Blitz & Pairing Questions

Running Chess Tournaments
1

At USCF Blitz events: If a player touches a piece that cannot be legally moved, do they lose? I ruled no, but others said yes.

When using team sensitivity setting, is it possible to turn it off for just a portion of the section? In this case, one team monopolized the 2-2 score group, so SwissSys caught that. After taking off team sensitivity, the top boards (3-1 scores) had several teammates playing each other. Would the TD have had the right to manually re-pair those top board pairings? It was not the final round.

2

The TD has the right to do that but are you suggesting those players play out of their score group? That might prove a hard pill to take for the players in the lower score group.

3

Thanks for the quick answer Tim. No, the other players would be in the same score groups.

4

I can’t believe that the USCF would have a rule that compels a player to make an illegal move. Perhaps I am mistaken, but I would never rule that way. There should be no penalty for touching a piece that has no legal moves, unless it is done deliberately, of course, to annoy the opponent.

As far as pairings are concerned, the TD is ultimately responsible for all pairings, no matter what the computer says. The TD not only has the right but the obligation to fix pairings that are incorrect, for whatever reason. I can go either way on whether teammates being paired is an incorrect pairing, but there is no reason, within the scoregroup, not to change the pairings in that situation, IMO.

Alex Relyea

5

Alex and Tim - thanks for your knowledgeable comments. I need to check this forum more often because I have a lot to learn!

6

WinTD has an “only if necessary” setting for the teammate pairing, which will pair teammates within a score group if there is no other way to pair the score group while avoiding pairing teammates; or you can set a threshold point value above which teammates can be paired, again, if there is no other way to pair those score groups. However, one team monopolizing the 2-2 scoregroup doesn’t seem to be that big a deal. Dropping some 2-2’s to play 1.5-2.5’s or 1-3’s isn’t likely to affect any prize winners.

Tom Doan

7

It might effect those team scores for the 1.5-2.5’s or 1-3’s.

8

How is it possible to “monopolize” the 2-2 score group? That group is right in the middle, and normally would be the largest group.

I have my usual plea: Let’s see the crosstable.

Bill Smythe

9

If there are team awards, then there would really be no good reason for pairing 2-2 teammates. When I do a scholastic tournament with no team awards, I’ll generally avoid pairing teammates only if

(a) the teammates are monopolizing the perfect scores, so any pairings would be “pick a victim”, or (possibly)
(b) the teammates are monopolizing the bottom score group(s).

For (b), I would generally ask permission of the coach. It seems to be about 50-50 between coaches who would rather have their kids play someone outside their school, even it they’re overmatched, or those who would rather seem some of the kids win something, even if it’s just against each other.

10

SwissSys has the option to use the plus-two rule for pairing teammates in individual-team tournaments (28N1, pair teammates if necessary in plus 2 or higher scoregroups). But in practice, I usually have more team problems in the bottom scoregroups than the top ones; it’s not uncommon to have scholastic teams of all new players.

In the middle and bottom scoregroups, SwissSys will do whatever it takes to avoid pairing teammates. I give the program a shot at the pairings first, and then look at the consequences. If you have to drop down a couple of 1-3’s to play some 0-4’s, that’s often preferable to pairing teammates. If it drops down a 1.5-2.5, that’s pretty sketchy unless maybe the 1.5 had a full-point bye. If you get 2-2’s vs 0-4’s, which can easily happen in a small section, then it’s pretty clear you need to do something.

Rather than manually re-pairing, I temporarily delete the team codes for the teammates in the problem scoregroup and let SwissSys re-pair the section. If you were using WinTD, you could turn on the “only if necessary” option (which works for all scoregroups) without having to play around with the team codes. I don’t have a lot of experience with WinTD; I suppose you could leave that option on all the time and not worry about it. But then you lose the possibility of dropping down a couple of 0.5-3.5’s to avoid pairing two 0-4 teammates. The rule book implies that you can be more flexible with the teammate pairings in the lower scoregroups, which I completely agree with, but it’s hard to put that into an algorithm.

Fortunately these problems don’t usually happen out of the blue. If you have a small section with either a relatively large team, or a team where most of the players have low ratings or no ratings, that’s a red flag. You can usually let the coaches and/or players and/or parents know about problems a round or two ahead of time, and avoid riots. As Tom Doan said, often 0-4 players (and their coaches) don’t really mind playing teammates, because their team is guaranteed a point.

Teams with a lot of high-rated players who end up in the top scoregroups aren’t really a problem, because they usually know what to expect, and the rules make it clear what you need to do.

11

Even though they know what to expect, and why, complaints still come when you warn the coaches of what you will be doing. There was one tournament this year where a coach appealed to a spectating NTD when I paired two teammates in the final round (the team had three of the four perfect scores going into the round). The coach said that they don’t do things like that at a national tournament. Unfortunately for his argument, I cited the 2005 NYA as a tournament where such a pairing was deliberately done (I did discuss it with the NYA chief TD before doing it then). By the way, the spectating NTD said he was retired and would not be able to support the coach when the coach came to see me.

In another tournament this year a different coach who has faced this problem before still requested looking for a legitimate alternate pairing (the team had four of the five perfect scores, so two teammates had to be paired).

One common complaint is that the team will only be able to garner one point from the game and other teams thus have a better chance to catch up. My counterargument is that the team is guaranteed at least one point from the game and other teams have that much higher a bar to reach to catch up. In the second tournament mentioned above, that pairing ended up clinching first place for the team going into the final round (the top four scores were used for the team score, and as it turned out they ended up five games ahead of the second place team).

My non-scientific feeling is that the two arguments roughly counterbalance each other.

12

For the upper score groups, there is at least a definite rule you can point to (although there are variations). 28N1c says “For score groups of plus two or greater … players should not be removed from their score group to avoid playing those from the same team.” We always follow that rule, and set SwissSys to follow it, and that usually quiets most of the arguments.

At our recent state championships, we had an argument from an opposing coach that teammates should have been paired, but weren’t. In the fifth (next to last) round, two of the three 4.0’s were teammates. SwissSys dropped down one of the teammates (which was the natural pairing anyway), but the only 3.5 was also a teammate. SwissSys skipped the 3.5 and paired the 4.0 with the top 3.0. It comes down to how much weight you give to individual competition vs. team competition, and it’s a subjective thing. Note that even if you consider 3.5 as a scoregroup for the purposes of the rule, it’s not a plus two scoregroup. (As it turned out, the 3.0 was the highest-rated player in the section, much higher than the 3.5, and the 4.0 lost the game.)

To me, the Swiss system was designed for individual competition, and the team idea is sort of grafted on. There will always be conflicts. Unfortunately, to many scholastic players and parents, team competition is everything, and individual competition is a mildly interesting sidelight.

13

That last statement isn’t true, is it? After 4 rounds, 3.5 means 3.5-0.5, making it in fact a plus-three scoregroup.

So SwisSys apparently removed the 4.0 from his (new) scoregroup (3.5) to avoid pairing teammates, in violation of the rule you cite. I’d say the complaining coach had a valid point.

I suppose, though, it could be argued that since the 4.0 had already been removed from his original scoregroup (4.0) for another reason (odd number in group), that he was no longer in his scoregroup anyway, so how could he be removed from it?

As usual, I’d like to see the whole crosstable, complete with color histories and team affiliations.

Bill Smythe

14

Yes, of course Bill is right, 3.5 is a plus three scoregroup. The gray area is what happens once you leave the original scoregroup. I’ll try to dig up the crosstable.

15

This is the crosstable information for the top scoregroups. Colors, as you can see, are not an issue.

Ann, 4.0, 2203, Team A, BWBB
Barbara, 4.0, 1945, Team B, BWBW
Cathy, 4.0, 1757, Team A, BWWB

Deirdre, 3.5, 1939, Team A, WBWB

Ellen, 3.0, 2242, Team C, WBWB (top-rated 3.0)

None of these players have played each other.

SwissSys pairings:
Ann vs. Barbara
Ellen vs. Cathy

IMHO, SwissSys’s (is that enough s’s?) rationale was, once Cathy left the top scoregroup, to find the next scoregroup where she could be paired. In retrospect, I think this follows the letter of the law, but possibly not the intent. Does a scoregroup, for the purposes of the rule, refer strictly to the original groups, before dropping odd players?

16

I agree. If a player shouldn’t be paired out of his group merely to avoid team pairings, then it makes even less sense to pair him out of his group again merely to avoid team pairings, even if the original out-pairing was done for another reason.

A better wording might be: “28N1c. … A player with a score of plus two or greater … should not be treated as an odd player [29D] merely to avoid playing another player from the same team.”

You still didn’t furnish the entire crosstable, so of course there remains the possibility that other factors (such as serious pairing problems within the 3.0 scoregroup) could have prompted the SwisSys pairings.

Bill Smythe

17

I don’t know if this would be helpful or not, but SwissSys (at least in version 7) has a pairing logic module under Tools in the pairing chart view. This should show why it has paired as it has.

18

Unfortunately, the pairing logic does not show any changes made to avoid rematches or avoid teammates; it seems to be confined to improvements made for color. That is, the “natural” pairings that are the starting point for the logic diagram have already been corrected for rematches and teammates. It would be very helpful in this case if it worked as you suggested. As it is, it shows Ellen vs. Cathy as the “natural” pairing.

19

With WinTD, if you pair this allowing teammates to be paired at or above 3.0 (which is how +2 and the like are implemented), it would pair Ann with Barbara and Cathy with Deirdre (barring some problems preventing the 3.0’s from being paired normally). The “logic” that’s built in to the algorithm for scoring pairings is that no unnecessary drops among the 3.0 and above score groups are permitted if they can be avoided by pairing teammates.

Tom Doan

20

Question: What would happen if the whole scenario occurred in a score group one point lower? Such as:

Ann, 3.0-1.0, 2203, Team A, BWBB
Barbara, 3.0-1.0, 1945, Team B, BWBW
Cathy, 3.0-1.0, 1757, Team A, BWWB

Deirdre, 2.5-1.5, 1939, Team A, WBWB

Ellen, 2.0-2.0, 2242, Team C, WBWB (top-rated 2.0-2.0)

After pairing Ann vs Barbara, would we now:

(a) pair Cathy vs Deirdre, on the grounds that Cathy is plus-two, or

(b) avoid pairing Cathy vs Deirdre, on the grounds that Deirdre is not plus-two, or

(c) avoid pairing Cathy vs Deirdre, on the grounds that Cathy’s new scoregroup is not plus-two?

Bill Smythe