Tuesday, January 20, 2015

In which we ponder change


Members of the VCA will recall—if they have memories like iron vises—a discussion between CP and me about the number of MJP categories. I’ve gotten pretty used to 6 categories, with 6 being a strike and conflicts being separate. Tabroom allows you to set as many categories as you want. I’ve experimented with 5 for small pools, but found that unnecessary when I’ve compared similar pools with 6.

Palmer’s argument in favor of more categories is simple, that more categories allow for closer mutuality. Imagine 60 judges broken down into 6 categories of 10. My 1 can be 9 away from your 1, and in a 1-2, it can be up to 19 away. With 60 and 9 categories, my 1 is 6 or so away from your 1, and in a 1-2, it can be up to 13. And keep in mind that the scale slides, as if you’re looking at the numbers with a literal slide rule, which is the entire basis of ordinal MJP, another thing entirely. (We’ll get there shortly.) These numbers are clearly mathematically better, and CP’s argument is based on the undeniable math. In practice though, it may or may not work out that way. If you have fewer 1s you have less likelihood of mutual 1s, so you’re more often doing 1-2s (and 2-3s and 3-4s). At this point, you may or may not be getting the benefit of the math anymore. I don’t know. You’d have to look at it knowing not only the ranks in the 1-9 tournament but what the coaches would have ranked in a 1-6 scenario. Impossible.

So is it worth going against the norm? I mean, I wouldn’t do ordinals which, following the math of the slide rule, probably gives you the closest mutuality, because the field is not really familiar with the idea. After all, we’ve only been doing MJP regularly for a couple of years, and there is still a significant percentage of schools who simply don’t pref, for whatever reason. I used to do a whole campaign trying to get them to do it, on the assumption (a good one) that these were more conservative schools who believed (wrongly) that MJP favored circuit styles, which it only did if the more conservative schools didn't pref, a perfect example of a self-fulfilling prophecy. I worked with a number of people to come up with elementary categorization of judges as traditional, circuit or newly trained. We did everything we could, but at some point, something becomes standard practice and it’s no longer our responsibility to insure that everyone understands what we’re doing. Let’s face it: schools that don’t pref now certainly wouldn’t do it if only we went to ordinals. We’d probably have about the same buy-in eventually that we have now. Lord knows, I’ve really wanted to experiment with ordinals because I do believe that it probably renders better mutuality. But here’s the thing. In practice, if the difference between 6 and 9 isn’t all that much and not really demonstrable (even though we know it has to be true that 9 is better), is the difference between ordinals and 9 and/or 6 any more demonstrable or, in fact, all that much? I ask this because you’ve got to take into consideration the users. If I can prove in theory that ordinals is better, does that really matter when I can’t prove it in practice? Users don’t like change, unless they get a direct, measurable benefit. It doesn’t matter what product the users are evaluating. If they don’t see something in it for themselves, they won’t do whatever is necessary to take up the product. That’s why those conservative schools remain resistant to MJP. They don’t see the benefit to themselves of trying to figure out all these judges they’ve never heard of, even with our little crib sheet of Trad/Circ/New. As for everyone else, we’ve got them on board with MJP now, except for the ones who regularly query why they got a 4 and why didn’t their opponent. Is the benefit of a different system—and ordinals is a radically different system while 9 vs. 6 is only a slightly different system—worth the hassle? Do we think users, i.e., debate coaches, are clamoring for it?

Mutuality only promises one thing: that you and your opponent think similarly about a judge. Ordinals probably gives you the closest possible mutuality, but in the end is it all that much different from what we’re already doing to warrant the havoc of change (and all change is havoc)? So many coaches now seem to be convinced that better MJP numbers equates with better results, as if their debaters aren’t good enough to just look at the judge they’ve got and pick up that ballot, period. Everything else is just playing with the data because we can. Should we nurture coaches’ worst competitive instincts? Maybe this would happen. As we move into any newer, deeper system, we lose the older, not-so-deep people. LD has already lost the buy-in of a lot of folks because of its arcane, non-resolutional styles. Should we add to that the most complicated ranking system possible, the one that requires encyclopedic understanding of every pool every week (unless it’s the same old deadbeat college judges traveling from circuit tournament to circuit tournament, the familiarity with which is also in the $ircuit coach’s favor)? You’ve got to draw the line somewhere.

I say, for the time being, we draw the line where we are now. There may be theoretical ways of doing it better, but are they practical? Aren’t we better off locking in, at least for a while, best practices that stay put, rather than always throwing new stuff at people?

This is not a plea from a hyper-conservative for hyper-conservatism. It’s just the ramblings of a realist suggesting that every change made has repercussions, and we need to study and understand the repercussions before we make the next change. 6-step MJP is settling in. How has that affected the activity, if it has at all? I want to know the answer to that before moving to something else.  

No comments: