What makes a good tiebreaker?

[bradleykirksey]

I didn't know where to put this, so I guess I'll put it here. I was a little surprised that there isn't like a "Writing for people who don't write no good" or "writing theory" subforum but that's life. Can't have an infinite number I guess. I've worked on a few sets over some period of time and I've ended up packetizing a lot of them. It's a lonely life, but it allows me to be snide every time I hear two lit tossups in a row. But it means that I've thought more about tiebreakers in quiz bowl matches than is ever healthy for a human being, or even most marsupials.

Honestly, every set ever made has had some tossups that were better, worse, easier, and harder than others. You don't want to put a great tossup there in TBs, because then only two teams will hear it, probably. But you don't want to put a bad tossup there, because then it's like "the pack wins" sudden death no matter who gets it right. If it's too easy, you risk a buzzer race determining an entire match. If you pick the hard tossups for TBs, it means you risk burning through multiple questions by bottom bracket teams on a TB and you need to write more.

I think ideally, a tiebreaker should have

1) As many buzzpoints as possible. Tossups aren't evenly buzz-in-able on every word, and there are some points where you can expect a buzz. And I guess you'd want as many as possible on a TB, because you really want to differentiate good teams here.

2) Of a sufficient quality. I've heard more than one writer suggest sliding down a bad TU to tiebreakers, rather than go through awkwardness of outright rejecting one. But it should probably be a little closer (70/30??) to the weaker end because you don't want to bury a gem (which can be harder to predict than I initially expected, but maybe just because I'm bad at editing) but also, you don't want a really weak tossup determining a whole tournament.

3) It should shy away from canon expanding questions. But also, if possible, stock answers and clues. But it in that order. Dropping TB questions is depressing, it's worse for writers, it's bad in general. But also it means that no one is going to get to see this cool, awesome thing you found that you want included in the canon, if you include a canon-expanding question down there. Conversely, quiz bowl matches shouldn't be won on stock-clues-for-The-Great-Gatsby-knowing, but that's both more permissible and easier to avoid, in my mind.

I think that writing that out as helped me think it through, but I'd really love to know what actual competent people think. What do you guys think?

Or am I overthinking something an RNG could do better?

[Skepticism and Animal Feed]

I think the single worst thing you could do is use a tiebreaker that is impossible. If it doesn't get answered, then, congratulations: you've just delayed the tournament! I have always favored accessible answerlines for tiebreakers.

[Tejas]

I think a lot of this just boils to “write good tossups” and the only difference from the rest of the set is your third point, that you don’t want to make them too hard to convert. I understand why people would put tossups they think are worse as some of the later tiebreakers, but doing so means they’re not producing the best possible set.

[Ike]

All of Bradley's points seem reasonable to me. Historically, writers often just dumped their worst questions into tiebreakers since they don't have enough time to polish every aspect of their set. I'd argue that rather than putting an extra tossup into each packet, (thus creating something like 15 tossups worth of tiebreakers,) it's best to set aside 5 additional tossups in an additional mini-packet. That way, you can minimize the work done to ensure that they are not your worst questions.

If you're in a packet submission tournament, and you want to create such a tiebreaker document. You can also leave a note saying which school wrote the packet from which the tossup was taken. That's for the case that you have a tiebreaker match involving said school, then you can just read one of the other tossups from the tiebreaker packet.

[Stained Diviner]

Some of this depends on how many extra tossups you put in each packet. If there is one, then Bruce is correct, and you actually should put that Gatsby tossup there. If there is more than one, then you can put the tossups in pyramidal order (which isn't an actual concept, but hopefully that makes sense), and you can start out with a low conversion tossup. That does not excuse impossible tossups, which do not have a place.

Low-scoring matches are a little more likely to be tied than high-scoring matches, so consider the possibility that there will be a tie between two teams that are at the bottom of the field. That doesn't mean you should start off with a stock clue, but it does mean you want somebody buzzing in by the time the question is over.

[CPiGuy]

I'd argue that rather than putting an extra tossup into each packet, (thus creating something like 15 tossups worth of tiebreakers,) it's best to set aside 5 additional tossups in an additional mini-packet.

I strongly agree with this. Writing a tiebreaker for each round is pretty wasteful because tie games are pretty rare in quizbowl. Having a mini tiebreakers packet also makes it very easy to compensate if a tiebreaker goes dead -- just read the next one!

[ThisIsMyUsername]

Like others, I heartily agree with Bradley's three criteria for a good tiebreaker, as stated in his original post. (Although, I should note that his first two criteria should be true of any question that makes its way into a set!)

If you're in the position in which you're carefully considering which tossup to use as a tiebreaker, I would say that something has already gone wrong. If your tournament is a housewrite, you should have designed the subdistribution of the packets to be self-standing, and chosen your tiebreaker answerlines separately. (To be clear, I'm not arguing against anyone in this thread. I am just pointing out a flaw in some conventional characterizations of this process.) Likewise, I suggest that editors of packet-submission tournaments write the tiebreakers themselves, rather than using submitted questions.

Another important point: While Bruce is right that tiebreakers should not be harder than the median difficulty for the tournament, I'll diverge by saying that it's not necessarily a bad idea if they are capable of going dead or being answered by only one of the two teams. (That is, it is not necessary to write a "hard- / medium-level tossup on an ACF Fall answerline," unless that is the general tenor of your set.) This is because these tossups are used not only when the final score is tied but also as a means of resolving protests. (Indeed, I think that is their more common use.) This can involve the tossup being read to only one team; or a situation where Team 1 doesn't need to convert the tossup themselves, so long as Team 2 fails to convert it. While reading a tossup to only one team undoubtedly confers a huge advantage upon them (an advantage that many object to), it should not guarantee that they will win.