sbraunfeld wrote:Check wikipedia for alternate answerlines. In particular, p-infinity should have been accepted for the Prufer group.
Answers similar to Z-p-infinity and C-p-infinity were acceptable.
sbraunfeld wrote:Giving more clues provides some robustness.
I'm going to get better at this whole editing thing. Thanks for the feedback (everyone, not just Sam).
The other bonus part with odd construction was Szemeredi (also a good answerline). Again, we have fairly little information, namely "what used to be called the Erdos-Turan theorem is now named for this guy" (One could have moved the mention of Furstenberg to this part, or described the regularity lemma). Furthermore, players have to recall the leadin (this is the second bonus part) for a statement of the Erdos-Turan theorem (which actually is describing a different conjecture of Erdos and Turan than what Szemeredi proved). For hard bonus answerlines, one generally doesn't need to be coy/stingy with information, and making players recall information from significantly earlier in the question seems poor.
There are a couple of directions I could have gone to make the bonus a little more effective:
1. I could have mentioned something like the Szemeredi regularity lemma.
2. I could have switched Erdos/Szemeredi around and made Erdos the hard part (there are plenty of solid ways to do this).
The presentation of the material was suboptimal and probably made the bonus a little harder to 30 than it should have. Sorry.
sbraunfeld wrote: The 2n choose n question. It is far harder to do computations in-game then when writing a question. My experience with this question is a rather extreme illustration of this point. I am well aware that the max size of a Sperner family of n elements is (n choose n/2), so on the Sperner family clue, I did the calculation to get (2n choose n), and then I think somehow combined these two things to arrive at (n choose n). I realized this was nonsense, and spent more or less the rest of the question trying to figure out what had gone wrong. Now, I think the computation involved here was sufficiently trivial that this was a perfectly fine clue, and a significant take-away here is that I apparently can't fart and chew gum at the same time, but also that computations are really hard, particularly when trying to process other clues, which may themselves be computational (and the remaining computational clues were far less trivial than this one). Quizbowl seems to have agreed that computational math is ill-suited for tossups, and I'm unsure why this exception was made. Also, I would add that, unlike in high school, (advanced) collegiate math doesn't just consist of computations, so it isn't even really like such questions are "keeping it real"; descriptions of proofs strikes me as the better way to achieve that.
My intention throughout the course of editing was to use every submission that I could. The computation was a triviality but I didn't think to remove it, per se. I thought the submission was really creative and so in the end I wanted to reward Max's creativity by using his question. That submission was made later in the writing process so I ended up having to push for a little more math to be in the set. I was successful in this regard. Computational math tossups are bad quizbowl, surely - however this wasn't necessarily one of those.
Theoretically, is it wrong to use such a clue? For instance, let's assume we have a hypothetical tossup on the number zero. Somewhere in the middle clues there is a simple clue using Cauchy's integral formula which points to the answer to the tossup being zero. A simple computation has to be made. Is it wrong for the clue to be used?
sbraunfeld wrote: Terminology is important, and although paraphrasing is nice, it shouldn't come at the cost of correctness. Jacob mentions some wording problems on the first and last parts of the Prufer bonus. I assume he's referring to the use of "union" in the first part, and "direct product" in the second part, which should be "direct limit" and "direct sum", respectively. (In particular, replacing "direct sum" with "direct product" meant that, by a cardinality argument, the answer to the last part would have to be the trivial group, which was very confusing.) In general, be very careful when deviating from given wording in definitions or theorems.
When looking back through my notes, I noted that the group itself could be found by taking the infinite union of an ascending chain of cyclic groups. This is in fact a direct limit and could have been made clearer. On the other hand, regarding the divisible groups clue, Rotman has it as a direct sum whereas I do have it listed as direct product in my notes. My prof has taught the grad level course for 30+ years and is a notable group theorist so I figured my notes were as good as any. I wasn't trying to skate around either issue.
sbraunfeld wrote:P.S. Several clues in the editor's Coxeter question contained material very close to clues to Joelle's submitted reflection tossup (which were replaced in the editing process). This seems poor, in case the miraculous had occurred, and we played in the finals. Was this issue noted?
As it turns out, the Coxeter tossup was written before the submission came in. As I wanted to use as many submitted tossups as possible, I figured I would take out the Coxeter and Weyl clues and give the reflections tossup a more numerical flavor. Of course, I had several algebraic reflection clues in my editor's tossup. Wanting to use the submission to reward the player for writing it, I modified it. As I mentioned upthread, I really couldn't think of many ways to go about writing a CO-level math reflection tossup that didn't specifically use Coxeter clues, however I did remember the stuff about Householder transformations so I used that. Even though I didn't want to throw the reflections tossup away, it probably would have been a smart editorial decision to do so as there was already material of a related nature in the tournament. The material in the final reflection and Coxeter tossups didn't overlap, however some of the material overlapped with the submission. Coincidence? Absolutely. But then again I surmise this kind of thing happens all the time in packet submission tournaments. While the questions were of a similar nature (silly me), there weren't any overlapping clues. Was it a poor choice to include both in the set? For diversity's sake, probably so.
In addition, Coxeter was in the tiebreaker packet and not the finals.
I've included some attachments so people know where I got the clues from. If you are afraid of math, don't download the files.