Should ACF tossups be shorter?

[Santa Claus]

Glancing briefly at this year's ACF sets, Regionals and Nationals tossups were around 8 lines. Would it be better to have questions which were roughly one line shorter?

Some arguments for this shortening:

• Fewer clues means less writing to do.

This strikes me as particular relevant due to packet sub. Every discussion of the practice I can think of has mentioned how daunting teams find it. Packet sub is a valuable institution, even in the reduced state that it currently exists in, and if we're going to commit to keeping it then reasonable steps should be taken to ensure that the barrier to entry is as low as possible (without compromising the exercise).

Editors have to do a lot of writing from scratch as well, so they benefit also. Optimistically we can say that 13% less writing is necessary (though of course this is very optimistic).

• Fewer clues means less editing to do.

The precise amount of "less" is again debatable - there are obvious costs to achieving concise questions - but this should at least be mentioned.

I feel like the task of editing has gotten a lot harder over the years as standards have risen, something that has probably contributed to the task being split among progressively more and more people. I don't think it's fair to simply expect editors to keep stepping up to this challenge - despite all the new names working on Winter and Regionals, it feels like only a few people are routinely willing to take on the high-stress, high-effort jobs like editing Nats. I have hope that the situation will improve and this will resolve itself, but I feel like it behooves the community to make things easier if possible.

• It is implied that there would be a small impact if lead-ins and other early clues were pared down, making them obvious candidates for removal.

The buzz point data that exists implies that the majority of first clues are not buzzed on by any players in the field - this includes at Regionals, which has both larger fields and (theoretically) easier questions. The simplest interpretation is that these clues could simply be removed.

• Shorter questions are played faster.

A nice treat for readers and players alike.

• Editors often have a sense of what can be cut.

I don't wish to overgeneralize, but in my experience I've had many more situations where I struggled to find the clues to finish a question than I have questions which were brimming to the point I had to pare them down. Additionally (and perhaps consequently), I've always known what the weakest clue was and it was often the one which was shoehorned in for length.

Some arguments against:

• It is not possible to just remove the lead-in/second clue from every tossup.

I am personally convinced that early clues at higher difficulties are "stochastic" rather than strictly pyramidal. I certainly don't mean to suggest to editors are not trying to order clues as best as they can, only that if multiple clues are below a certain conversion threshold then it becomes largely random whether a player will know one over another. This is, in mind, an argument for keeping the clues - without concrete data that simply doesn't exist, it's not really clear which (if any) clues can be removed.

• Early clues provide something to both the players hearing them and later readers of sets.

Many players say that clues they don't buzz are still useful for providing context and what not. Seems perfectly reasonable, if hard to judge.

On the second point: I enjoying reading packets after the fact but I think it is silly to prioritize that. Then again, I don't think anyone is seriously arguing this.

• This would mean ruining perfectly good questions.

It seems mathematically certain that not every question could be shortened without compromising its integrity. This means things like ruining pyramidality, introducing cliffs, breaking sentence structure, etc. I guess these could just be left at 8 lines, but I do also question what fraction of questions would really fall into this category.

I do think it's worth noting that the thought experiment of "remove one line from existing questions" is not a perfect model: having shorter questions would simply produce different questions, some of which would closely resemble longer questions with one clue removed. I don't expect this to have the extreme impact on answerspace that result from the strict limits of ICT, but perhaps some ideas would become more feasible.

• There's no guarantee that this would actually improve things like lead-in difficulty.

This is one advantage that the "remove one line" model has: it explicitly handles the problem of lead-ins being too hard by explicitly removing them. This could be replicated in the mind of the editor, but it's a hard problem in general.

If not done somewhat deliberately, 7 line tossups could even have the same issues around lead-ins that 8 line tossups have. Perhaps then we'd argue for 6 line, then 5 line, and so on.

Conclusion
It seems like a lot of the arguments on either side are not based on hard facts - how could they be, when there are so few? In lieu of such empirical evidence, we might as well yell about it. Here is my piece; I welcome others.

[Mike Bentley]

In general, I think quizbowl would play pretty similar if tossups are shorter. We already see this in ICT finishing places being reasonably similar to those of ACF Nationals.

As a writer and editor, I do enjoy having more space to work. I personally find it harder and more work to craft something like a 6-line tossup than a really long one (which is why ARCHIVE will be on the longer side). You need to think about each clue a lot more. But as a player, it's rare that I play a shorter event like BHSU or ICT and feel like I really needed that extra space to have a good time / decide the winner fairly.

[Santa Claus]

I have been reminded of a previous thread in this vein by Victor Prieto in 2017, which has lots of relevant discussion as well. Victor was kind enough to provide some links to other, older threads in the original post as well: 2009, 2012, 2012.

[theMoMA]

Questions could probably be a little shorter. That said, I don't think eight lines (or seven, for matter) is an unreasonable length for a tossup. In any event, almost any eight-line tossup can be edited down to seven while keeping all of the clues and most of their substance, because most writers and editors are not very good at getting the most out of the space allotted. In that vein, I'm not sure that it's actually less work to have shorter questions; if you want those questions to be good, you'll be working pretty hard to pack in as much material as possible into seven lines, or eight lines, or whatever number of lines you have. I don't think that mechanically removing one of the first two clues in each tossup would be the best way to shorten any tournament's length.

In my experience, chatty bonuses are a more common problem than an extra line of tossup material, because bonuses take longer to play, have multiple prompts and multiple answer lines that create stops and starts and make eyeballing length difficult, and are generally just easy to forget about length-limiting. Limiting bonuses to seven or eight lines of prompt (roughly two for the lead-in and two for each prompt) would go a long way toward tightening up round lengths. (More than removing a line in each tossup, which would likely save just 150-200 seconds of gameplay time.) It doesn't even have to be a hard limit, if there are a couple bonuses that deserve a few extra lines to reach their full potential.

It's kind of amusing to read threads that imply (although I doubt this is Kevin's ultimate point) that it would be nice if the quizbowl were shorter so it would just be over sooner. Yes, it's bad when questions are so long that they're not appropriate for the audience or the readers. Yes, not every event has to have long questions with many difficult clues to luxuriate in. But there are lots of players who enjoy playing lengthy questions with challenging early clues. It's fun to hear questions that are trying to demonstrate an interesting conceit or set the table with clues that are more contextual than imminently buzzable, and that's just a totally different playing experience than a tournament where you have to be ready to buzz right away on every single tossup. This kind of quizbowl becomes more a thinking person's than a remembering person's game, if that makes sense. I certainly enjoy both facets of quizbowl, but I especially enjoy the thinking side of the game when the questions are entertaining and well written (and this is what I strive for when writing events such as CO Trash, as well).

[tpmorrison]

For what it’s worth, both of the Regionals that I’ve worked on (2021 and 2023) have had seven-line tossup caps. When questions pushed beyond this, it was almost always either because proofreaders extended the wording a bit for clarity or because pronunciation guides added a line or so.

[eygotem]

It's also important to consider that some answers could only be tossed up with shorter questions since there wouldn't be enough good clues to fill out a nats length tossup, an issue I've run into in the past. Also, especially for lower bracket teams, it would be nice not to have to listen to facts about a topic no one in the room knows about for as long as we do currently (which also takes some pressure off writers with optimizing answer lines, since a dead tossup wouldnt affect the experience as much). The 500 character tossups of ICT may be hard to do for everything but the present standard for ACF and housewrites is at the other end of the extreme I feel (even if things have improved from the 13 liners of a decade ago). I do agree that bonus length is a bigger issue though, it's such a slog especially as a reader when every part is so verbose.

[naan/steak-holding toll]

Seven lines seems like plenty to me at most difficulty levels. If cutting lines also meant more games, then I could see a good case for making Winter and Regs into six line affairs. I recall editing DEES in 2014 and not having too much of a an issue making six line regular difficulty questions, though they weren't the most creative.

I do think seven lines is very helpful for giving a bit of time to tee up context or throw in some fun anecdotes here and there.

[gyre and gimble]

But there are lots of players who enjoy playing lengthy questions with challenging early clues. It's fun to hear questions that are trying to demonstrate an interesting conceit or set the table with clues that are more contextual than imminently buzzable, and that's just a totally different playing experience than a tournament where you have to be ready to buzz right away on every single tossup. This kind of quizbowl becomes more a thinking person's than a remembering person's game, if that makes sense. I certainly enjoy both facets of quizbowl, but I especially enjoy the thinking side of the game when the questions are entertaining and well written (and this is what I strive for when writing events such as CO Trash, as well).

This feels important at higher difficulties. To condense Nats- or higher level tossups into 7 lines would devalue the skill of lateral thinking and "figuring out" questions. I think these are very important skill. I'm not sure if people still do this, but back when I was an active player, the comment "you figured it out!" used to be a sort of backhanded compliment. I always found that annoying and hope that attitude has changed. You figure stuff out before others do because you know more than them. It is a legitimate demonstration of knowledge, and even deeper understanding of the subject matter.

It seems to me, though, that in order to properly reward this skill, questions need to be long enough for players to have time to process clues, narrow down possible answers, and make educated guesses. And the harder the difficulty, the more time that's needed to do that. I would hate to see Nats or CO tossups reduced to less than 8 lines for at least this reason.

[eygotem]

Idk having played my share of NAQT sets it's very possible to test lateral thinking with shorter tossups, if slightly harder. Surely we don't need eight whole lines to establish context especially when many teams might not know enough to even buzz in before the giveaway if at all.

[theMoMA]

I think people are talking at cross purposes to some degree. Seven lines is probably better for ACF Regionals, and eight lines is probably better for ACF Nationals and Chicago Open. Six lines is probably better for ACF Fall. 500 characters (roughly 5 lines), with more questions and more categories, is probably better for SCT and ICT. I've edited each of these tournaments to exactly those length limits (2008 ACF Fall, 2014 Chicago Open, 2017 ACF Regionals, 2018 ACF Nationals, and numerous SCT/ICTs).

The reason these length limits are better for these tournaments is not because of some grand principle encoded within quizbowl. ACF Fall gets six lines because it's meant to feel like a tournament with relatively easy and brisk yet still circuit-style tossups. ACF Regionals gets seven lines because it's a meant to feel like a tournament with moderately difficult tossups that can sometimes be gotten quickly but often need to be worked at. And ACF Nationals and Chicago Open get eight lines because they're meant to feel like slow, contemplative, difficult tournaments where each converted tossup is an achievement unto itself. And NAQT events are their own genre, meant to feel fast-paced and high-adrenaline, with all of the contemplation and action over each question condensed into a much narrower time and space.

These tournaments are not meant to feel like that because quizbowl has data saying that this is what the players really want or some other objective optimizing principle. Rather, they feel like that because it's what it seems like they should feel like, based on the historical progression that has produced the game as it exists today. You could write CO with three-line tossups and one-part bonuses and be done by 2, and one of the best teams would probably still win, but it wouldn't feel like CO, because CO has always been a long, grueling, and contemplative event that rewards thinking and endurance as much as speed and recall; consequently, that's what the players enjoy and the editors strive to reproduce. Similarly, you could write a much longer and more intense version of ACF Fall, and the best teams would almost certainly still win, but you would confound the expectations of the hosts and players in the process.

It's worth noting that there's nothing objectively "short" about a six-line tossup (which, as I'm sure we've all seen, appears impossibly long to any tyro) or "long" about an eight-line tossup. These things are purely relative to the standards that now prevail after decades of editing and posting. So if I went to a CO now and played seven-line tossups, I'd probably think things seemed a bit brisk and terse for my liking, because I'm used to enjoying a certain kind of CO experience that's at least a little more languid. If I had to read eight-line tossups at Fall, I'd certainly register my displeasure for the extra vocal strain. But it's all relative and arbitrary, and I'm sure you could write a CO that's a little shorter or a Fall that's a little longer and mostly hit on the right feel. I'm skeptical, however, of the idea that there's an objective right answer for these length questions, and I'm very skeptical of the idea that it's always better to condense and shorten. These events have roughly these lengths because that's what people expect, and I don't think it's safe to assume that they'd be better or more popular if they were cut down to a smaller size.

[Borrowing 100,000 Arrows]

I think Andrew makes a lot of really great points, but I'll just add two more. First, while it's true that ICT's order of finish mirrors the ACF Nats order of finish, upsets are much more common at ICT. It's basically unheard for a third bracket team to beat a top bracket team at Nats, but it's fairly common at ICT. The NAQT distribution obviously plays the biggest role here, but I'd argue that shorter questions do too. They increase the number of buzzer races, which makes occasional wild upsets more likely. I think this is fine at ICT, it adds to the unique feel of the tournament, but I wouldn't be stoked if both ICT and Nats were like this. Second, there are many categories like literature where condensing the question often makes the question worse. One of my main beefs with ICT is that it's often hard to adequately contextualize a scene in a novel, play, or short story in a five line tossup. This might just be me, but I'm much more likely to get a good buzz on something I've actually read at ACF Nats or CO because of the extra context.

[ErikC]

I think question length caps end up effecting the kind of clues you end up hearing/writing, not just the number of clues in a tossup. You end up with less adjectives, less explanations, and more 1:1 proper noun, capitalized word matching. There's nothing wrong with ICT having a lot of that, but clues that need some time to fully explain or the use of adjectives that help describe something might be too hard to include at a 6 line cap for harder difficulties.

7 lines for 3 dots or higher seems a little short to consistently produce tossups with good pyramids and avoid cliffs as well. Some kinds of questions need more time to get across the info they need to. I'm really skeptical that cutting important space from a Regionals tossup would really change the experience compared to the wait between games and the length of bonuses that typically aren't cut off like tossups are.

[jmarvin_]

I've already written at length in length's defense in other threads, so I will do us all a favor and refrain from retreading things in such detail here. I do want to put another name on the proverbial petition, though, endorsing 1) that length is not necessarily bad, 2) that writing shorter questions (even for a submission packet) is harder to do well than longer ones (and harder for the editors to improve, in the submission case, with less to sculpt), and 3) that I and many others find the "long," thoughtful, context-heavy style of quizbowl characteristic of Nats and CO to be the most fun version of the game, with unique benefits. Basically, Andrew has my thanks for already lucidly saying everything I would otherwise have taken the time to argue. Another thing which I haven't seen mentioned here is that the lack of buzzes on a lot of lead-ins is a misleading observation: oftentimes, one can reasonably extrapolate from the experiences of top level players, the rare strong specialists that will know these clues nevertheless wait to buzz because, without any powermarking and against a team that isn't threatening super strong buzzes against you on that category, it is strategically better to hear a bit more and confirm than to go for it. I don't think one can take the lack of lead-in buzzes on a given tossup to signal either that nobody knows the clue, or that nobody is getting anything out of hearing it (and Kevin was correct to point out that reading the packets afterwards is another source of value, one that people underrate, in my estimation).

On that note: if we're debating ways to make ACF questions more fun, my suggestion is powermarking (which my previous poll has shown to be supported, at least for some ACF sets, by a majority of quizbowlers surveyed)!

[Cheynem]

A lead-in is very rarely worthless--many times, they provide invaluable pieces of context, help to get players in the right time frame/location/etc., or at the very least, are entertaining and interesting. That's not to say that 10-line tossups featuring fascinating lead-ins are a great idea, or that there aren't tossups whose lead-ins are boring, uninteresting, and unhelpful. But as John and Andrew have pointed out, lead-ins, even if buzzpoints don't reflect this, can be very helpful.

I still don't like powers in ACF.

[benchapman]

While it's obviously good that longer questions can have more content for interested players to look up after the fact, I don't think we should prioritize that (relatively minor) benefit at the expense of the more tangible improvement to the gameplay experience itself from not-super-long questions. While I had a blast playing both CMST and BHSU and think they were both great sets, BHSU felt like less of a drag (even though I did worse) because of its strong length control vs. the excessively long 10/11-line tossups in CMST. The current ACF length limits (7 for Regionals, 8 for Nationals) work pretty well in my opinion as a player and we, at the very least, shouldn't make them longer.

[Santa Claus]

It seems that most of the posts have focused on the playing experience of shorter questions. While this is an important concern, plenty of ink has been spilled about this in the past and it always boils down to "it's subjective". I actually meant for this post to focus on the trade-offs for editors/writers - as I mentioned in some of my points, it seems like expectations on editors have been monotonically increasing over time and I imagine that contributes to long term attrition.

Since I'm pivoting the topic mid-thread, I'll do a quick recap of the editing-related parts of other posts for readers:

As a writer and editor, I do enjoy having more space to work. I personally find it harder and more work to craft something like a 6-line tossup than a really long one (which is why ARCHIVE will be on the longer side). You need to think about each clue a lot more.

Questions could probably be a little shorter. That said, I don't think eight lines (or seven, for matter) is an unreasonable length for a tossup. In any event, almost any eight-line tossup can be edited down to seven while keeping all of the clues and most of their substance, because most writers and editors are not very good at getting the most out of the space allotted. In that vein, I'm not sure that it's actually less work to have shorter questions; if you want those questions to be good, you'll be working pretty hard to pack in as much material as possible into seven lines, or eight lines, or whatever number of lines you have. I don't think that mechanically removing one of the first two clues in each tossup would be the best way to shorten any tournament's length.

It's also important to consider that some answers could only be tossed up with shorter questions since there wouldn't be enough good clues to fill out a nats length tossup, an issue I've run into in the past.

I recall editing DEES in 2014 and not having too much of a an issue making six line regular difficulty questions, though they weren't the most creative.

2) that writing shorter questions (even for a submission packet) is harder to do well than longer ones (and harder for the editors to improve, in the submission case, with less to sculpt)

For the purposes of this discussion (and because my phone lost my original draft), I'm going to take a slightly-more-extreme-than-my-real-opinion stance. I'll put this in a little box so it's clear that I'm being a little hot-headed:

All this posturing about shorter questions being harder to write is ridiculous. We're not talking about ACF adopting NAQT length constraints - it's one line out of eight (seven for Regionals), which is still several hundred more characters than ICT. In my mind, the simplest solution is to remove about a line's worth of clues, typically from the lead-in, and be slightly more concise. This immediately saves a ton of time: finding and editing early clues comprises a huge amount of the time I've spent on questions below nationals difficulty. The worst case of "rewriting questions to be shorter while keeping every clue" is not going to be necessary most of the time because so many lead-ins are unnecessarily hard. There aren't many concrete numbers, but what does exist make it clear that huge numbers are impossible; I can't really believe any argument that says "oh people are just waiting" because while waiting is solid strategy any early clue someone 100% knew would absolutely get buzzed on for clout. Clues which serve as context are obviously good, but they're second to clues that serve as actual clues - forcing editors to forage in the wilderness for novel clues so they can serve as vague ideas in a general direction seems like a poor trade-off.

[Borrowing 100,000 Arrows]

Kevin, I think that you're conflating a lot of things. First, I think you're conflating the time it takes to find fresh, buzzable clues with the time it takes to write longer questions. If you want to reward real knowledge and not just packet study, you're often going to have to spend a lot of time digging for fresh clues, period. It doesn't matter if the tossup is five lines or eight lines, it will take some time. [EDIT: As Jordan's comment reminds me, the difficulty of finding fresh early clues is literally why I and many other writers have Google docs filled with fresh, novel clues/ideas we've encountered.] Second, the fact that people don't buzz on a lead-in isn't necessarily evidence that they don't know the clue. Sometimes it takes a few seconds to process/remember the clue, or maybe you want an additional clue for confirmation. I can think of countless examples of both from this year's Nats. Especially at a tournament like ACF Nats, I'm trying my absolute hardest to not neg my categories, because the margin for error when you're playing top teams is really low. At a lowstakes tournament I might occasionally make risky first-line buzzes for clout as you say, but I'm not playing that way at Nats.

[naan/steak-holding toll]

Some context is due regarding my comment about editing DEES in 2014. Naturally, I was a less experienced editor at the time and also a player heavily focused on memorizing discrete facts and clue-connections; I was also primarily a history and beliefs player. These are categories that, like all others, benefit from the presence of context, but probably lose a bit less in their absence than other categories I've picked up some competence on over the years, especially literature and music. Describing historical events (without names), patterns, artifacts, etc. uniquely is, in my opinion, harder than doing the same for literature (I'd rate music about equally-hard, assuming equivalent baseline knowledge); on the other hand, omitting these descriptions also does not alter the play-experience of history as much as condensing literature to superficial plot descriptions, names, titles, etc. In music, requiring fewer clues makes some tossups possible, but makes many others impossible by making an acceptable pyramid with descriptions of non-nameable moments extremely challenging to produce.

I stand by my statement that shorter tossups are indeed possible at many college events; however, I'd strongly encourage writers who want to produce such events to critically think over how they're using their clues. This is especially relevant for questions aimed a stronger audience: I learned this the hard way with Missouri Open 2015, where many had structural playability challenges from rapidly switching from descriptions to name-drops, before I got a lot more experience by writing for NAQT and getting sent a ton of REWRITEs by more experienced editors.

---

As a random aside, the description of visual artworks deserves something of a comeback in quizbowl; none of us are clamoring to write corners-of-paintings bowl anymore and there's a real paucity of single-artwork tossups these days.

[theMoMA]

Kevin, I realize your position is meant to be rhetorically extreme, but I think it's based on mistaken assumptions about how people play and edit.

First, I just don't think you're correct about how players approach early clues at difficult events. Sure, it's fun and memorable when you're able to nail a question on the first or second clue, and it feels great to show off that knowledge and daring. I recall a CO game (annoyingly, a very consequential opening-round matchup) in which I happened to get roughly first-line buzzes on two very difficult answer lines, and we won a close game because I was dialed in and confident that my buzzes had very little chance of going awry. But I don't think I'm alone in saying that there are plenty of buzzes I could make with maybe 60 or 75% confidence that I just do not make because I'm playing to win the game and not to amass superficially good points per game or power numbers. Occasionally, of course, it's smart to make a very early buzz with an educated guess against a very strong player in a particular category. But generally it's not a winning play to take yourself and your teammates out of the rest of the question unless you're sure you're right.

Second, I just don't think that your understanding of tossup writing is realistic. You seem to be advancing a "bean dip" model for the construction of a tossup, in which the clues are layered, one on another, until they stand to the proper altitude above the answer line. You can certainly write a question that way, but it's probably going to be loose and sloppy, and while it might be "good enough," it's not going to be great. In my opinion, if you want to write a really good tossup, you should dredge up four or five really good clues, and then flip and chip and chisel and work each of them so that it fits perfectly with the others exactly into the space allotted (whether that's 500 characters or eight lines).

Finally, I just want to raise once again the continuing objection to this notion that a difficult first clue is unnecessary or impossible or whatever. It may seem strange, but please consider that some people like a difficult first clue, whether it's because it's an interesting piece of information that sets the table for the rest of the question or is otherwise just a pleasure to hear, or because it's nice to ease into a tossup and get a little context without having to worry that the first clue is going to be the site of contested buzzing in most cases. Or because not buzzing on 99 lead-ins makes buzzing on the 100th all the more memorable. There are a lot of reasons why quizbowl as it exists is enjoyable, and I don't think any of that stands or falls on the narrow criterion of whether each line of the tossup is generating a ton of buzzes.

[Cheynem]

I would also purport that if we made Nats tossups six lines instead of seven (or what have you) that most lead-ins would still generate pretty few buzzes.

[Jem Casey]

I whole-heartedly endorse Andrew’s posts in this thread and hope they will be remembered and quoted during all future iterations of this topic.

Anyway, I'm a little surprised by Angry Kevin's suggestion that scrounging around for leadins is a major time sink for writers. In my experience at the 3-dot level (which may not be representative, either), the first sentence or two is disproportionately likely to be quickest to write, because the prospect of getting its content into quizbowl is the reason I’m writing the tossup in the first place! I'd guess that, more so than any other part of the question, college qb leadins are likely to bear traces of the organic train of thought that led to the tossup, not a protracted, contrived post-hoc research project--at least, that’s the case for at least half of mine. Frankly, with a stricter line cap I’d probably try to fit those pet clues in anyway, and the gradient of the tossup would suffer somewhat as a result.

Granted, I've spent plenty of time struggling to find a clue or two to get some conceit to 7 lines, but the missing pieces are just as often medium or late clues--needs that wouldn’t go away if the top of the pyramid was lopped off. And if you do find yourself coming up empty on leadins that the field might plausibly buzz on, the answer you’ve picked may be too hard for the level you’re writing at; re-evaluating it, or just keeping the tossup on the shorter side, could be a good idea.

I’m also not sure what to make of Angry Kevin’s (a) insistence that buzz data can prove something like “a huge numbers [of leadins] are impossible” and (b) broad dismissal of others’ claims to derive some use or enjoyment from leadins they don’t buzz on. Sure, most sets have a smattering of ridiculous leadins, but many truly notable leadin clues don't get buzzed on, either. For a 7-line buzz curve, a first line should only be buzzed by something like 1 in 100 players, so it makes sense that some thoughtful, well-calibrated efforts to hit that small target come up short. And re: (b), the hunch or context I get from a leadin is a decisive factor in my later buzz (including in apparent “buzzer races”) multiple times at every tournament I play. Judging from some of the above posts, this doesn't seem to be particularly unusual.

That said, here are a couple points more in sympathy with the pro-”shorter questions” side:

• Personal tastes can change with community best practices--at least, mine do. As an underclassman, I enjoyed writing, reading, and playing jam-packed, lengthy questions and chafed at length constraints. But editors like the 2016 ACF Regionals team showed that writing great tossups under a 7 line cap was possible, and after working on sets with similar standards I came to appreciate controlled length as a tool for shaping better questions. More than just being practical, this appreciation became aesthetic, to the point that I now wince a bit when I read otherwise great old tossups that run to 9-10 lines for no clear reason. So if conventions change and caps become 1 line shorter, I might come to enjoy those standards as well. I’m just skeptical that the change would bring close to the same benefit to players and writers that the standardization of line caps did originally (notice that we haven’t had any 9 line tossups on Alvar Aalto at Regionals in a minute).
• Regardless of length, 3-dot questions can be easier. Too many tossups steer too far away from canonical material for the first several lines; more so than unconverted leadins, I’d be concerned about unconverted 2nd or even 3rd sentences, both of which do happen. As writers, we could stand to be stricter with ourselves when selecting clues; if you can’t make up a pretty compelling story about why multiple people in the field might know a clue, it’s probable that no one will. Likewise, there’s no shame in a couple extra people buzzing on your tossup’s second sentence, even if they’re getting there off (heavens!) packet-based studying.

[ThisIsMyUsername]

I'm having trouble following part of Andrew's argument, in as much as I feel that he has articulated a strong premise for the opposing side, while drawing the opposite conclusion. In his first post, he says:

In any event, almost any eight-line tossup can be edited down to seven while keeping all of the clues and most of their substance, because most writers and editors are not very good at getting the most out of the space allotted.

He then goes on to to defend several virtues of longer questions. But most of these, such as "trying to demonstrate an interesting conceit or set the table with clues that are more contextual than imminently buzzable," seem like things that should be found to equal degrees in the eight-line question and its hypothetical seven-line condensed version, if indeed the two have essentially the same content, as he suggests.

After that, both Andrew and Stephen argue for an additional virtue of longer questions, one that the eight-line version would possess that the seven-line version wouldn't: giving the player more time to think. But if I were following that logic (by treating that as a virtue unto itself), it seems like I (as a writer/editor) should never seek to streamline the wording of any tossup I write/edit, as long as it is under the tournament's length limit and the extra verbiage does not impede comprehension. It seems like deliberately inefficiently conveying the same information would, in fact, be good, because it would allow for more thinking time!

To be clear, I think their general argument in favor of eight-line questions is defensible (although I strongly disagree with it), if one rejects Andrew's initial premise--if one thinks that most eight-line tossups use their space super efficiently, such that any reduction in length requires compromising or jettisoning a clue. However, if one accepts Andrew initial premise, their argument requires making padding into an automatic virtue, which seems absurd.

My own position is that we should retire the eight-line limit at the upper level. Now that good quizbowl-specific character counters are available outside of NAQT, we should shift to having the limit be a character count. (I think all of us who have edited enough questions are familiar with rewording questions in dumb ways to get around where Microsoft Word places the line breaks, when trying to meet line limits.) This then leads to the question: what should that character count be at different levels?

If you had asked me during the tournament what BHSU's length limit was, I don't think I would have guessed that it was a soft cap of seven lines (as an editor has just confirmed for me that it was). The question pace didn't feel too fast at all. But looking through the first packet of BHSU and the first packet of this year's ACF Regionals, the average character count for tossups in the former appears to be shorter, around 800 characters or less. Likewise, it appears that the players didn't realize that 2022 Nats had a shorter character limit than usual (875 for 75% of questions), until I announced it after the tournament was over. These are anecdotal data points, of course, but what they suggest to me is that the aesthetic criteria that Andrew, Stephen, et al. are advocating for are (in the subjective perception of players) likely to be satisfied at a lower character limit than what we've been using (e.g. a 7- or 7.5-line CO tossup frequently won't feel too short to someone who doesn't know that it's only 7 or 7.5 lines). Indeed, unless you're hearing questions go to the end all the time, it's quite difficult to know how long the questions are (unless the difference is extreme, as in ACF vs NAQT lengths), since you don't know how many words follow your buzzpoint!

If a 3.5-dot tournament can use a high preponderance of <800-character tossups without compromising gameplay, 2- and 3-dot tournaments should be able to follow suit easily. And if a Nats can maintain a soft limit of 875 characters (one that I made soft only because ACF policy at the time forbade me changing the hard cap), there's no reason a Nats or a CO can't maintain a hard cap of 875 characters throughout, or even 850 characters.

[gyre and gimble]

After that, both Andrew and Stephen argue for an additional virtue of longer questions, one that the eight-line version would possess that the seven-line version wouldn't: giving the player more time to think. But if I were following that logic (by treating that as a virtue unto itself), it seems like I (as a writer/editor) should never seek to streamline the wording of any tossup I write/edit, as long as it is under the tournament's length limit and the extra verbiage does not impede comprehension. It seems like deliberately inefficiently conveying the same information would, in fact, be good, because it would allow for more thinking time!

To clarify my position, I didn't mean to advocate for merely increasing time as a physical variable. What I like about longer tossups is "room" to think, by virtue of more clues that can either give you a higher chance of recognizing something, or build context. I think this room is necessary to help people to buzz early more often on things about which they have deep knowledge.

[theMoMA]

I don't agree with basically any of the characterizations in John's post. Length limits are obviously arbitrary, and I don't think there's any inherent "virtue" (a word pulled from who knows where) to a longer question, or a question of any length, other than that some questions lengths generally feel about right for a certain difficulty level. I'm on the record above and in my editing about what I generally think are the line limits that feel about right, but I also acknowledged that it's arbitrary and there's wiggle room.

I don't understand why John is treating question length as a "virtue unto itself." There's nothing in this thread that would suggest anyone is taking that position. No one has advocated going wild and writing 15-line tossups. No one has said that questions should be long because longness itself is the desired characteristic. Rather, the thrust seems to be that question length should be in the right ballpark given the history and expected feel of a particular event, but it's not an exact measure; as I said, a seven-line CO would probably feel a little brisk to me, given my expectations for that CO, but "I'm sure you could write a CO that's a little shorter or a Fall that's a little longer and mostly hit on the right feel." The thing I like is not the length of the question, but the feeling of a satisfying difficult event that promotes the thinking side of the game.

I think it's a poor reading to suggest that anyone has advanced "long questions" as a sufficient condition for producing that effect. A tournament that engages the thinking side of the game emerges from a confluence of factors, which include (but are not limited to) questions that are long enough to allow some manner of contemplation without fear that every question is so imminently buzzable that you have to prepare for battle on each line. To me, those factors also include meaty and dense clues that are interesting, well characterized, and well written, in addition to a question length that feels right for the event.

I don't think I've played an event where I thought the clues were so jam-packed together that clue density was itself a problem, but maybe that would be possible at the very extreme end of the spectrum. I like hearing a clue I almost know and then thinking in two tracks (the clue in one, the ongoing clues in another) and trying to piece them together to get there. I don't like sitting through six lines worth of clues that sprawl over eight or hearing worthless clues because it allows "thinking time" or whatever. If John believes these are somehow positions in tension, I don't agree with that.

I don't think there's any merit to John's argument that it's impossible to defend writing an eight-line tossups if you accept the premise that most eight-line questions can be condensed to seven lines without much loss of information. When I write an eight-line question, it's not a temporarily embarrassed seven-line question, but rather eight lines that have been condensed and worked to fit the allotted space. I think most questions are written to a length instead of edited into that length (see my discussion of the "bean dip" model above), but it's plainly possible to produce tight (or loose) questions to any size. (For what it's worth, I'm strongly opposed to the position that the first line is worthless so you might as well throw it out to shorten the question, because I think you can and should pound the air bubbles out of your question before you do that; seven lines and no fizz is better than six lines of clues and a line of fizz.)

I also don't see the point of a character count. That works for NAQT because it's integrated into the central database for writing and editing and there are standard rules (handled automatically) for things like whether punctuation or pronunciation guides count against the limit. For any tournament using a setup like that, sure, use a character count, because it's going to be easier than anything else. But if you're forced to get out the measuring stick for each question, I don't think there's any real point to that when length limits are arbitrary and the difference of a half line here or there is minimal, as John says. If it's much easier to eyeball it with line limits, that's good enough.

[ThisIsMyUsername]

Andrew, I'm utterly baffled at how you derived basically most of that long post from what I said. Perhaps the rest of this discussion is best had in another medium?

I responded briefly and directly to two quotations from you, plus one summary that I would think is unobjectionable (that a benefit of eight-line questions is that they give players more time to think). Perhaps I have misinterpreted those quotes? (If so, your recent post doesn’t really say how.)

Unsurprisingly, your recent post quotes only my words “virtue” and “virtue unto itself” (and even then, interprets them rather eccentrically). If you actually quoted or engaged directly with what I said, I don't see how you could have arrived at the interpretations that you did. In particular,

No one has advocated going wild and writing 15-line tossups.

I think it's a poor reading to suggest that anyone has advanced "long questions" as a sufficient condition for producing that effect.

bear no resemblance to anything I wrote.

I interpreted you as saying the following three things:

(1) Almost any eight-line question could be condensed into a seven-line question while keeping all of their clues.
(2) The reason they tend to be eight lines rather than seven is “because most writers and editors are not very good at getting the most out of the space allotted.”
(3) Nonetheless, at certain difficulty levels, we should continue to write eight-line questions rather than seven-line questions.

This must mean that there is something about the eight-line version that is better than its seven-line counterpart. I used the word “virtue” in describing this. The Collins Dictionary says, “The virtue of something is an advantage or benefit that it has, especially in comparison with something else.” I’m not sure why you find this word objectionable (and call it "a word pulled from who knows where”). It seems like exactly the right word, assuming you are saying that eight-line questions have advantages over seven-line questions in certain circumstances. I don’t know how to read your post otherwise.

The numbered points also seem to entail that length is a unique vehicle for delivering this added value. I used the phrase “virtue unto itself.” The Collins Dictionary says, “If you say that something is, for example, a world unto itself or a place unto itself, you mean that it has special qualities that it does not share with other, similar things.” As far as I can tell, that is the only definition it provides for the phrase "unto itself." Judging from your response, you seemed to interpret this to mean “for its own sake.” I can’t find a dictionary that defines “unto itself” in this way, but perhaps I am unaware of some alternative meaning of the phrase.

The question I posed is what this virtue (or benefit, or whatever other synonym you like) is. I said the only plausible candidate I think you proposed is the longer question provides more time to think through the question. As I said, I think this is perfectly defensible if the room to think is created by adding another clue that has the potential to be buzzed on; but it is not when it’s created by a less compact wording of the same clues. In other words, I think (3) is tenable, but not if (1) and (2) are true.

I see that you address this directly in one paragraph of your recent post. You say:

I don't think there's any merit to John's argument that it's impossible to defend writing an eight-line tossups if you accept the premise that most eight-line questions can be condensed to seven lines without much loss of information. When I write an eight-line question, it's not a temporarily embarrassed seven-line question, but rather eight lines that have been condensed and worked to fit the allotted space. I think most questions are written to a length instead of edited into that length (see my discussion of the "bean dip" model above) [...]

This suggests that eight-line questions are optimized to fit the eight lines. This seems to me to directly contradict (2), which (as I stated) I treated as one of your premises.

I also don't see the point of a character count. That works for NAQT because it's integrated into the central database for writing and editing and there are standard rules (handled automatically) for things like whether punctuation or pronunciation guides count against the limit. For any tournament using a setup like that, sure, use a character count, because it's going to be easier than anything else. But if you're forced to get out the measuring stick for each question, I don't think there's any real point to that when length limits are arbitrary and the difference of a half line here or there is minimal, as John says. If it's much easier to eyeball it with line limits, that's good enough.

I explicitly offered one reason already: If you use line counts, you get annoying situations where you have to move a long word to a different place in the tossup, because it’s throwing off where the line breaks are. I at least implied another reason: that the ideal number of words to maximize efficiency while still preserving the aesthetic qualities you and others seek may be between seven and eight lines, and therefore can only be captured by character count. I’ll add a third. If you use line count, then you can’t add pronunciation guides before question text is finalized. If you do, you have to remove the guides, check the lines, and then add them back in. With a character counter trained to ignore pronunciation guides, you don’t need to do this.

[Santa Claus]

This thread ended up going off the rails (several times) so I’ll just make the overarching post I should have made when I first posted.

Expectations on writers and editors have increased more or less monotonically over time (at least at the difficulties we’re talking about here). There’s variation here and there, but the quality that players expect from a modern project is so much higher than it was a decade ago (maybe even on shorter scales too). The development and subsequent popularization of databases means that reverse clue lookup has gone from “convenient” to “obligatory”. The increased availability of online resources means that questions can (and are often expected to) be meticulously researched enough to pass scrutiny from subject-matter experts in the field. This has been mitigated somewhat by the fact that editors are rarely expected to edit more than one 4/4 category, if even that. On a per-question basis, though, the point remains.

In my eyes there are a couple routes that quiz bowl writing/editing can go in the long-term:

• there is sufficient growth and interest from successive generations that no special effort is required to maintain existing projects and/or expand
• at some point it reaches a tipping point where there is not enough new blood active to replace retiring editors, forcing quiz bowl to contract or make special effort to correct this
• things go bad fast and the game implodes due to no writers

The third is certainly too pessimistic, but the first strikes me as a little too optimistic - it seems like the second is a real possibility (these are obviously reductive and vague predictions but this is just a rhetorical device anyways). I’d be surprised if the game died in ten years, let alone five, but it seems foolhardy to never make any changes on the assumption that things will turn out fine.

If we consider the hypothetical goal of “decreasing writer load”, what are the options available?

1. lower expectations of quality
2. increase number of writers/editors
3. reduce the number or length of questions

I don’t think option 1 is very likely (and it’s not like anyone could convince the community to put the toothpaste back in the tube anyways). Option 2 is already going on (though there are various practical limits to how far it can go) so I don’t think it requires any special consideration. Of option 3, I’ve never heard of anyone suggesting a less-than-20/20 format, or lowering the number of games played in an average tournament; both strike me as obviously bad anyways. That leaves question length as a logical place for reform (and hence this thread).

My main goal was to have people come up with their best arguments for whether shorter questions would be practical so that someone who actually had influence (like the board of ACF) could do their own deliberating if/when the issue came to a head, with this thread as a starting point. Previous posts advocating for shorter questions have focused on tangible things like the impact on game speed; my focus this time is editor welfare.

I feel like I’ve already summarized the relevant discussion in my previous posts, so I’ll leave it here.

sidebar 1

I earlier suggested that the easiest way to shorten questions would be to remove lead-ins, as those take a lot of time to write. Some have suggested that writers should have lists of clues ready for use, or that lead-ins should be things which one specifically wants to include. While I agree that these are important sources of clues, I find it hard to believe that these would comprise a significant part of any writing project. Maybe others have had different experiences than I have, but I find that after the half-completion mark I’ve run out of non-overlapping lead-in ideas and I again have to return to the clue mines. A lot of the time I have to start a tossup from an answerline or vague conceit to ensure that subdistros are met - there’s no guarantee I’ll have something set aside in that scenario.
Edit: rereading the thread I see much of this was already mentioned - that’s what I get for writing based on memory.

Besides, I don’t think I was advocating for only cutting leadins or something - they just represent a straightforwardly cuttable set of clues. If one’s already picked out you could just as easily cut another clue and still see a benefit.

sidebar 2

I feel like we all got collectively stuck on how enjoyable long questions are to play or write. I’m pretty sure we’re all on the same chapter, if not the same page there - after all, most of us are playing CO this year (where I will also be reading a bunch of long, self-indulgent questions) - but these arguments over subjective things alway devolve into “I like it” vs. “I don’t like it”.

[kievanrustic]

In the abstract and if the goal is to decrease writing time, I'd imagine that looser length requirements are faster to write than strict limits.

I'd echo the previous point that leadins and giveaways are likely the fastest parts to write as the former gives the writer the most leeway with question space while the latter is a natural description of the answerline. It's the middle with concerns of a smooth difficulty gradient and the place of the clues in the context of previously written quiz bowl that, at least for me, takes the longest to research and write.

[Sam]

I'm curious where the range of length fits into this. From casual empiricism, if the cap is eight lines, there are unlikely to be many that are five lines, even though that's not a crazy number in an absolute sense. Players and writers have come to expect now that there will be variation in difficultly within a tournament, hopefully evenly spread across categories, would it be a problem if the same were true of length?

I guess this is a question of magnitude: there already is variation in length, not every tossup is exactly the same number of characters. And it's not as if variation in difficulty is a carte blanche for including IS-set level questions in Chicago Open. But I wonder how allowing for more length variation within a set would play out.

[Adventure Temple Trail]

I'm curious where the range of length fits into this. From casual empiricism, if the cap is eight lines, there are unlikely to be many that are five lines, even though that's not a crazy number in an absolute sense. Players and writers have come to expect now that there will be variation in difficultly within a tournament, hopefully evenly spread across categories, would it be a problem if the same were true of length?

I guess this is a question of magnitude: there already is variation in length, not every tossup is exactly the same number of characters. And it's not as if variation in difficulty is a carte blanche for including IS-set level questions in Chicago Open. But I wonder how allowing for more length variation within a set would play out.

At least in my own writing/editing experiences, I've felt a sort of Parkinson's law analogue: question content expands to fit the cap allotted. The usual temptation is to share as many cool, useful clues as possible, and the number of such clues is usually* greater than the space available for them.

As such, I doubt that "allowing more length variation" would actually result in much length variation; if more length variation is an outcome we want, it's wisest to try imposing some top-level distribution of lengths, with a program or script keeping track/enforcing it for you.

*I do think some amount of length variation is worth pursuing, precisely because this is usually true but not always true. As others have noted before, there are some tossups that are workable at ICT but not Nats, because there are enough clues to fill five lines but not eight.

[The King's Flight to the Scots]

I feel like that much length variation would be confusing to players. The difficulty structure of a 5-line tossup is going to be very different from the structure of an 8-line tossup; it'll be jarring to play the 5-line tossup and hear a clue in line 2 that you'd expect to hear in line 4 or 5. Generally if the cap is 7 lines, I think the large majority of questions should be at least a full 6?

[Sam]

*I do think some amount of length variation is worth pursuing, precisely because this is usually true but not always true. As others have noted before, there are some tossups that are workable at ICT but not Nats, because there are enough clues to fill five lines but not eight.

Yes, I think this is correct: not variation for its own sake, but when it could expand the set of askable answer lines. Within that category there are two types of answers I can think of: 1) those where the answer line is specific enough that there are legitimately just not that many facts about it in the world (the only things that are coming to mind are humorously bad common links, but there are probably some good ones as well), and 2) those where it's just very, very hard, the kinds of tossups where everything is in power.

The scenario I have in mind is not distributions saying "you need 20% to be 5 lines, 30% 6 lines, etc" but say an editor gets a question or has an idea for a question that's interesting but in one of the categories above. There are different strategies: if it's in category #1, the editor could make it a bonus instead. If it's #2, they could be to cut it entirely, or change the answer line but still use a lot of the same clues (excerpts from Nicanor Parra's poetry go in a "Chile" lead-in now), or leave some of them in as they are and be okay with the top 1, 2, 5% of tossups stretching difficulty. But maybe they could also just make it five or six lines, if they think that's what it will bear.

I feel like that much length variation would be confusing to players. The difficulty structure of a 5-line tossup is going to be very different from the structure of an 8-line tossup; it'll be jarring to play the 5-line tossup and hear a clue in line 2 that you'd expect to hear in line 4 or 5. Generally if the cap is 7 lines, I think the large majority of questions should be at least a full 6?

I agree with this, a majority of questions should still be pretty close to the cap for the reasons Matt Jackson outlined, it's fun and educational to include more stuff if you can.