MJP cont'd: Stefan clarifies

Stefan's latest:

1 -- Sorry, R3 should be down 1, down 0, down 2. Some people would prefer down 2, down 1, down 0, but I think it is unlikely that people who start 0-2 will clear. There certainly will be exceptions, but as a general rule...

2/3 -- I think I can answer these two together --

I prefer more categories (9/ordinals) to fewer categories (6/3) because more categories makes it more likely that mutuality is protected. I also think it is why 1-2 can be a bit of a whine.

For example, in a 1/1 match on a 6 point scale, I may draw my top rated 1 and you may draw your lowest rated 1. In a 100 judge pool, that's a big spread. Similarly, in a 1/2 match, I may draw my lowest 1 and you may draw your top 2 -- basically, no difference. An ordinal ranking would make that obvious (it also creates many more placement gradations that you know are pretty close to mutual).

As for the percentage of the pool that people should debate in front of, I think each individual person/team should have to debate in front of up to half of the judges (until they are eliminated from clearing). What do to do with the others? (1) I assume everyone's preferences aren't the same, leaving the other 50% to judge other people (they would be in the top 1/2 of someone else's preference sheet). (2) They can judge people below the break (and possibly even those way above it). And 50% is just my own personal preference. I think tournament directors can decide on their own, but they should publish it before the tournament. Wake thinks it should be 95%.

[Side note: I also think elim panels of 5-7 judges are a good idea because they make more judges feel included with little harm to competitive fairness].

I also agree that if you have fewer categories, that makes it easier to give mutuals. That is simply true. The problem I see, however, is that that really just creates an illusion of mutuality. For example, if you make everyone rate 1/2 the judges as 1s, then it appears that you are giving most people 1s in most debates, but that randomly sacrifices mutuality and really just protects the appearance of both mutuality and preference. So, I think you should give as many categories as you can (within technical constraints), ask people to divide the judges evenly on their preference sheets, and place judges with the goal trying to protect as much preference and mutuality as you can until people are eliminated or far above the break.