Monday, May 16, 2011

Doubling the brackets

This is the theory. The bracket should not be set every round, but every two rounds. That’s what we discussed on TVFT last week. Let’s look at it.

First, there’s the neg bias. Let’s not worry about the cause of the neg bias; there’s differing opinions on that. But the evidence points to it clearly. In random rounds, or at tournaments of relatively inexperienced debaters, the better debater usually wins. It doesn’t matter if the better debater goes aff, neg or a little bit of both. Class will out. Side doesn’t matter much, except that perhaps less experienced debaters have trouble with ARs. But less experienced debaters aren’t so hot with negatives, either. It all irons out.

As we start hitting brackets, however, we start, by design, evening out the skills of the two debaters. The evidence says that if two evenly matched debaters hit, the neg is more likely to win. No matter how you interpret the evidence, there is little question that the neg does show a statistical advantage.

Bietz’s thought was that the third and fourth rounds should be paired together, based on results of rounds one and two, rather than sequentially. In other words, after the first two rounds, we have brackets of 2-0, 1-1 and 0-2. These brackets should now debate within their bracket on both sides, not just their next due side. If we don’t do this, it means that round 4 is more closely bracketed than round 3, and the neg in round 4 has a bigger advantage because of being more evenly matched. Pairing 3 and 4 together minimizes that potential advantage. It’s not perfect, but it does make intuitive sense.

How you determine the bracket then comes into question. Normally we go by speaker points, but if it’s just two rounds, that’s a pretty dicey construct, because there’s not much to go on and it can be skewed. Opposition wins? That is, if you won two, but your opponents lost their two, you had an easier draw than, theoretically, the person who won two whose opponents both won their other round. You should hit someone with a roughly equivalent previous draw. And while speaker points are arbitrary, until at least there’s mathematically enough of them to drop a couple, winning and losing is clear-cut.

As Bietz says, this is the stuff that keeps us up nights. I mean, while you’re out there eating Cheetos and shooting craps and, occasionally, debating, we’re trying to keep things honest and tab the best possible way we can. Granted, sometimes what we do is also a crapshoot, and we like Cheetos as much as the next person, but it is a different business. Too inside-baseball for the room?

5 comments:

Unknown said...

Another interesting thing is that pairing round 4 based on rounds 1 and 2 it makes it so that the bracket you are in is not based on whether you affirmed or negated in round 3. If you negated in round 3, you're getting put in a harder bracket, and you debate someone who won an aff round (ooh, ahh), theoretically a more difficult draw (because they have shown more skill by winning that aff round). If you won your aff round in round 3, you get an "easier" draw, because it's just someone who won the neg round - not the aff round. You are then benefited for affirming in odd rounds. I'm not sure how much sense that makes, but I think it's probably a good thing that rounds are paired based on everyone having an even number of each side, not based on some people having negated more than affirmed and vice versa.

pjwexler said...

Typically, I dunno. Bringing back lag pairing from the tabbing-by-hand days as a proactive 'good' thing seems dicey. It would also seem to institutionalize the idea that the neg bias is an inherent part of the activity, rather than a factor caused by current strategies and mindsets.

If bias has become inherent, than maybe in rounds 7-8 this would be needed. But if the neg bias is so great that we have to do something this radical as early rounds 3-4, to me that suggests greater problems in the activity.

Unknown said...

I don't think it's true that this assumes neg bias is an inherent part of the activity. Rather, I think that it tries to make the pairings more fair in terms of side bias in general; if, on a certain resolution, the aff seems to have an advantage, this has the same effect. If a certain debate just likes to affirm more than she likes to negate and her opponent feels the same way, this also has the same effect. The idea, as I see it, is to make pairings based on only instances in which debaters have debated the same number of each side.

More importantly, though, I think that we should try to correct for biases that we see in the activity, no matter how much it is caused by current strategies and mindsets, not just ignore them and say they should be fixed in other ways. This is the way that strategies are evolving, and so if that means altering tabbing so as to make sure that the best debater is coming out on top and that pairings are not affected by such biases, then I think that's what we should do.

pjwexler said...

I hear, what Sophie is saying. I just think that the neg bias problem can probably be fixed through better Aff strategies.

If that is not true, maybe in three years or so it might be wise to alter the pairing guidelines. I'm reluctant to alter them based on trends over a comparatively brief period.

True, I may just be channeling Eddie Burke. But in the event if an Aff renaissance, the shift would be for naught. And preferences and techniques DO change.

Unknown said...

In the event of an aff renaissance, I think that what I said is still true. The point is not to correct for a neg bias that exists in the activity at this moment (although it will have that effect), the point is to correct for the randomness that occurs when some people debate one side in odd rounds and some debate the other side.