Some Thoughts on Bonuses
Posted: Mon Jan 17, 2022 8:33 pm
After tournaments, more often than not, I end up talking with people about bonuses: what functions they serve, what knowledge they test, how they should be written. In this post I'll share what I see as bonus best practices and gesture a bit towards their future. All of my examples will be hypothetical (I'm not a fan of the j'accuse! school of forum posting).
I'll begin by stating my core postulate: quizbowl is first and foremost a game. The rest of my post will assume that the primary function of bonuses is to differentiate between teams competively, at three different levels of skill. There are other nice things about bonuses--leadins are usually where writers stash interesting facts or context that they can't fit elsewhere. But these considerations are secondary to making gradated, answerable questions.
Bonus intricacy, moderating, and cognitive load
In an effort to write more creative or educational questions, writers try structure their questions around unique themes and include as many interesting clues on the subject as possible. The problem with this focus is that these questions often take up much more space. They tend to balloon or employ a succession of intricate clauses to thematically link bonus parts. Sometimes, in an effort to make an easy part "interesting," you end up with questions like:
1) Moderators have finite amounts of energy. Tournaments with systematically long and complex bonuses drain readers, who begin to make more mistakes. They lose their voices, mispronounce words, and mis-emphasize things. That's not fun for players either! Not coincidentally, NAQT sets are the most fun to read.
2) Intricate bonuses increase cognitive load, thereby making questions unecessarily hard to answer. We've all seen the good team that sits through 3 lines of "After...that...which..., while..." only to go "what are they asking for?". Quizbowl is a difficult game that requires a lot of sustained thinking over a long period of time. Part of being an empathetic writer is imaging oneself as a tired player, listening and trying to keep track of all the clues being thrown at them (this is where reading your own questions out loud helps). Note relatively short questions can also unduly increase cognitive load by trying to fit too many clauses or clues into a short amount of space.
3) Intricate bonuses disproportionately punish newer players. And here, I mean the players who haven't yet become skilled at cutting through the quizbowlese, sifting through extraneous information to realize the writer is just asking for the guy who wrote "The Raven." Most easy tournaments do a pretty good job at this, but there is room for improvement.
I am of the opinion that there is rarely reason to go over two lines for a bonus part, and that it is usually possible to fit interesting, gettable clues within 1.5 lines---with exceptions, like scientific formulas and score clues. Someday I'd like to see a tournament try a hard cap of 2 lines with at least one part of each bonus required to stay on one line. My guess is that sort of limitation would improve quizbowl writing by forcing tighter clues.
What are easy parts for?
It is a truth universally acknowledged that easy parts should legitimately test knowledge. Easy parts are there to differentiate between the bottom of the field. Compare two possible easy parts for a bonus making a connection between the Latvian novel The Electioneer's Daughter and Anna Karenina:
Most teams probably know Anna Karenina threw herself in front of a train. A related type of bonus asks you to make the obvious guess without concretely knowing any of the clues:
This example might seem absurd, but I swear I've heard a similar question in just about every recent academic tournament. Usually it happens when an author has a niche theme they don't want to break (Latvian literature, prehistoric Japanese artifacts, Romanian mythology, etc.). In these circumstances, I would suggest writers choose the lesser of two evils and find a way to make the easy part on Tolstoy. The answer can even still be an object, event, substance, etc.---vary your answerline types, of course. But make sure you're 1) testing knowledge of the actual subject 2) giving a concrete clue you expect most of the field to know.
I'd like to emphasize that I don't think these sorts of easy parts make the game more interesting for the bottom of the field---they're not the ones burnt out on Tolstoy questions. I think these questions are written out of the fear that very good quizbowlers who have heard it all before will criticize the bonus for being boring. Newer players are usually just happy to get points for knowing something.
Most of quizbowl has recognized that it is better to write a tossup on a simple answerline with creative clues than a tossup with a creative but confusing answerline. I'm not sure we've paid similar attention to bonuses. We should be clear that there is nothing wrong with a simple easy part. In some ways, it takes a lot of confidence to write "Anna Karenina is by this author of War and Peace." Don't let the need to do something new get in the way of writing a bonus that distinguishes between bottom bracket teams.
Bonuses and the End of Quizstory
I've talked a bit about how bonuses are currently written, and I've probably said a lot of things you already know about writing simple, direct questions. But any bonus discussion eventually runs into the big problem: why does a mostly redundant knowledge test take up more than half of game time? By some accounts, bonuses decide less than 10 percent of games, and are about 1/12 as predictive of ICT finishes as tossups (source: some threads I read once but can't find).
Arguments in favor of the current three-part bonus usually focus on the teamwork aspect, providing opportunities for weaker players to contribute, testing knowledge that doesn't fit well into pyramidal tossups, and quizbowl's uniquely educational ethos (i.e., bonuses give more space for writers to provide important context and fun facts). These are all good reasons for keeping bonuses as they are, and I don't think the state of the game today is exactly dire. But given how much time we put into bonuses versus their importance for deciding games, I think it's worth considering ways to keep their positive aspects while improving game efficiency.
We seem to have reached the End of Quizstory, where we've become focused on producing the best possible version of a fixed, immutable game. There's no real reason why this needs to be the case. I would personally enjoy playing academic housewrites that take big risks with the format (successful distributional experiments like FLOPEN have already taken some steps in this direction). Here are some barely thought-out mechanical adjustments:
Many very good writers and editors have made these most of points elsewhere, and I don't think I've put down any unique insights here. I would love to see further discussion on the broader purpose of quizbowl bonuses, and how we can be writing them better.
I'll begin by stating my core postulate: quizbowl is first and foremost a game. The rest of my post will assume that the primary function of bonuses is to differentiate between teams competively, at three different levels of skill. There are other nice things about bonuses--leadins are usually where writers stash interesting facts or context that they can't fit elsewhere. But these considerations are secondary to making gradated, answerable questions.
Bonus intricacy, moderating, and cognitive load
In an effort to write more creative or educational questions, writers try structure their questions around unique themes and include as many interesting clues on the subject as possible. The problem with this focus is that these questions often take up much more space. They tend to balloon or employ a succession of intricate clauses to thematically link bonus parts. Sometimes, in an effort to make an easy part "interesting," you end up with questions like:
This sort of long exercise in capital-matching is common enough, and usually elicits a groan from someone in the room. It's fine to keep players interested by mixing in some extra clues, or to give them something to look up later. But long, intricate bonuses create competitive problems:This modern-day country traces its founding to the ancient king Dudley Princesymbol, who used flaming weasels to repel the Numenorians from his ancient capital of Labradoria, therebye fulfilling Jim Bob Beaverton's prophecy that "fire will be stronger than water." For 10 points each:
[10e] Name this country, where, in the 1950s, Dudley Princesymbol's legacy divided the Borgesians and the Quixotians at French-backed universities overseen by Leopold Churchill. Dirk Lastrade's Pulizter-winning history Restless City recounts how Dudleyite paramilitaries used pitchforks and torches to clash with the Borgesian-aligned LPLA in this country's capital of Coolesville.
1) Moderators have finite amounts of energy. Tournaments with systematically long and complex bonuses drain readers, who begin to make more mistakes. They lose their voices, mispronounce words, and mis-emphasize things. That's not fun for players either! Not coincidentally, NAQT sets are the most fun to read.
2) Intricate bonuses increase cognitive load, thereby making questions unecessarily hard to answer. We've all seen the good team that sits through 3 lines of "After...that...which..., while..." only to go "what are they asking for?". Quizbowl is a difficult game that requires a lot of sustained thinking over a long period of time. Part of being an empathetic writer is imaging oneself as a tired player, listening and trying to keep track of all the clues being thrown at them (this is where reading your own questions out loud helps). Note relatively short questions can also unduly increase cognitive load by trying to fit too many clauses or clues into a short amount of space.
3) Intricate bonuses disproportionately punish newer players. And here, I mean the players who haven't yet become skilled at cutting through the quizbowlese, sifting through extraneous information to realize the writer is just asking for the guy who wrote "The Raven." Most easy tournaments do a pretty good job at this, but there is room for improvement.
I am of the opinion that there is rarely reason to go over two lines for a bonus part, and that it is usually possible to fit interesting, gettable clues within 1.5 lines---with exceptions, like scientific formulas and score clues. Someday I'd like to see a tournament try a hard cap of 2 lines with at least one part of each bonus required to stay on one line. My guess is that sort of limitation would improve quizbowl writing by forcing tighter clues.
What are easy parts for?
It is a truth universally acknowledged that easy parts should legitimately test knowledge. Easy parts are there to differentiate between the bottom of the field. Compare two possible easy parts for a bonus making a connection between the Latvian novel The Electioneer's Daughter and Anna Karenina:
[10e] In The Electioneer's Daughter, Sylvia throws herself in front of a train, much like this author's character Anna Karenina.
ANSWER: Leo Tolstoy
The first style of bonus is "boring," but it accurately sorts bottom bracket teams based on whether they know who Leo Tolstoy is. The second style of bonus asks you to know what vehicles run on tracks, and then throws in a non-literature clue. You probably haven't and won't see this type of bonus for Anna Karenina specifically, but you've definitely seen it for other topics and subjects.[10e] Both Sylvia in The Electioneer's Daughter and Tolstoy's Anna Karenina throw themselves on the tracks of this sort of vehicle, which in Russia run on a "Trans-Siberian" structure.
ANSWER: a train
Most teams probably know Anna Karenina threw herself in front of a train. A related type of bonus asks you to make the obvious guess without concretely knowing any of the clues:
Coffee of course is the obvious guess here, but I don't actually know anything about Fire on the Neva. For all I know, Otto is a Satanist who drinks a fresh cup of goat's blood every morning. What this bonus asks the player to do is successfully recognize that it's the easy part, and not out-think their way past the obvious guess. Instead of differentiating literature knowledge, this bonus differentiates between teams who can guess what someone might drink in the morning and teams who go wrong somewhere along the way ("uh, tea has caffeine I guess").[10e] The scarcity of the plant used to make this substance is a recurring motif throughout The Electioneer's Daughter. In Lando Daugavins' novel Fire on the Neva, Otto drinks this substance to wake up in the morning.
ANSWER: coffee
This example might seem absurd, but I swear I've heard a similar question in just about every recent academic tournament. Usually it happens when an author has a niche theme they don't want to break (Latvian literature, prehistoric Japanese artifacts, Romanian mythology, etc.). In these circumstances, I would suggest writers choose the lesser of two evils and find a way to make the easy part on Tolstoy. The answer can even still be an object, event, substance, etc.---vary your answerline types, of course. But make sure you're 1) testing knowledge of the actual subject 2) giving a concrete clue you expect most of the field to know.
I'd like to emphasize that I don't think these sorts of easy parts make the game more interesting for the bottom of the field---they're not the ones burnt out on Tolstoy questions. I think these questions are written out of the fear that very good quizbowlers who have heard it all before will criticize the bonus for being boring. Newer players are usually just happy to get points for knowing something.
Most of quizbowl has recognized that it is better to write a tossup on a simple answerline with creative clues than a tossup with a creative but confusing answerline. I'm not sure we've paid similar attention to bonuses. We should be clear that there is nothing wrong with a simple easy part. In some ways, it takes a lot of confidence to write "Anna Karenina is by this author of War and Peace." Don't let the need to do something new get in the way of writing a bonus that distinguishes between bottom bracket teams.
Bonuses and the End of Quizstory
I've talked a bit about how bonuses are currently written, and I've probably said a lot of things you already know about writing simple, direct questions. But any bonus discussion eventually runs into the big problem: why does a mostly redundant knowledge test take up more than half of game time? By some accounts, bonuses decide less than 10 percent of games, and are about 1/12 as predictive of ICT finishes as tossups (source: some threads I read once but can't find).
Arguments in favor of the current three-part bonus usually focus on the teamwork aspect, providing opportunities for weaker players to contribute, testing knowledge that doesn't fit well into pyramidal tossups, and quizbowl's uniquely educational ethos (i.e., bonuses give more space for writers to provide important context and fun facts). These are all good reasons for keeping bonuses as they are, and I don't think the state of the game today is exactly dire. But given how much time we put into bonuses versus their importance for deciding games, I think it's worth considering ways to keep their positive aspects while improving game efficiency.
We seem to have reached the End of Quizstory, where we've become focused on producing the best possible version of a fixed, immutable game. There's no real reason why this needs to be the case. I would personally enjoy playing academic housewrites that take big risks with the format (successful distributional experiments like FLOPEN have already taken some steps in this direction). Here are some barely thought-out mechanical adjustments:
- Try out one-part and two-part bonuses. Even combine this with an increase in tossups, and more distributional space for "Other" questions.
- Allow conferring after the other team negs, while cutting back on bonuses. For many players--like Reach for the Top vets--this is actually very intuitive. You get more of the pace of tossups, while keeping the inclusive aspects of bonuses.
- Play around with point values. Maybe tossups should be worth zero points: they'd be plenty valuable anyway. Introduce two-part bonuses worth 20 points each.
Many very good writers and editors have made these most of points elsewhere, and I don't think I've put down any unique insights here. I would love to see further discussion on the broader purpose of quizbowl bonuses, and how we can be writing them better.