Join Bridge Winners
Qualifying Swiss Discussion - Part 1

Let me preface this by saying that I am a "true believer" in Swiss as a format. I believe that it is more likely to produce the "right" winner than either a round robin or a knockout in a given time constraint. Having said that, I believe that it is not a particularly good way to get a valid relative ranking of teams unless a lot of rounds are played. Round Robin is a much better format for getting a relative ranking, though not as good as Swiss (or KO) for determining a winner. When you have limited time and a large number of entries, RR is not a possible choice.

When you run a qualifying event you are looking for a relative ranking, an ordering of the strength of teams, with the intention of allowing the strongest "N" entrants continue to the next phase of the event out of an initial field of "M" contestants. What you want, therefore, is an RR among the teams contending for the last qualifying positions. You aren't really concerned with who wins. 

How many rounds should be played?

It was observed elsewhere that "a swiss should have R rounds where R is the log base 2 of the number of teams plus 2". Another person (correctly IMHO) responded that this is the minimum number of rounds that should be played. My own (ancient) simulations suggested that one really wanted about 1.75 to 2.25 times the log base 2 of the number of teams. But these desirable lengths are to produce a WINNER  not for a qualifying event.

I think to achieve the "RR around the last qualifying positions" the more rounds the better. I suspect the pairing methodology should also be changed. But let's leave that for the second discussion, please. As the first issue, how many rounds should be played. A secondary question might be what the tradeoff is between the number of boards per match and the number of matches. 

9 Comments
Getting Comments... loading...
.

Bottom Home Top