The Official Rules of Chess discus only two tournament types: the round robin, and the Swiss system. Why is it that the round robin is so popular, and why is it used so rarely for other kinds of tourney?
The Swiss is sort of an elimination tourney in which no one ever gets eliminated. For the second round, first-round winners play other first-round winners and losers play losers. In subsequent rounds, players are grouped by won/loss record, with, for example, players with two wins and two losses facing each other in the fifth round.
The Swiss system makes sense whenever the limiting factor is the number of rounds to be played. With a bit enough room and lots of chess sets, there’s no reason why some players should have to drop out of the tourney early, or to wait around for a match to become available. It would seem a natural choice not just for chess, but for other games like backgammon. Why shouldn’t backgammon players enjoy the same maximum participation that chess players do?
[apologies to Jonathan Steinberg, author of Why Switzerland?, which is the best book I know for understanding Switzerland, the country.]
The Swiss System is not unknown for backgammon tourneys, but it’s not the norm. Let me suggest three reasons for this:
- Running a Swiss system tourney is complicated. Lots of non-chess players don’t really understand it, and perhaps are also inclined to think that the complexity hides a lot of opportunity for the tourney director to favor friends and punish enemies;
- Swiss systems (and round robins also) produce lots and lots of ties, and ties are a pain in the neck;
- Swiss systems require one round to finish (or at least nearly finish) before the next one can begin, and this seems inefficient; and
- Some backgammon players think that playing for prize money is an essential part of any tourney, and so don’t want to keep on playing once they’ve lost enough games to ensure that they’re out of the money.
As for point one, it’s true that Swiss systems are more complicated than bracketed elimination tourneys, and there’s a learning curve to running them. Rules 27 and 28, which govern Swiss system tournaments together occupy 45 pages of the USCF rulebook, whereas rule 29, about round robins, needs only two pages.
One should bear in mind, however, that running a Swiss tourney for backgammon should be much less complicated than running a Swiss for chess. In chess, it’s important to give each player nearly the same number of opportunities to play white, with the advantage of the first move. Chess is a very high-skill game in which one should expect lots of skill progression. And chess tourneys (at least the ones run by the USCF) also organize play according to numerical ratings that represent skill, and are handled so as to introduce a kind of seeding into the tourney. Backgammon doesn’t have to worry about white and black, has a relatively low level of skill progression, and is almost never seeded.
Ties are a problem for backgammon, too, and perhaps an even bigger one. In chess there are draws, which means that there will be more different records. after one round of chess, there are 1-0 players, 0-1 players, and 1/2-1/2 players who got a draw. At backgammon, there are only the first two groups.
The need to finish one round before starting the next is also a problem, though there are some advantages to it, also. One of the maddening things about backgammon tourneys is that you often have to hang around waiting for your next match because you don’t know when your opponent will become available. If you know that the next round is going to start as a time certain, you know whether you have time for a walk (or even a nap) to clear your head.
Finally, the need to be playing for money is a real issue with some players, though not with others. It’s often the better players who feel this way – weaker players who are not expecting to win money often still want to play several rounds.
In the next few posts, I’ll explore ways to deal with these problems.
6 thoughts on “Why Switzerland?”
IIRC High School and College Forensics (Speech & Debate) competitions utilize a modified Swiss that goes like this:
Rounds 1 & 2: Random draw
Rounds 3 & 4: Ranked draw based on results after Round 2.
Rounds 5 & 6: Ranked draw based on results after Round 4.
Rounds 7-10: Seeded Top 16 SE Bracket based on results after Round 6.
I thought this was exceptionally FLEXIBLE: it allows for any number of teams to enter (including odds), gives some schedule notice to the teams, and ramps up/tones down difficulty based on prior results. Finishing with a straight bracket (especially a relatively large one like this) is a great way of “smoothing over” any schedule inequities…the drop line was usually somewhere in the 4-2 teams. 17th place was a huge bummer, but you definitely weren’t one of the best performers on the weekend.
I’ve always got this format and the Schenkel (https://www.curlingcalendar.com/wiki/61) in my back pocket in case I need them – both are extremely flexible, and only require you to work the schedule once or twice, usually with an overnight in between.
Pokemon also runs a Swiss-plus-Single-Elim system for their tourneys. Larger tourneys use Flights for the Swiss portion, but these are only rearranged in the case of multi-day tourneys. Found under Tournament Operations Procedures at https://www.pokemon.com/us/play-pokemon/about/tournaments-rules-and-resources/
The example of Schenkel metzgerism linked to seems kind of odd to me: Yes, a standard Swiss with 30 teams would take 5 rounds, but I’m thinking that a 2-3 or 3-2 setup would serve better than a 2-2-1.
Curling drawmasters avoid scheduling three games in a day. Most players sweep at least a mile per game and walk another 2-3 miles over two hours. It’s not a super-demanding sport, but it is a workout, and players are usually exhausted by the end of their second game of the day.
Even more important is logistics – you can only play as many games each day as you have sheets*timeslots. You can’t add sheets without tearing everything out. You can’t add timeslots (draws) without volunteers/employees to staff the place. Teams (usually) need rest between games, and it’s better to give equal rest to both teams. 6 draws is about the maximum sane number in a day (~14 hours), and you want to finish early Sunday so people can travel back home for Monday work. The maximum number of games I can get in my club over a weekend is 75.
We can’t do the best “theoretical” brackets, but optimizing bracket theory with the physical limitations we have is what I find most interesting.
Ok, Logistics of curling granted, so 3-2 wouldn’t work.
But I still don’t see how an set of groupings would be required for a single round. Saturday night could be used as a buffer to arrange cross-flight matches necessary to avoid repeat opponents in Round 5, but otherwise the Saturday flights could play on Sunday intact. They already partly do so, insofar as Loimaa Country 2010 had 60% of Flight F carry over to Flight I and 60% of Flight D carry over to Flight G; whereas Leppävirta Summer Bonspiel 2015, with 16 teams, had only 2 teams (out of 8) promote from Saturday’s low Flight (D) to Sunday’s Top Flight (F) and with 2 teams(out of 8) being relegated from Flight C (Saturday High) to Flight E (Sunday Low). And, with how the draws past the 1st round are determined, it doesn’t look like anyone who’s not already in the Top flight on Saturday has a chance at the Title.
(This is not meant to knock participation considerations: the very existence of “low” flights on Sunday speak to the value that curling sets on participation.)
The first line of the second paragraph should have read “But I still don’t see how an entirely new set of groupings would be required for a single round.” I guess using “” doesn’t work. Or wordpress uses HTML code…
You’re right, it’s totally pointless – I think it’s the conventional way to operate it in Europe. I’d never plan to copycat the Schenkel (I’ve never had to use it), just start with it and build a reasonable/effective tournament from there. The key concept is 2 games in random groups, 2 games “seeded” groups, then a final game.
My great concern is that some day we have 33 teams show up (instead of 32). Breaking them into three groups of 11 and playing two games apiece is the starting point for that. Afterwards I’d have to make some hard decisions with our schedule because there’s no guarantee there won’t be three teams sitting at 4-0.