To Ask Or Not To Ask (the Audience)

Revisiting Who Wants to Be a Millionaire’s trickiest lifeline in a Cedric the Entertainer world.

Jackie Fuchs (Fox) on Who Wants to Be a Millionaire

In game show parlance, I got “dursted.”

But then relying on someone else for an answer on has always been an iffy business on Who Wants to Be a Millionaire.

Just ask comedian Rudy, who was a contestant on the show in February 2000.

Reber was in the hot seat with $250,000. If he answered the next question correctly, he would earn $500,000 and see the million dollar question. If he missed, he would drop to $32,000 and be out.

The question: Which Hollywood figure directed Michael Jackson’s 1988 video “Bad”?

Reber didn’t know. But he had his phone-a-friend “lifeline,” which he used to call fellow comedian, Will Durst. Durst told Reber he knew for a fact that it was John Landis. Reber took Durst’s advice and made it his final answer.

That misstep cost Reber $218,000 and bought Durst game-show immortality as a verb. (For the record, Landis directed Jackson’s “Thriller” video; Martin Scorsese directed “Bad”).

The phone-a-friend is long gone, along with the 50-50 and various other lifelines which have followed over the years. They’ve been replaced by the more reliable “jump-the-question” (or “JTQ”), which contestants can use up to twice to bypass a question altogether, along any money they might have gotten from it.

A more devious change is in the order of the first ten questions (Round 1), which is now random. Contestants aren’t told the difficulty level of their questions. As a result, it can be hard for a contestant to gauge just how many difficult ones they’ve seen, and whether any “wicked weed outs” are likely to come.

Yet despite these and other changes, one thing has endured: the ask-the-audience (ATA) lifeline, which has been there from the start. It’s interactive. It’s dramatic. Best of all, it’s unpredictable.

Or is it?

On the day of my show, before the morning’s taping began, Millionaire’s executive producer, Rich Sirop, showed the contestants some sample questions. The idea was to teach us what made for a good ATA question. But the guidelines were somewhat vague, and I had little time to digest the information. I was up almost right away, the second contestant on the first of five episodes that would tape that day.

I breezed quickly through five of my first six questions, using a jump on the third to be safe.

I was feeling pretty confident as question 7 appeared on-screen:

Boasting it "helps sell records," what rap act put their own "explicit content" sticker on 1987's "Rhyme Pays"?

A : Public Enemy B: LL Cool J

C: Ice-T D. Beastie Boys

Here was the category I’d dreaded going in. My first inclination was to use my second jump. But three of my first six questions had seemed really easy, and even the one I’d jumped hadn’t been that hard. Which meant a there were probably hard questions still to come. I might need that jump. And since I’d gotten low random dollar amounts to that point, the question was worth between $5,000 and $25,000.

A little less knowledge would have served me well. If I wasn’t so familiar with these artists, I would have used my last JTQ and not thought about it twice. But rock radio had played the Beastie Boys a lot and I was able to rule them out. And I had negotiated deals for LL Cool J.

1987 might be pushing the envelope time-wise, but I was not a big fan of rap. Maybe “Rhyme Pays” had been a huge album. If so, the audience might know. And if they didn’t, the distribution of their responses should be more-or-less random, in which case I’d use my last jump. Sure, I’d be burning my ATA. But it would give me a chance to save the more valuable jump.

The audience’s response surprised me:

A - 47%

B - 10%

C - 22%

D - 21%

Neither the overwhelming majority I’d hoped for, nor the random distribution I expected. Still, Public Enemy was the top response by more than a 2-1 margin. And a full one-quarter of the audience had picked them over the next closest choice, Ice-T. I decided it was enough.

As durstings go, mine wasn’t nearly as bad as Reber’s. But it did mean I lost money on my appearance, since Millionaire doesn’t pay for contestants’ hotel or airfare. The $1,000 consolation prize might be nice for locals, but it doesn’t cover the cost of cross-country airfare and two nights in a New York City hotel. Not to mention the costs of a trip to Las Vegas for the audition.

So what exactly went wrong?

Was it just that 1987 was too far back after all? Or is that Millionaire audiences are like me, knowledgeable about music generally, but not big on rap?

To find out I turned to the WWTBAM Bored, where a group of loyal fans posts unofficial transcripts of each show. I analyzed the 100 most recent Round 1 ATA questions, sorting them by category, difficulty level, and the time period the question addressed.

The audience’s overall success rate was 77%, down from a reported 91% during the PAF/50-50 era.

As expected, difficulty level proved the best predictor of audience accuracy. At levels 1-5, the audience was 100%. But when you’re on the show, determining difficulty isn’t easy, especially early in the round. Were there any other guidelines that might have helped?

As it turns out, there were.

The audience was almost always right when:

1.At least 50% of the audience agreed on the top response (93% accuracy); OR

2.There was a 20% or greater difference between the top two responses (92% accuracy).

And they were usually wrong whenever:

1.Less than 40% agreed on the top answer (15% accuracy); OR

2.Less than 20% separated the top two responses (38% accuracy).

Category and time period proved less relevant than one might expect. Only two categories stood out as problematic. The first was history, on which the audience was a dismal zero for three.

Their other bad category was – you guessed it -- rap. They were one for three, and the one they got right was also a film question.

Perhaps most notably, the audience did well on other music questions, going 11 for 14.

So is rap Millionaire’s contestant killer, or is something else at work?

I was told afterward by some people at my taping that the audience missed every question they were asked that morning. They also told me that they’d been misled by the words “their” and “act” in my question into thinking the answer had to be a group.

Maybe it was the audience, then. Maybe it was the show’s writers. More likely, however, the questions were just too difficult. And my guidelines would have weeded out the audience’s errors, even in my case, which was a pretty close call.

In retrospect, my decision to go with the audience wasn’t terrible. There was a 25% margin between the top two responses, and the highest was one I hadn’t ruled out.

So, is rap really the problem, or does something magic happen at 50%? I’ll leave that one to academicians to sort out. But as considerable as Sirop’s wisdom is, I’d have gladly traded it for my numbers.

Future WWTBAM contestants, you’re welcome. But if it all goes wrong, don’t say I didn’t warn you. The ATA can be fickle. And a mind is a terrible thing to Durst.