Best Way to Run a Quad Tournament

I will be running a Quad tournament once a month starting in March. What is the best way to run a quad tournament (how to do the pairings, colors)?

The USCF rulebook says “the preferred pairing table for quads is as follows” (Player’s numbers are assigned in order of rating, not randomly as in larger round robins.) Round one: 1-4, 2-3, round two: 3-1, 4-2, round 3: 1-2, 3-4 (colors by toss in this round)". What is the rationale for doing it this way? Also, is there a way to do the pairings like this in SwissSys without doing the pairings manually?

Also, it says “When the total number of entries is not divisible by 4, the director may create a 3-round swiss among the lowest 5 to 7 players.” Why do the lowest 6 to 7 players automatically make up the 3-round swiss? Wouldn’t it make more sense to make the 3-round swiss be whichever group makes all the groups the most even?

The terms “preferred” and “may” give the TD the flexibility to organize the quad+swiss tournaments the way he wishes. He can shuffle the ratings in a quad so that the highest rated player is not #1. This has the advantage of varying the colors for players if the same people show up all of the time. The top 4, etc. in each quad can be ordered according to when they enter the tournament or just placed randomly. I have done this a number of times to make the event more interesting and less predictable.

While it is not usual to have the Swiss section be anything but the last section, I can see where it would be equitable to group the four lowest rated in their own quad and have the Swiss section more rating balanced. It is your tournament. If you announce in advance what you are going to do, there should not be a problem in breaking the mold a little to make the sections more equal.

I’ve never heard of a quad done anything other than “by the book”. I’d expect that if you were going to announce “Quads” and do anything different, you should announce that in all advance publicity as a major variation.

SwissSys will handle round robin pairings correctly, but what is difficult about doing them by hand?

Alex Relyea

Many times clubs and TDs let each quad proceed at “their own pace” if they finishing more quickly
than their time control allotment. And simply posting the pairing chart for each quad lets the players
do this.

Rob Jones

Are you trying to make groups of equal strength, or are you grouping top to bottom? If the latter, I’m not sure what “makes all the groups most even” means. Defining and computing “evenness” can also be quite a challenge (see viewtopic.php?f=2&t=19596). With standard quads, you group top to bottom and put the extras into a Swiss. This has the advantage that the lowest rated players are playing the Swiss, so you are more likely to be able to either get in a full extra round, or at least can cross round pair to deal with an odd number of players. The higher rated players are less likely to get enough games done quickly enough to allow those types of adjustments.

The rationale (note the e at the end!) for the recommended method is that 1-2 don’t meet until the last round. That’s the only real advantage of this over the standard random draws. I would recommend the latter if you’re just assigning people to quads based upon ratings rather than based upon some type of ranking earned in preliminary competition.

A lot of tournament rules (such as due color and byes) give “preference” to higher rated players, and I see this as another example of that. The higher rated players “deserve” more to have a true quad that they signed up for rather than a random small Swiss.

Whether you think that is fair or not is another matter, but I am used enough to the principle of “higher rated players are forced to deal with the fallout of exceptional circumstances less” that it seems reasonable to me.

I think what Mr. Smith means by making the sections most even is that in some situations it might make more sense to have a middle section or higher section be the swiss instead of always being the lowest section.

For example, say we have fourteen players (two quads of four and a swiss of six)

The top 4 players are all rated above 1800.
The next 6 highest players are all rated between 1400 and 1600.
The next highest rated player (11th overall) is rated 1100.

If we follow the lowest six players in a swiss guideline, in the lowest section there would be two players rated over 1400 and then the other four players at or under 1100, but if we make the middle section six players instead, then in all three sections all of the players would be rated more closely together.

Of course, the director can do this now, so no change in the rulebook is necessary.

If that’s what he means, then it’s completely wrong-headed. The Swiss is better able to deal with a range of abilities. The last thing you would want to do is have quads with wide ranges and a Swiss with a narrow one.

Well, I think the purpose would be that there isn’t a big range in ANY section.

In this VERY hypothetical situation you would have:

Quad 1: 1850, 1830, 1810, 1800
Quad 2: 1600, 1575, 1550, 1525
Swiss: 1450, 1400, 1100, 1050, 1000, 950

Or you could have

Quad 1: 1850, 1830, 1810, 1800
Swiss: 1600, 1575, 1550, 1525, 1450, 1400
Quad 2: 1100, 1050, 1000, 950.

In the latter case, no two players in a section are more than 200 points apart

Yeah, right. I’m sure the 1450 and 1400 would be happy as clams that the TD did something completely unexpected and (in practice) thoroughly arbitrary.

I wouldn’t go so far as to say “I’m sure”, but I wouldn’t be surprised to hear that they actually were “happy as clams”. I know you were being a teeny bit sarcastic, but if the event doesn’t have any prize money then some players would enjoy the challenge more than some relatively easy victories.

That said, it may be the the other four class C players that are upset because they didn’t get the quad they expected to get under the rules. Some of the class E players may be upset in not getting the challenge they expected.

As stated earlier, if you are going to vary from the standard make sure it is clearly announced in advance or is clearly and obviously agreed to on site with no intimidation or embarrassment used to get that agreement.

Well some people like to play up and might cherish the competition of players rated 100-200 points higher than them rather than 300-400 points below. $ usually isn’t as large a factor in quad tournaments which are typically low EF, low prizes, players often play for the experience and rating points rather than the money or a chance at 1st place.

But, I’ll echo Mr. Relyea’s point earlier that any organizer who would do this would need to put this variation in the advance publicity, that the director could assign the swiss to a group other than the bottom one.

I’ll stop putting words in Mr. Smith’s mouth now, as he may have meant something entirely different.

The main point is that there are good reasons for grouping from top to bottom and putting the extras in the bottom group. Since the grouping is, in effect, “pairing” the tournament, participants should expect that it will be done according to some reasonably well-defined process (as with pairing a Swiss) rather than something arbitrary that the TD comes up with. If all six players in the proposed middle section were OK with playing a Swiss, then I would say go ahead with that (I don’t think the bottom four would have any standing to object since they would be playing the advertised quad), but I don’t think I would leave that up to only the two that the TD was thinking of floating up.

Suppose the turnout for your quads was as follows:

5 masters

4 C players

4 E players

Does it make sense to throw the lowest rated master in with 3 of the C players and the lowest rated C player in with the 4 E players in a Swiss or 5 player RR?

I’ve never run quads where the “odd-man” Swiss wasn’t in the bottom section. It’s an interesting idea, and I can certainly see where it could have practical application. However, I would think all the affected participants (so, everyone but players in the quad(s) that are above the proposed redistribution) would need to agree without coercion or reservation to changing the order before I would try it.

Moreover, if these quads are being used for some sort of qualifying or grand prix-type structure (a seasonal or annual cumulative competition), I would think uniformity in player distribution would increase in importance.

Ask one of the masters to not play. With the five player Swiss, you have three of the masters playing only two rounds. Other than hoping that everyone would consent to running a 13 player Swiss (and I’m guessing the top four masters wouldn’t), I’m not sure I can think of anything that would work reasonably. Probably the best lesson is not to advertise quads unless you expect a better-distributed set of strengths.

Try to avoid having the same players in the same quad month after month.

Micah, first and foremost, thank you for organizing an event. Every one helps advance the cause.

I wouldn’t try to reinvent the wheel when it comes to assigning sections. Players come to a quad with certain expectations, and groups of four until the bottom section is one of them. If you deviate from this, you should have a compelling reason (only experience I’ve had: siblings in the same quad, ranked 1 and 2, wherein the bottom player in the next quad up is just tens of points above the top ranked sibling, and no player objects to a change when asked).

Similarly, when it comes to pairings, you can’t go wrong with the recommended method. Nevertheless, if you have a pairing program, you may want to consider having the program do the pairings so that the computer does the last round toss and prints the pairings for all three rounds in advance. This is my practice because I view leaving the toss to the players as inviting shenanigans. I’m the director, I get to pair. Your mileage may vary. If you have no program and you have to have the players toss, so be it.

Good luck!

I co-sign Mr. Price’s post above - especially the first paragraph. Organizers are hard to find. Thanks to Mr. Smith for taking on a normally thankless task.

When I do the pairings as a round robin, it assigns the pairing numbers randomly. Is there a way to have the pairing numbers by rating? Also, it seems to always give the top 2 pairing numbers White in the first round and Black is the second round. Is there a way to make this random?

Yes, this is what I meant.