Tiebreaker Distro

This forum is for discussing tournament formats, question styles, strategy, and such.

Tiebreaker Distro

Postby More Weight » Sun Jul 23, 2017 4:28 am

After losing two (2, count em boys) games in 5 minutes thanks to tiebreakers, I've been thinking about alternative distributions for tiebreakers specifically. Fast glossary. Everyone understands that some topics are "fraudable," for example if a question is hesitant to mention gender and some aviation clues come up (Amelia Earhart), or if some Alice in Wonderland clues come up along with "this poem"(Jabberwocky). It's also common knowledge that stock clues are ways that students with strong memory skills but no desire to learn can excel at the game. That being said, what if tiebreaker questions were only current events questions?

There exist no stock clues for current events because ideally they have happened fairly recently, and they're somewhat difficult to guess without knowing the perspective of the people in the clues. And that's a good thing to encourage, right? For players to actually stay up to date with major political figures and world happenings, as well as the impacts of such, as opposed to just grinding on Quizlet for days at a time?

Alternatively, having questions of any subject specially made to not include any stock-ish material? "Real knowledge" has become the opposite of "stock clue," and involves more holistic learning. If you power a biological research process because you happened to have worked on a project and had to personally perform it, good for you. That is about 30x more respectable than hearing "Allred-Rochow scale" and collecting 15 points, or hearing chemistry words and the words "this state" and buzzing equilibrium.

This might not matter because tiebreakers are fairly rare happenings, but I wanted to talk about alternative ways of determining which team is truly more knowledgeable in the case of a tie. And though I think 2 of 3 rock-paper-scissors is truly the best tiebreaker method, I want your thoughts on this.
Andy Kong
Hoover High School
Hoover, Alabama
More Weight
Kimahri
 
Posts: 4
Joined: Tue Jan 13, 2015 9:49 pm

Re: Tiebreaker Distro

Postby 2017 in amusement parks » Sun Jul 23, 2017 1:13 pm

Why should the tiebreaker be any different from the rest of the match? Every tossup in an ultimately tied match matters, so why should we nitpick tossup 21?
Michael Borecki
Middlesex Middle 13,
Darien (co-captain) 17,
Bowdoin College 21
2017 in amusement parks
Rikku
 
Posts: 329
Joined: Thu Nov 15, 2012 6:45 pm

Re: Tiebreaker Distro

Postby Theodore » Sun Jul 23, 2017 1:18 pm

A lot of what you said applies to all questions in general, not just tiebreakers. Stock clues and transparent questions should be avoided in all questions. If the tossups you mentioned appeared in the regulation first 20 TUs, they would still be considered poorly-written questions, and would cost you the same amount of points, In fact, they would cost you more as TB TUs don't have bonuses, but you wouldn't see them as game-deciding due to a psychological "recency bias" and/or the intense mindset you're in when a TB is played.

This brings up one of the arguments for having 1 TB: teams have already heard 20 TUs to differentiate themselves. In addition, 1 TB per pack (or fewer TBs overall in the TB TUs document) saves time for readers and writers.

However, 3 TBs allows for 1 literature, 1 history, and 1 science TB each. For this reason I prefer this (as somebody who has won and lost numerous games on tiebreakers). That being said, I understand the strain on writers, so I don't mind when there is only 1 TB. You shouldn't view it as the TB costing you the game, as you could've beaten the opposition to any of the TUs they got in the previous 20.

Whether 3 TBs should be 1 lit, 1 history, and 1 science is a separate debate. It often becomes a slippery slope questioning why the Quizbowl distribution is the way it is -- not to say that isn't worth questioning from time to time. The reason why all TBs aren't current events are likely the same reasons why lit, history, and science each make up a much larger proportion of the QB distribution than current events in the first 20 tossups.

From my experience (disclaimer: this is from a couple years ago, and is not guaranteed to be true anymore), if the writers write all their tiebreaker tossups in a separate document, they will either write all lit, then all history, then all science in that order, or will cycle through 1 lit, 1 history, 1 science, and so on. This usually means that unless the TD randomly chooses one (but often they just choose the first one for simplicity), if you know you are the first game of a tournament to hear a tiebreaker tossup, then it's probably a literature tossup. Writers should keep this in mind and randomize the order of their tiebreaker tossups, whether dispersed (1 at the end of each pack) or compiled into one document.
Ted (Zefeng) Gan

Carleton University, 2019
Colonel By Secondary School, 2015

Former Executive Member of the Ontario Quizbowl Association (ONQBA), 2015-16
Former NAQT Writer, 2013-2016
User avatar
Theodore
Lulu
 
Posts: 99
Joined: Thu Mar 21, 2013 5:40 pm
Location: Ottawa, ON, Canada

Re: Tiebreaker Distro

Postby More Weight » Sun Jul 23, 2017 4:40 pm

2017 in amusement parks wrote:Why should the tiebreaker be any different from the rest of the match? Every tossup in an ultimately tied match matters, so why should we nitpick tossup 21?

You're right, I was viewing the final TU as a culmination of the game instead of as just another question.

Theodore wrote:However, 3 TBs allows for 1 literature, 1 history, and 1 science TB each. For this reason I prefer this (as somebody who has won and lost numerous games on tiebreakers). That being said, I understand the strain on writers, so I don't mind when there is only 1 TB.

I do like this idea though. It seems fairer to teams with both breadth and depth, but there is the possibility of tying the game again. If one team gets two 10s and a neg, and the other team powers the last one, the game is no more resolved. Associated bonuses would just increase the outcomes and not necessarily the fairness of the TB. Simplicity seems the best method of resolution.
Andy Kong
Hoover High School
Hoover, Alabama
More Weight
Kimahri
 
Posts: 4
Joined: Tue Jan 13, 2015 9:49 pm

Re: Tiebreaker Distro

Postby benmillerbenmiller » Sun Jul 23, 2017 6:28 pm

I think if your goal is to avoid fraudable or otherwise cheap TB questions, then making them all current events is probably not the way to go. With the exception of trash, current events is the least "academic" aspect of the distribution and one of the easiest categories to power without any great knowledge of the topic. I get the frustration with making all TBs history/science/lit, especially for those who specialize in other subjects, but I think it is relatively fair to arbitrate ties using the categories which constitute the heart of the game. A possible solution to the fraudibility problem is to make TBs more creative, perhaps through common links or diverse clue selection, but most sets are understandably disinclined to place their most unique questions where most people will never see them.
Ben Miller
Montgomery Blair High School '18
"William the Bastard"
User avatar
benmillerbenmiller
Lulu
 
Posts: 26
Joined: Fri Feb 17, 2017 6:25 pm

Re: Tiebreaker Distro

Postby Skepticism and Animal Feed » Sun Jul 23, 2017 10:41 pm

I'm not aware of any deep theory that's been written on tiebreakers. Below, I will talk a bit about the approach that I took to tiebreakers back when I was writing and editing lots of questions (which was a long time ago, but I like to think some of the ideas might still apply today).

In my opinion, the most important thing a tiebreaker tossup has to do is be answered. If the tiebreaker goes dead, that's just annoying for everyone and has the potential to delay the tournament. So first and foremost I tried to make my tiebreaker tossups accessible - it was the place for easier answerlines. For example, once I wrote a tiebreaker tossup from the mythology category and the answerline was "blue". It was all about gods or items that were blue. There were hard clues at the beginning, but by golly somebody was going to answer it by the end because everyone has heard of the color blue.

I also preferred to write tiebreakers from the "Big 3" categories: Literature, History, Science. My reasoning there is that these are probably the most popular and well known categories (and, because they each make up a big percentage of the set, they're the categories players are most incentivized to study) and as such tiebreaker tossups from these categories were, in my view, more likely to be answered. I realize that my example from above is an RMP tossup, so I didn't always stick to this, but I tried to make the bulk of my tiebreaker tossups from the Big 3.

The unfortunate reality is that tiebreaker tossups (and questions in tiebreaker packets) tend to be rejects that were put in there last minute after being kicked out of another location in the set. That's bad.

If you haven't written the tiebreakers for your set yet and you are low on time, one approach is to just make a separate document with a few tiebreaker tossups on it. Then, when tiebreakers are needed, read questions from that document. Odds are, you won't need more than just a few tiebreakers over the course of the tournament, just keep track of which teams have heard which tiebreakers. Obviously this takes more time and is not ideal, but if the alternative is to have low quality tiebreakers at the end of each packet it's certainly better.

Ultimately, I don't think things like "is this answer fraudable" are any more important in tiebreakers than any other tossup.

EDIT: to be clear I'm talking about 1 TB packets above. If you're doing 3 TBs per packet I'd say diversify beyond the Big 3, but not a lot.
Bruce
Harvard '10 / UChicago '07 / Roycemore School '04
ACF Member emeritus
My guide to using Wikipedia as a question source
User avatar
Skepticism and Animal Feed
Auron
 
Posts: 3093
Joined: Sat Oct 30, 2004 11:47 pm
Location: Arlington, VA

Re: Tiebreaker Distro

Postby Skepticism and Animal Feed » Sun Jul 23, 2017 10:56 pm

Theodore wrote:teams have already heard 20 TUs to differentiate themselves


I think this is a very powerful rebuttal to a lot of nitpicks about tiebreakers. Over a pretty decent sample size of questions these two teams have proven to be evenly matched: there's not really a "wrong" outcome so long as the question meets basic good quizbowl principals.
Bruce
Harvard '10 / UChicago '07 / Roycemore School '04
ACF Member emeritus
My guide to using Wikipedia as a question source
User avatar
Skepticism and Animal Feed
Auron
 
Posts: 3093
Joined: Sat Oct 30, 2004 11:47 pm
Location: Arlington, VA


Return to Theory

Who is online

Users browsing this forum: No registered users and 1 guest