The Quizbowl Resource Center

By popular request.

Cody wrote:I would also like to have a public discussion of the extremely unfair practices of: carrying over all prelim games and allowing a team down 2+ games to enter a finals and win the tournament by winning just two games.

Well, what exactly is NAQT's rationale for this practice?

I'd also like to talk about the practice of playing off placement ties with half-packets, when enough full packets were available. I could understand if Yale vs UVA had to be an advantaged final, but since that was going to burn only one packet anyway, there was absolutely no reason that UMD vs Chicago/Penn vs Chicago couldn't have each been played on a full packet.

vinteuil wrote:Well, what exactly is NAQT's rationale for this practice?

Primarily that "dropping a prelim game to a relatively weaker team is a real measure of team strength."

The Quest for the Historical Mukherjesus wrote:I'd also like to talk about the practice of playing off placement ties with half-packets, when enough full packets were available. I could understand if Yale vs UVA had to be an advantaged final, but since that was going to burn only one packet anyway, there was absolutely no reason that UMD vs Chicago/Penn vs Chicago couldn't have each been played on a full packet.

Noted, and we will consider adjustments to this policy in the future. (Note here: it's very possible that second-place ties might need half-packets because the alternative is running out of packets.)

bt_green_warbler wrote:Noted, and we will consider adjustments to this policy in the future. (Note here: it's very possible that second-place ties might need half-packets because the alternative is running out of packets.)

I completely understand. People may disagree, but I think having a conditional policy in which full packets are played if available (switching to half if not) makes sense.

The practice I'd be most interested in seeing a justification of is "If one team has an advantage, it will need to win one game while the other team will need to win two", which in theory could produce even stranger final outcomes than the 2005 D2 situation (e.g., if the first-place team somehow emerged from the playoffs with a three-game lead over the second-place team, they played a two-game final, and the second-place team won, the winner of the tournament would have a worse record than the overall second-place team).

The Quest for the Historical Mukherjesus wrote:I'd also like to talk about the practice of playing off placement ties with half-packets, when enough full packets were available. I could understand if Yale vs UVA had to be an advantaged final, but since that was going to burn only one packet anyway, there was absolutely no reason that UMD vs Chicago/Penn vs Chicago couldn't have each been played on a full packet.

Also, the non-timing of the half packets is pretty much an admission that the clock is dumb. I suppose the relevant model is one in which there's a tradeoff between variability of the outcome due to vagaries of which questions are asked in which part of the packet/half of the match vs. something worthwhile in the "excitement" of timed quizbowl. In that case, the judgment that entails the outcome "time full-packet matches but not half-packet matches" is a completely arbitrary decision that match outcomes decided on half packets (rather than a single ten minute "half") would be too noisy a signal but that twenty minutes is a long enough time to play quizbowl such that the noise is reduced below some threshold. Pretty much nothing convincingly locates that threshold on the interval (10 minutes, 20 minutes). And obviously, not timing those half-matches gives them a very different feel from the rest of ICT.

Also, why was the seeding among the three teams with tied records decided on PPG rather than PPB, when they'd faced different fields?

bt_green_warbler wrote:Primarily that "dropping a prelim game to a relatively weaker team is a real measure of team strength."

The problem with this is that the other teams in the playoff bracket did not play the same prelim teams, so it's unfair to carry over the games against teams they don't have in common. This is pretty much analogous to the well-accepted reason you don't use PPG to compare teams who didn't play common opponents: it has become a meaningless statistic [of course, apparently NAQT doesn't really believe this either....].

This comes into play even if you have perfect brackets, of course, but it's an especially huge problem at ICT because the brackets are always a huge mess and are never seeded anywhere close to "almost correct" (which is more of a subject for the other thread, to be fair).

Tees-Exe Line wrote: Also, the non-timing of the half packets is pretty much an admission that the clock is dumb. I suppose the relevant model is one in which there's a tradeoff between variability of the outcome due to vagaries of which questions are asked in which part of the packet/half of the match vs. something worthwhile in the "excitement" of timed quizbowl. In that case, the judgment that entails the outcome "time full-packet matches but not half-packet matches" is a completely arbitrary decision that match outcomes decided on half packets (rather than a single ten minute "half") would be too noisy a signal but that twenty minutes is a long enough time to play quizbowl such that the noise is reduced below some threshold. Pretty much nothing convincingly locates that threshold on the interval (10 minutes, 20 minutes). And obviously, not timing those half-matches gives them a very different feel from the rest of ICT.

Forgive me if I'm missing something here, but I think that I disagree. Not every tournament is going to have DI ICT readers. UCF was at two tournaments at the same site in the not distant past where the readers were slow. In the NAQT tournament, neither UCF A nor B heard a full 18 tossups per game (B never heard more than 19 tossups and A heard more than than only once). At the ACF tournament, we cleared the entire pack like we were supposed to. Despite a delay at the NAQT tournament waiting for a no-show tournament, we still finished something like two hours before we did at the ACF tournament.

When you're at ICT with ICT readers and the events can take the entire day without a problem, and the match is going on at around the same time as a more important championship match, you can kill an extra 3 or 4 minutes and no one cares. If you're untiming the entire tournament from beginning to end (which I'm not opposed to) you can stretch the overall tournament time a couple of hours.

Tees-Exe Line wrote: Also, why was the seeding among the three teams with tied records decided on PPG rather than PPB, when they'd faced different fields?

That's always bothered me too. That and the playing half-packets when you can play a full pack are the two things coming out of this thread that I think are really worth reconsidering, personally.