Week 251: Prep Work like we used to do it.
12 Rounds Charted
During the past week, I have spent my time reviewing the material I have on this year’s episodes. I took apart each episode from this year and did some comparison work against last year’s episodes. We are exactly at the point where our episode aired last year, so we have 12 episodes of evidence. You can look at the evidence as easily as I can, as they’re all available online.
Before we go into this, I want to revisit a couple of points as to my theories of how this sort of analysis matters for studying past results.
Everything in quiz bowl repeats.
Everything repeats, but frequency of repeats matters more. Things that repeat over seasons can be prepared for. Things that the writer repeated in the season already may come up again.
Answer selection is much tighter in televised competition because they want to avoid all manner of ambiguity and a round of stumping questions is not good television. Even worse is a round of answers that the audience hasn’t heard of.
The writer isn’t writing the whole season in one sitting. They will ship the questions in sets throughout the year. They may repeat themselves accidentally between shipments, or they may split similar questions between shipments so the audience and editors don’t realize it.
If the writer can’t fit a category into their particular question format, it won’t show up. And that means you don’t need to teach it for this event.
The writer can show their preferences to write by what they go after first.
I watched them straight through in two three-hour sessions, which allows you to notice patterns that go from round to round, and from shipment to shipment. Last year I had settled on the idea that it was a six-episode shipment, meaning we got the last round of shipment 2. The one piece of information I had gotten from the email was that our taping was episode 2618 this year, versus 2512 last year. So for our purposes it’s the same spot in the shipment. Last year that meant it was the hardest round in the shipment and the 60-second round was an experimental format, which depressed the score in the round.
Surface details
There are three changes to scoring of rounds this year. First, all teams begin with 100 points. This avoids needing to display negative numbers to the audience, and makes the rounds look a little closer. Second, the second round of picture based questions has lost a question. This also avoids negative numbers by making the worst loss you can take in that round -80, This may make it easier for the writer. The third change is in the final round, a decision to award a bonus 25 points if the team pulls two 30’s on 30-20-10 questions, guessing both questions correctly. This makes it more difficult to have a runaway finish, as a team with a 100 point lead could now be caught, even if they don’t know either answer (they’d lose 10 on each question by answering incorrectly on the last clue, and their opponent could pull 85) but the chances of a team pulling that off from behind is still very slim. I think these changes will modify the psychology of the game, rather than actual results.
To a certain degree there has been a small correction to last year’s questions in terms of subject matter. And to that certain degree I think it’s been a reaction to trying to work in the choice of media last year created. As I described it last year there were a number of questions which relied on using clues from the image presented and clues from the moderator’s reading. That made it easier to do questions on fine arts, geography, and math, but harder to do questions on literature, music, and history.
There’s been less art, and more music. I kind of expected that, partly because the writer did go pretty deep into art last year, and barely touched composers for the bulk of the season. The format’s focus on having a visual to go with most questions hurt composers, and mostly moved them to the 30-20-10 rounds.
They’ve actually made an effort to include literature in the visual questions since last year. I’ve seen three questions using book covers and masked out titles, and two questions which used the first page of a book, which means first lines are visible. That’s a pretty good improvement which helps teams.
I’ve noticed that the question have had longer reading in this season. That’s a return to form where some of the previous questions had the image giving essential information not included in the text. That feels like expediency to get better results for all teams. I may have appreciated the effort, but I also recognized that it is an effort, and that you can’t do that forever. Including the clue which reinforces the image is probably necessary for all teams to be able to thrive.
As with last year I’ve seen a few rounds with subject rounds that have repeated. State Flags and Current Events appeared twice, and I think that those were originally written as a single ten-question lightning round, split in two to fit the show’s format. I’m going to be especially vigilant for those patterns, since they usually are split across shipments. The one obfuscation they’ve employed this year which makes me think they realized this problem is that the subject rounds’ subjects are much more generic. We’ve had lots of “History” (no modifier to limit it) “Science”, and “Literature” as the subject round choice, which almost makes me think the 60-second rounds which needed 30 questions to be written, have been modified to fit demand. This also makes it more difficult to get a real feel for the subject of the round, but it makes it much easier to know which subjects you want to avoid.
I’ve also noticed that questions this year have made multiple uses of the same image. We’ve had two questions where a diamond was the image, and two questions where Shakespeare was the image, but each case had different answers. I think this is a practicality of getting things done. It’s easier to have an image cleared once for television, and write multiple questions around it, than to find the right image for each question. It seems like a necessary reduction in effort by the writer and editor. What I have not recognized so far and would expect to be a similar expediency, would be to use images cleared in the previous year used for things this year.
We’ve had two subjects rounds of current events, which were actual current events, and not pictures of longtime newsmakers. That’s something new, and something I have to be concerned with. A round of current events questions to the team covering the last month would probably cover a one in six chance of hitting it.
The math hasn’t changed and I feel confident that I’ve nailed the material I mentioned on the Pocketmod I gave to the team. I’ll run through it in my training material, but at this point it’s been three points of review on the things that matter. And their font is reasonable this year, so we’ve already sighted the factorial in a math question.
Science questions have actually improved this year because they’ve figured out a reasonable source for diagrams. This means two of my things I expected to show up last year came up in spades this year: Bones were a subject, and they’ve used an actual periodic table for questions.
The one thing I expected to see in small amounts this year, but only saw once, was AI-generated images. I had to figure that eventually an AI image demonstrating something would come into play. The only place I’ve seen it so far was in an attempt to have an image of the Sun Wukong, where the answer was monkey. I’m a little concerned for that possibility, but it was effective in putting an image forward that was a viable clue that could be used.
Changes to the ‘distribution’
I say that because I know there isn’t a distribution per se, but there is a choice of the writer in what subjects they choose to emphasize and move forward.
Social History — Of the 12 60-second rounds I’ve seen half of them have been history, and half of those were what fell under Social History in the distribution I’m most familiar with. The first two episodes had rounds focused on Women in History, and then the Civil Rights Movement. The eleventh episode used an unusual decades format to split between teams, the first team got the ‘40s and ‘50s, the second the ‘60s and 70s, and then the ‘80s and ‘90s. The subjects in the first two episodes 60-second rounds were covered again in the eleventh. That’s definitely a shift of emphasis, and it’s a shift in showing what you want to do with the opening rounds. I have to take that into account.
The interesting bit of this is that the 60-second round is where you can slide out of the televised answerspace more easily, because the question goes by faster, and because it doesn’t require an image to be displayed. There’s no expectation that all ten questions are of the same difficulty for each team, only that the teams hit roughly the same number of hard or answer expanding questions.
Space — Last year, I was expecting a lot of space questions and astronomy questions based on the material of previous years, only to not see it in the episodes aired prior to our match. And though we did get one question from images from the Webb telescope, we were anomalous in our sample size. The Space and astronomy category was the third show’s 60-second round and it dug deep into history of the space program. The manned space program also made many appearances in the Decades round, and in the picture rounds as well.
This seemed like a conscious choice, and one that was to put things forward. It was probably a play to the expected audience, but more to play to the producers of the program, who might have seen the absence of astronomy questions last year in the same way I did, unexpected.
The only 60 second rounds which really clanged were one where you had to give square roots of perfect squares where the answers ranged from 2 to 50, and a second one which gave numbers and you had to give the president. Both of those came out of very orderly information, but were pretty rough. The former was really easy if you had learned a simple rule (the last digit of a perfect square indicate the last digit of the root, and if you knew that 10x10=100, 20x20=400, you could place the tens digit), and sucked time if you didn’t. The other was how tv shows had given last clues for presidents for decades, but I’ve never seen it done that way, and it read very awkwardly without associated pictures. Both of these did what the “complete the title” lightning round did for us last year, it shortened the field and capped the range on that round. If we get one like that, we know what we’re in for.
Stylistic changes
Three remaining oddities from questions that may reflect stylistic changes. The stacking of all the presidents←→ number relation into one round, and then not really using that clue ever again sure feels like a “we’re not going to do this anymore” decision. I’m okay with that, as I’ve tried to use other clues to teach my team the order of presidents (relative position, succession, being in a predecessor’s cabinet, or running after leaving office.) If they aren’t using the tool, that’s fine. Teams will adjust, we’ve already adjusted.
They’ve also seemingly made a conscious decision to not use state capitals as a universal last clue for states. Three questions used senators and governors, past and present, for the link to the state as an answer, and a further two used a format of [college] in [this city] to clue the state. I’m pleased by that, even if it undercuts one of the earliest drills I put into practices. Note this does not apply to nations, those are still reliably pointing to capitals as giveaways.
The third stylistic change has been to get geography out of the political entities and mountains game, and broaden out answers into other categories. I saw two questions on rivers, which pointed to tributaries which emptied into rivers as clues. This is kind of a classic late clue of the circuit which has migrated back into televised questions, and it’s something I can easily prepare the team for.
Even more than the choice of emphasis in the early rounds, the stylistic changes and the reinforcing of image clues in the text indicate to me some modification of the writer staff, most likely a broadening of the base. Last year was surprising for how they were able to commit to the bit and push a vision into the format, and these 12 episodes have seemed like a bit of return to the previous format. But with that I think it’s more likely that additional or different editors have been brought in.
I think we’ve got a good shot, we’ve got the advantage of experience now, and many of the teams from last year have already been matched against each other and bowed out in the first round. I think I have a better handle on the writer(s)’ intentions this year than last, and I feel less panicked about it two weeks out than I did four weeks out last time. Of course, it’s very easy to feel that on New Year’s Day. Plenty of time for things to go wrong.