Sunday, February 01, 2004

Formats of Rounds - I

Setting a good quiz, apart from just the raw content of questions, depends a great deal on the structural framework of a quiz. Or put less glibly, formatting of different rounds is extremely important to how it is received by audience and participants alike. A fine balance needs to be struck between retaining the interest of the spectators while being fair to the participants. ("Fairness" will be a constant refrain on these pages, for a majority of such formats tend to be tuned towards providing semi-gladiatorial pleasure for the audience, and letting the contestants fume and wish for fortune cookies to crumble their way.)

D & P and Forward Biasing

If you wander into a quiz and see the organisers pull out the dinosaur called the Direct and Passing method, then you can be sure that (a) they have never participated in good quizzes themselves (b) Do not believe in modern concepts such as being fair to participants or (the most likely) (c) lack the common sense to spot the inherent flaws in the system.

The D & P method has been discarded long ago by all well-meaning and informed quizzers. Those who persist in its use would do well to consider these points. The D&P method (also known as the 10/5 system or the conventional passing system) essentially consists of offering a team a direct attempt at a question for a value, say 10 points. If that team fails to answer correctly, the question passes to the next in order, but only for a lesser value, usually half, in this case 5. The explanation usually given for it is that it is easier for the teams following as possible options have been ruled out. Superficially, this may be a valid point, but veteran quizzers will easily provide a contradictory argument. This is the fact that on many occasions, the answers provided by a team create a certain trend of answering (especially if the quizmaster seems to respond to that answer in a manner that suggests it may be in the general direction of the right answer) that consciously or otherwise is followed by the rest. Being misled by the previous answers is a trap that all quizzers fall into quite frequently.

For instance, at Chakravyuh 2001, a visual connect consisted of linking images varying from that of the sun, moon, some of the planets, people such as the Curies & Einstein, a map of a Swedish district, Berkeley and so on. It was marked by a large number and wide variety of these images. The first team to take a shot, went by the more recognisable faces of Einstein, Rutherford etc to guess atomic programmes and World Wars. All the rest of the teams veered in that same direction, and ended up as a result, way off the mark - clearly the first team had infected their responses. The answer, by the way, was that these items had inspired the nomenclature of various chemical elements. I call this Forward Biasing (well, it's inspired :-) )
So clearly, providing different values to a question isn't right. Perhaps this is some latent socialist streak among enlightened quizzers, but it is fair to expect that all answers have the same value attached to them, and there is no need for a team only to get more because they got first crack at that question.

Another pain point about the D&P method concerns the other major pillar of the method: deciding the recipient of the next question. Here, since the order is static, the recipients of a question are pre-decided in a round. Other events do not influence this decision of which question goes to which team. Coupled with the 10-5 system, this can have disastrous effects on the outcome of a team, which has grounded the chances of many a good team. Consider this: Team A gets a question, can't answer, qn passes to B who gets it right and hence scores 5 points. The next question (and this is the killer) goes again to Team B who answer it direct to get 10 points. Suddenly, they are up to 15 with none of the others (except A) having touched a ball, to use a metaphor. Additionally, it is a common occurence in this method that a team answers say 5 questions, all from passes to get 25. Another team may have answered just 3 questions, all direct, and hence surpasses the former to be on 30. In the longer run, teams can suffer just because they answered after someone else, though they might have had to fight harder.

Another personal experience may highlight the above point. In Chakravyuuh 2002, Gaurav asked a question which ran something on the lines of:

Which TV programme has featured the guest voices of the following people: (I don't remember any specific names, mostly a list of actors & actresses).

Easily the most obvious answer to this was The Simpsons cartoons. Harish & I were fifth in line, so like the others, we thought it was a gimme because it had to be The Simpsons. But this was Chakravyuuh. The first team answered that, and to the general amazement of all, that wasn't right. This made everyone else scramble for alternative guesses, which weren't easy to come by as our first thoughts were on the lines of other cartoons, another classic case of Forward Bias. Soon it was to be our turn, and we had to think of something good, especially in the wake of a Guess-For-All situation. Finally it struck us - Frasier was the right answer - it featured guest voices who call into Frasier Crane's Radio show. It helped that like other BC quizzers, we followed the credits of films/serials and it did stick in. It did turn out to be correct answer. I can tell you that I'd have been extremely cross if we had got only 5 points for that toughie because some idiot decreed that was what we could get for a passed question.

So to be fair, the best and easiest method is to do away with different weights for the same question. Each right answer must have the same value attached to it - this has the simple result that the team that answers the most questions will win the quiz. Additionally, in those hostile places where order of teams are decided by underhand methods, the D&P method may make it easier to rig questions as the 2nd question in each round will always go to Team B (but of course, oil-riggers are smarter than this and will always carve their own way out :-) )

So what are the alternatives? By far the most popular and the most accepted one is called Infinite Rebounds (a.k.a Infinite Bounds). A discussion on it next time.

Comments/Criticisms/Commendations/Contributions welcome as usual

No comments: