## Sunday, January 26, 2014

### Mutual judge preference

I read Jim Menick's description of mutual judge preference (M.J.P.) and was surprised about how he would assign judges (the same idea by Mr. Menick also appeared in the Rostrum, vol. 88, issue 2, fall 2013). For those who do not know, mutual judge preference gives every debater at a tournament an opportunity to rate every judge, usually in six categories from 1, most preferred judge, to 6, a strike. After each pairing is set, the tournament uses these ratings to assign judges to each round.

The part that was surprising was how Mr. Menick would rank the ratings, as it were. In order, he would rank a 1-1 (where both debaters rate the judge as 1) as the best option, of course, followed by 2-2, 3-3, 4-4, then 5-5. Then he would circle back to rank a 1-2 as the next best option -- where one debater gives a judge a rating of 1 and the opponent rates that same judge as a 2. Then the order would presumably go 2-3, 3-4, 4-5, then 1-3, 2-4, 3-5, and finally 1-4, 2-5, and 1-5. Mr. Menick's stated objective is to maximize mutuality, that is to say, any possible judge assignment with a difference of 0 in ratings (even a 5-5) is preferable to any judge assignment with a difference of 1 in rating (e.g., a 1-2).

Huh?

I had always understood the rankings would go something like this: 1-1, 1-2, 2-2, 1-3, 2-3, 3-3, 1-4, 2-4, 3-4, 4-4, 1-5, 2-5, 3-5, 4-5, then 5-5. I thought the goal is to maximize each debater's judge preference but not at the cost of screwing the opponent, and thus, a 1-5 judge assignment is ranked below a 4-4 judge assignment. Am I wrong here? Is this not what most people assume mutual judge preference to mean?

As a debater, why do I fill out the judge preference sheet? Under his system, my five most likely options are to get a judge I rated 1, 2, 3, 4, or 5. That seems like a lot of time to invest in considering the judging pool for... not any benefit to me, the debater. Under Russell's system, my six most likely options are to get a judge I rated 1, 2, or 3. That seems worth the time, to get the best half of the judging pool. Actually, if I were a debater and knew Mr. Menick's system was being used, I would not even both consider filling out the judge preference sheet at all. Remember, he points out that if only one debater has filled in the sheet, the judge assigned for that pairing defaults to that debater's picks. So consider the options:

In the two highlighted scenarios, I am better off NOT filling in the preference sheet. Only in the scenario that my opponent and I disagree about every judge am I better off filling in the sheet. How likely is this to happen? Mr. Menick seems to think that the main problem is debating style leading to incompatible judging preferences. I would agree that this happens a lot. But there are also terrible judges out there whom nearly every debater would agree are bad, so the last kind of scenario in the table is truly rare. As a competitor, I am going to hope that my opponents -- even if they have a different style -- will protect me from the truly awful judges and NOT bother to fill in my own judge preference sheet. After all, as the slightly-disagree scenario shows, filling in the sheet only WORSENS my chance of getting a judge I like!

Mr. Menick argues his prioritization of mutuality will lead to debaters getting judges they do not like and force them to adapt. Sure, that is likely true. But why bother with ratings at all? Why not just randomly assign judges? Randomly assigning judges forces judge adaptation at a lot less effort.

Before M.J.P., the tab room decided who the good judges were, and assigned the best judges to the break rounds. M.J.P. was just supposed to avoid the tab room's bias on this issue. So why is M.J.P. a rationale to foist judges I have said I would hate onto me?

Update:

Mr. Menick posted again about M.J.P. I cannot tell whether he was directly responding to me or not, since there were no specific references. Anyway, I think he makes a good case for adapting to lots of different kinds of judges. Which continues to baffle me because it is an argument against any form of M.J.P., not really an argument for his particular form of M.J.P. In other words, if I agree with him, I really ought to prefer random judge assignment, period.