From the department of irreproducible results (I hope)

I remain befuddled by Round Robins.

I thought I was sooooo clever. An RR needs two judges per debate. 16 debaters = 8 rounds = 16 judges. We had 16 and 14. First reaction? Go with one judge. The problem with that is that it makes the RR look mostly like a very small batch of set prelims. Given the frequency with which judges split rounds at most RRs, 2 judges makes a big difference. Ballot count is a lot different from wins. Anyhow, in my remarkable cleverness, I figured, how about 1.5 judges per round? That is, sometimes you’d get one judge and sometimes you’d get two. It wasn’t perfectly fair in some respects, but given that I did it at random, it had the inherent fairness of being decided by pure chance. I simply set the computer to 2 judges, and put in 1 or 2, depending on the round. Worked like a charm.

I will point out that much of the RR tabbing process is by hand. I did all the pairings by hand, matching them against my spreadsheets, then I did all the judging by hand, because with this small number, often people are judged by the same judges but on the opposite side. That requires total override of TRPC. But what I count on TRPC for is managing the data once its entered, i.e., ballot count, speaker points and the like.

And then it snowed. While back in NY, people were strolling the boulevards in their bathing suits, up in Massachusetts the preferred means of strolling was behind a team of huskies. As a matter of fact, when we drove home, there was a very clear demarcation on the Massachusetts border where the snow ended and the sunbathing began. I blame Sarah Palin for this, but I’m not sure why. So one of the Round Robinskis was snowed in for a while. Which meant giving a bye. An easy ballot for one person, needless to say. But life is what it is.

But here’s what happened. Since TRPC was set for two judges, the bye meant two ballots, not one. Which never occurred to me, since there were already plenty of missing ballots because of the 1.5. Nor did the inherent disruption of speaker points occur to me. Maybe I was in a daze, but mostly it was so off the radar that it wasn’t until someone looked closely at the printouts that it made a difference. The good news was that the debater had won not by the one phantom bye ballot but by two ballots, including the phantom ballot. Which meant that it actually was a legitimate win. As far as speaker points were concerned, if one were to spend the next six years trying to solve it, there is a possibility that the skew was enough to change the top 3 results slightly. However, since the tournament mistakenly ordered 3 “Top Speaker” awards instead of first, second and third, each of the top 3 was, indeed, a top speaker. The announced order may be questionable, but not the underlying status of topness.

Another fine mess, in other words. In the past I’ve managed to make mistakes, which these days makes me incredibly scrupulous before and during and occasionally after, but seriously now, how many of us have ever given a bye in a 1.5 judge RR? All of this is, of course, O’C’s fault. While we were digging our vans out from the ice and snow with nothing but our fingernails and our true grit, he was lounging about in Las Vegas drinking champagne cocktails with Bette Midler. Next year, I’m going out to Vegas for the champagne cocktails, and we’ll let him cope with the never ending series of mishaps that plague the average winter debate tournament.