Monday, February 03, 2014

In which we consider some of CP's rebuttals on MJP practices

Let’s go back to the original intent of MJP, which is to move the selection of judges out of the hands of the tabroom and into the hands of the teams. Assuming that we agree that the teams’ hands are the better determiners, our issue is how those team hands will work.

My sole presumptions with MJP are that
1. We will create tiers for the judges
2. Debaters will organize their judges according to those tiers
3. Tab will assign mutually ranked judges according to those tiers using a fixed set of priorities (first preference to bubble rounds, second preference to in-competition rounds, third, or no, preference to out-of competition rounds).

There are no systemic limits to #1 above, and I think this is where we need to be clearer. (Here’s where Palmer discusses things at length: Lexington had 9 tiers, including strikes, and 66 judges. Each tier was 11% of the field, or 6 judges. Columbia, with 6 tiers, including strikes, had 51 judges. Each tier was 18%, or 9 judges (and 5 strikes). At Lexington, there were numerous 1-offs in in-competition rounds, because a match could not be found. At Columbia, there were almost no 1-offs in in-competition rounds because a match could not be found.

Here’s what Palmer says: Suppose one tournament has 4 categories and another 8; and the first delivers 100% mutual matchups, and the second has a number of one-offs on the pairing. The second tournament will have delivered the more mutual judging. The 4-category tournament will have many more matchups that are just as non-mutual as those 8 category one-offs. They will only nominally be mutual to the eye of the tabber. There will be more of them, too, as the tab system no longer knows which matchups would have been 1-2s in a 8 category system, and so it can do nothing to minimize them.

When you place a 1-2 judge in an 8 category tournament, you know what you’re doing and you know there’s no better choice. If that same tournament used 4 tiers, then you might place that same judge into the same debate, despite there being a more mutual option which is concealed by the broader, less precise categories. The pairing looks prettier, but at the expense of the missing data which might make it fairer.

If he has 6 in a category, with 1-offs, and I have 9, with no 1-offs, the span of judges according to his math is 12 (my top 1 against your bottom 2) whereas the span of judges for me is 9 (the unvarying mutuality). He’s only addressing the math for no 1-offs. Unless the number of 1-offs is inconsequential, a larger tier makes more mathematical sense. He’s absolutely right in that he may have more precise categories, but the precision is meaningless if he doesn’t have them all the time. Palmer’s numbers only work when there are no 1-offs.

I don’t really think Palmer’s argument is so much to support his math, though—and his math skills are no doubt better than mine—as for getting the most mutuality no matter how it’s done. While I hesitate to call something mutual if it is not mutual, we’re looking for the same goal, agreeable, unbiased judging assignments. I’m not trying to make it easier to tab, I’m just trying to make it work. If I wanted easy tabbing, I’d—I have no idea how to end this sentence. The oxymoron of “easy tabbing” is simply too profound.

I would say that we need to come up with the most precise categories we can, congruent with the various ideas of mutuality and choice overload and the like. If some places use ordinal assignments to solve the same problems we’re trying to solve with MJP, obviously there are other ways of approaching this. I would expect that rather than coming up with an arbitrary number of tiers, we would be better off trying to establish an arbitrary number of judges that make sense regardless of the number of tiers.

I want to throw something else in to this mix, though, before throwing out a number. At some point, we need to have rules in debate that force the activity along lines of certain expectations. Personally, I still believe that picking up a variety of judges ought to be a skill in the debater’s toolkit. The reason sonnets and haiku have limits is to obviate questions about form and concentrate the mind on substance. That may not be the best metaphor for somewhat restricted MJP tiers, but you get my drift. In LD I also believe in things like advocacies on both sides and the debating of resolutions, and while these may not be cutting edge ideas, I am not alone in holding them dear, and at some point putting the selection of the judges into the hands of the teams has to conform to some sort of standard, as does the activity as a whole. If you can’t go to a tournament with 50 judges and find 10 or so you want to be judged by, either change your ways or go to a different tournament. Maybe the problem here is the idea that a “1” in judge preferencing is comparable to, say, giving 5 stars in a movie review. All 1 judges are not 5 star judges. The 1 is not equivalent to inherent goodness; the 1 is a comparative against the field as a whole. If I tell you to watch 50 movies and rate them however you like, it is not the same as watching 50 movies and giving 10 of them 5 stars, 10 of them 4 stars, 10 of them 3 stars, and so forth. But that’s what MJP is. (Suggestions that MJP should be as many 1s as you want, as many 2s, etc., don’t work because then you do get into tab room management.)

I think, as I said above, maybe we’re better off with the idea of a floating number of tiers including a set number of judges. 6 seems to me too small, and the results, limited as they are to Lexington and therefore hardly definitive, show me too many instances of non-mutuality which may or may not balance the probability of more mutuality with the tier that Palmer supports. I offer 8 as a counterproposal. (I’d prefer 10, to tell you the truth, but that’s hardly compromising on my part, and I want to reach a deal here, and I don’t want to haggle. So I’ve pre-haggled: I offer 9, you offer 7, we compromise at 8.) With strikes accounted for, of course. So if you have 62 judges, we give you six strikes and 7 tiers. 96 judges? 12 tiers. 40 judges? 5 tiers + 4 strikes. At the point where you remove the number of judges in the pool from the equation, you’re eliminating the sonnet/haiku discipline I think a lot of coaches want in the activity.

Meanwhile, I’ll go into it next, but mostly what CP said about how he prefs was instructive to me. I wonder not so much how the pros do it, but how the amateurs can do it, going up against the pros. It’s an interesting subject.

No comments: