players with established ratings, with a tournament with the goal of 15 to 30 players. Having a open would be more fair for the unknowed players. As the more rounds in a open, the assignment of the rating would be less of a issue as the score of the player would take over.
It is a issue for the director. For the players they do not mind who they play with during that round. Players only make a issue when they feel they did not get due color, or the color order for them is out of balance.
If the tournament is a odd number of rounds, the colors could work out nice for the player and who they meet in the next round. The drop of the unknowed player or the next player in the score group. It could be a issue for the other players in that score group or even the rest of the lower boards after pairings are done.
Having a tournament designed to have odd number of rounds, you are going to give someone more blacks then whites for half of the players. Players do look at how the pairings are done more when having odd number of rounds. As a player I would, if having more blacks then white I would a little lot more. If you have a expert going 0-3 with one white and two blacks. The trash talk on the pairings will be the experts reason for the poor performance. If the expert went 3-0, having two blacks and one white or two whites and one black, the pairings will not even be a issue.
I still don’t see why you’d want to. Why not just drop the weakest player, which is probably the lowest player with a rating?
Granted, in small score groups, that may not always be possible or feasible. If the unrated has already played most of the players in the group, or if colors are just horrible, etc, you can make an exception and drop the unrated – don’t bother to post this “variation” in advance.
Or, if the unrated is obviously the weakest, go ahead and drop him. This would seldom happen, however. Maybe if he had a half-point bye and then lost all his other games in under 10 minutes, you could conclude he was the weakest.
I guess I’m a little bothered by why you would WANT to drop an unrated as standard procedure.
The software is coded that way and is not under my control. Do you know if Swiss Sys or one of the other Windows based programs handles this correctly? Not that it would help me because I am running a Linux only PC.
Even if you give a assignment of a rating the software still has this problem? The reason to ask, if you know a persons published rating would it still have that problem: drop the player down. Why not run a little test tournament and see if it would work. As how can the software know if the assignment rating and the published rating are not the same.
SwissSys normally drops the lowest-rated player (with many caveats, as has been pointed out). Obviously, if you want it to drop an unrated player, you can assign a rating to by typing it into the field. If you want to tinker with a pairing, however, it is usually easier to generate the pairings and use “Switch player.” Modifying the underlying data has the drawback that you may forget to change it back later.