Tuesday, January 21, 2014

In which we meditate more on MJP after Bigle X

Odds and ends on MJP.

You can have as many preference tiers as you want in the tabroom software. There is a school of thought that says the more tiers there are, the more accurate will be your personal assignment of rankings of the judges. I seriously question this. Your favorite ice cream is chocolate, once in a while you like strawberry, and coconut makes you break out in a rash. If I throw in choices raspberry and pistachio, you’re probably okay with that. When you go to the ice cream store, you'll have chocolate as always, but if they're out of chocolate, you'll take strawberry. If they're out of both chocolate and strawberry, you can live with raspberry or pistachio, but no matter what happens, you don't want coconut, not now, not ever. If I throw in raspberry, pistachio, cinnamon, rocky road, lemon, persimmon, pumpkin pie, etc., etc., I think I’m on the one hand creating a false sense of control by giving you more worthless choices in the raspberry-pistachio range, and at the same time, I’m moving away from mutuality because the idea is that cinnamon and rocky road are so similar that they’re virtually mutual choices, as compared to the more dramatic choices of chocolate and strawberry and then the couple of other. That is, instead of insisting on 1-1, 2-2, 3-3, 4-4, or 5-5, now 1-2, 2-3, 3-4, 4-5, 5-6, 6-7, and 7-8 become permissible (on a 9 tier scale) because of the mathematical difficulty of finding mutuality with there are a lot of possibilities. I go back to the basic statement that either it’s mutual or it isn’t, and at the point where it isn’t, call it something else like “Generally Close Opinions Judge Assignment” or something like that. I see clarity in 4 or 5 distinct tiers. The response that someone’s bottom 1s are more like 2s is beside the point. MJP does not give you perfection with a flow pad, it gives you a judge you’ve categorized similarly on both sides of the debate. The alternative is either a judge tab has decided to give you for their reasons, or a random judge. I say, keep it as clean as possible.

Does anyone other than me remember a time when people saw a schematic, saw who was judging, then went on to figure out how to win that judge? MJP should not make that go away, or at least make that sense that “the judge is who the judge is” go away. Assignments of high prefs should favor those still in competition, but that’s it. A lot of people are, I think, still having a problem with that. They’re also assigning way too much weight to the idea of preference in general. A pairing with a 1-5 judge is no guarantee, if you are on the 1 side, that you will win, or that if you are on the 5 side, you will lose. The ranking is only a measure of your preference for that judge. You also sort of have to go into the room and debate the round and convince the judge that you did a better job than your opponent. This is not the calculation of scientific instruments; it’s an opinion poll ranking other opinionators. Judging is not a precise science (ever hear of speaker points?), so how does ranking judges in some artificial order take on any scientific precision? MJP is simply allowing debaters to have an equal hand in judge selection. Any more than that is, well, science fiction.

Moving on. The field in general ranks judges roughly the same, but there are always schools that rank uniquely. I would say that maybe ten percent of the schools in a field, for whatever reason, have few if any matches at all, much less matched 1s. So in a panel, the best that it going to happen is that the math adds up. The problem with this, from some points of view, is that certain pairings get no 1s or 2s, and their entire panel is from their middle area. Of course, it’s from both teams’ middle areas, but be that as it may, the middle is the middle. It was suggested that, if just any 1-1 or 2-2 could be found, the others could be a 1-8 and an 8-1, and that would be fair. Well, mathematically, yes, but it was pointed out, in that case, it’s not really a panel anymore, since you know as certainly as you can that you’ve lost one of the ballots. That’s a good response to that question. Which also hearkens back to my comment above, how once upon a time people simply accepted the judges they got and dealt with them. Three judges of which you have a middling opinion would have happened then as it does now, except now at least both of you have the same middling opinion.

I had a discussion with someone who wondered why all extras, so to speak, defaulted as 1s. In other words, why couldn’t you have more 2s than 1s, or for that matter, all 1s and 3s, or whatever? While I found the possibilities intriguing enough, the unlikelihood of that working out for all the teams and still maintaining mutuality seemed extremely unlikely and led me to respond simply that it might not work in tab, which I don’t think is the greatest answer. Who cares about tab if they have to debate? We’re not there to make things easy for me. It took a while for the real answer to dawn on me. We set up artificial tiers of preference the same way we set up any structural limits on a competition: to ensure a fair, equal opportunity for everyone to break. Certainly forcing the usual 4 or 5 tiers on the field is arbitrary and restrictive, but it is equally arbitrary and restrictive for everyone. Each team has to pick a certain amount of judges out of each tier, with the only exceptions being either pick none at all, which is self-defeating because it puts you at the mercy of your opponent’s preferences, or put them any extras in as 1s, which simply means that you’re perfect happy to be judged by more than the given allotment at your top end, i.e., you like the pool and you’d accept any of them. The point is, we make rules for the sake of fairness. The more fair the competition, the better. The rule, so to speak, of 4 or 5 tiers is standard, clear and, ultimately, probably the fairest.

This does lead to the question of gaming the system, so to speak. I have heard people who claim to be able to do this, plus I have heard people come up with various strategies beyond the simple layering of their choices from top to bottom. As for the former, I don’t believe it. We’re talking debaters here, not Lucasian Chair mathematicians. And if there was a way to game the system, it would not be long before everyone knew it, and the system would be shortly re-rigged of its own accord and you wouldn’t be able to game it anymore. That is, everyone would know how to do it, and therefore everyone would do it, and it become the new system and the gaming would be over. As for not playing straight down the prefs, I’m only slightly less suspicious of this. I gather people will throw a 2 in as a 1 or vice versa, things like that, to catch their opponent’s real 2 with their 1 masquerading as a 2. Well, good luck with that. Prefs are a moving target, and everybody else is getting judges too, not just you and your opponent. If that masquerading 1 of yours is someone else’s real 1, you’ll probably end up with one of your real numbers too.

Of course, all of this game-playing and niggling goes back to denying the underlying reality that a preference number next to a judge’s name does not indicate a win before the round starts. MJP is a tabbing tool for assigning judges according to team preference rather than randomly or according to tab room preference. The more people try to make it something else, the more they either muck things up or, worse, the more they might actually succeed in making it something else. Do we want that? I mean, if you have something else that’s better, fine. Put up or shut up. But till then…

No comments: