A few thoughts on my categories based on the detailed stats:
I was overall quite happy with how both of my categories played, with pretty reasonable get rates (both at 87%), buzzpoints, and bonus conversion. Neg rates varied pretty widely from
memory with a 0% neg rate to a fairly-surprising 47% neg rate on
Los Angeles (I guess people reflex-buzzed Schoenberg -> Vienna?) but overall most tossups seemed to elicit a pretty normal neg rate.
In classical music, the questions on individual works seemed to play pretty well in some cases: I was quite happy with the very smooth distribution on both
Salome and
Mahler's 1st symphony, especially given that there is not an overwhelming number of past tossups on either of these answer lines. On the other hand, tossups cluing from "deeper" clues about famous composers, like
Haydn symphonies (which clued largely from the Paris symphonies) and
piano sonatas (clued largely from Schubert) seemed to not elicit many early buzzes. I suspect this may be because the large output of the Classical-era composers means that even pretty famous pieces have been clued comparatively infrequently, especially if there are more obvious choices (the London symphonies or Beethoven piano sonatas, in each of these two cases). The
lutes tossup may have been too novel/unusual of an answer line and did not play well - that question should probably have moved into European clues at least a line earlier.
The classical music bonuses generally played well, with quite a few hitting very close to the 90/50/10 ideal. The easy parts possibly could have been made slightly harder as a 94% overall get rate was among the highest of any category; that being said, I was glad that none of them played particularly difficult.
Organum played as a hard at this level, which was probably the only real surprise for me.
In social science the lead-ins and second lines seemed to generally be too difficult overall and garnered few buzzes (although the middle clues seem to have played well). This was not too surprising to me; social science is one of the most disparate 1/1s and there are relatively few deep experts on these topics. Some of the topics, like
common law,
mothers in anthropology, and the
Neolithic Revolution were also pretty novel, with few or any of these conceits having been tried before.
Difference-in-differences unsurprisingly played as one of the most difficult of the set with 41% conversion, although it still generated a pretty good buzz distribution.
The bonuses in this category were less successful, even though the aggregate conversion rates were still fine. Particularly bad was the
Needham/Landes/coal bonus where Needham (shockingly to me) only got 5% conversion and Landes 0%. The bonus parts on
markups (inspired by seeing
Kunal Sangani's job market talk),
strain theory and the
A-not-B error also got 0% conversion, more surprisingly in the latter two cases. A few bonuses, like the one centered on Afrobarometer and Latinobarómetro, may have also been a bit too random in topic choice, leading to confusion, although conversion numbers were OK.
An overall remark is that I was glad to see lots of buzzes on the pre-FTP in a lot of my tossups across both categories: I hope this means that the set was effectively able to distinguish teams across various skill levels. This I think should be the ideal for a set like Regionals is may still be played by less-experienced teams seeking to try their hand at the flagship 3-dot difficulty event of the year.