Saturday, September 05, 2009

Pounce and Bounce

A guest post by old BC alumnus Hirak Parikh, who now quizzes in Ann Arbor, Michigan. The original appeared on his blog here.

A veteran quizzer once told me that quizzing was not so much about knowing stuff, but was more about entertainment and having fun. I disagreed at that point thinking that such a remark was an act of high treason. Over the years, I have realised that quizzing is not a knowledge test. If it were so then quizzes would be like exams - timed and written. A written and timed test is the fairest way to judge who knows the most. All teams get the same questions, no question of order or luck. The team/person that got the highest number of correct answers would walk home with the prize. That's not the point, is it? The point is to have fun.

Over the years, people have obsessed over the best format that minimizes luck and ensures that the team that knows the most (read: best) wins. Quizzing innovations is an activity which has almost become a kind of cottage industry among quizzers with time on their hands. I am about to add more to that body of literature. That being said IMHO, my experience suggests that there is no point looking for the Holy Grail of a perfect format. A good quiz depends on ensuring three simple things in order of importance:

a) The format should be the 'modified infinite bounds with a midway reversal'. All questions have equal points, the next question to the team to the left/right of the team that answered it. If the question is unanswered then the original teams gets the next question. Order reverses halfway during the quiz.

b) A long quiz.
I think for 5-6 teams there should be at least >40 questions. A factor of 10 is ideal. Even the old system of rounds with pass-direct questions (full points for direct, and half for a pass question) would be okay provided the quiz was long enough. c) Questions, questions, questions.
The greatest evil is not the format, or the order, or marking scheme -- but bad questions. If questions are set correctly, spread over different topics evenly, and are of similar difficulty then the quiz will be fair and the best team 'should' win.

In practice, luck and order does play somewhat of a role even in the modified infinite bounds format, though b) should take care of it to some extent. The critical issue is to normalize questions in some fashion. Why normalize? Each quizmaster (QM) has his/her own personal strength (read: fetish) and quizzes as a result tend towards personal idiosyncrasies. The good QM is diligent about this and goes about setting questions keeping in mind those biases (knowing the QM and his/her strengths can help you work out the answer, cause you can guess what he/she knows, and how he/she sets questions). One way to do it, as is usually done for big quizzes, is to have two or more question-setters with divergent interests.

Despite the best intentions of the QM(s): all questions are not equal, some are more equal than others. A QM may think that a question is reasonably difficult, but may turn out to be a sitter. On the other hand, some questions are way too tough and end up being unanswered. Personally, if more than 10-15% of questions end up this way then the QM did a bad job. You cannot go about an weigh each question for difficulty.

The other aspect is that there should be some drama, some element of excitement in a quiz. I am a fan of 'buzzer-rounds' which have fallen out of favour in recent times. It provided that adrenalin-rush and rewarded quick recall and reflexes which are sadly missing from the current slow-cooking style. The current trend is away from the fireworks and some quizzes have written components at the start. This is bad, bad, bad.

Long story short. I tried out two innovations at the quizclub. Each team was given two wildcards to allow them to Pounce or Bounce a question.

Pounce: You can attempt a question out of turn. The team has to write down the question or tell the quizmaster before the question is attempted by any of the other teams in regular fashion. There are no negatives and a correct answer get full points.

Bounce: You can bounce a question to the team of your choice. If team doesn't answer it correctly they get -(full points), and if they do answer it correctly they get the full points. Regardless of the outcome, the team bouncing gets the next question.

The rationale behind the Pounce rule was to ensure that sitters can attempted by all. In the past, with great difficulty I have resisted urges to destroy the chair I was sitting on, or strangling the person who smirks when handed a sitter as a direct while I was left wringing my hands in despair. Often, close quizzes are decided on the basis of which team got slightly easier questions. This is where the Pounce comes in. Jump in on question out-of-queue. Grab a sitter. Of course, a team can misjudge the opponents knowledge and would end up with the question in the regular course of events.

The rationale behind the Bounce rule is to induce some excitement and additionally serves as a handicap for the obviously better team(s). There are always going to be a few questions that seem so unreasonable and tough that no one can answer them and a weaker team can either direct it towards the strongest team, or to their closest rival to level the playing field. Of course, if a team is really good and they can actually answer the question that seemed 'too tough' then all the better for them.

Innovations in practice:
This weekend when I tried these out and found the results mixed. No one team used the Bounce rule. They all played too nice, perhaps fearing retribution. The Pounce rule was used by all teams to good effect. Only once out of six times did a team not answer the question correctly. In all six cases, the team would not have got the question in the regular order and judged the moment of 'pouncing' correctly.

This was the casual Saturday quiz and not the best testing ground when there were only three teams and no one was too worried about winning or losing as there was nothing at stake. I curious to see how this works out in a longer, larger quiz more at stake. These rules do favour teams that can fake emotions of knowing or not knowing answers depending on the situation. A tight quiz can almost be like a poker game.
You can sometimes work out the answer being seeing who knows it.

2 comments:

Harish Kumar said...

Should try it out in the next Open quiz.

Anonymous said...

hi i have done a whole quiz at BCQC with the Pounce rule in every round...But negative 2 for each wrong pounce...to prevent pouncing for the heck of it...

I hope the members at BC rem it

Regards, Brijesh