Dig if you will this picture.... Ryan Seacrest intones "This...Is American Idol!", kicking off a new week's competition show. He banters with the audience for a bit, sends a shout-out to the
three four judges, and introduces the surviving contestants from the previous week. Nothing out of the ordinary yet...until Ryan unexpectedly announces that the show is over and the phone lines are now open. Vote for your favorite early and often, and we'll see you here tomorrow night when we announce the results. Everyone smiles, roll credits, fade to black.
For a moment, put aside the absurdity of this little thought exercise (chief among them that Fox and the producers would ever dream of foregoing an hour's worth of advertising revenue.) You'd probably agree that while the outcome of the vote might be highly unsatisfying, it surely can't be called unfair. In fact, it might be the most equitable elimination in the show's history. For once there were no silly themes that favored one contestant over another, no disadvantages stemming from the performance order, no biased intro clips, no hanky-panky with the lighting and camera angles, and no pimping or depimping from the judges. The Idols just stood there and waved. The elimination was decided solely by who happened to be the least popular among the voters before the show even started.
In the real world, of course, all of those "production" factors play a part, as do incidental matters like, you know, song choice, presentation, and singing quality -- i.e., "performance" factors. It seems obvious to say that the voting results from Smile & Wave Night would have been different...often monumentally different...if the contestants actually sang something.
Perhaps next week, we'll set up a different experiment using eight contestants you've never seen before in your life. That will allow us to eliminate "historical" factors in the decision, isolating only performance ones. The following week, we'll figure out a way to isolate production (i.e., pimping) factors. After that, we'll force the judges and producers to sing, and we'll put the contestants at the judges' table to rip them to shreds, which won't prove anything from a scientific standpoint of course, except we'd absolutely kill to see it....
But perhaps it's time we got to the point.
An enormous array of variables go into the making of an Idol champion. Performance quality plays a big role, of course, but as we all know (and as our off-season project, Camp Should-A-Been, helped demonstrate), it's far from the primary factor. The winning contestant has to project an image, build and extend a fanbase, stay on the judges' good side, get lucky here and there, and sing well enough to keep the train rolling through three-plus grueling months.
Is there a way to isolate and quantify these factors? For example, might we be able to predict with some accuracy how the voting would turn out if the producers really did stage a Smile & Wave Show one week? The relevant historical components might include demographics (age, sex, ethnicity), average performance rating, recent and/or early performance ratings, and number of times previously in the Bottom Three (and weeks since latest appearance), plus indirect evidence from stuff like recent Dial Idol numbers, Audition and Hollywood Week screen time, Google hits, number of forum fan postings, etc.
Over the past few months, we've been discussing such a project with our friend Leo, The Idol Guy. He recently published a study on his site showing that song age seems to play a role in survival patterns — all else being equal, contestants who choose younger material tend to outlast ones who rely on golden oldies.
Here at WhatNotToSing.com, meawhile, we showed how singing first increases your chances of elimination by 50% or more, and our Harvard correspondent Angela Antony extended this analysis to all performance positions, showing that some are clearly safer than others. Most Idol fans already know that a spectacularly bad performance by an established contestant is usually a little safer than a merely poor one (the "Mighty Mouse Syndrome"), but astoundingly, if there are two train wrecks on the same Finals night, woe be to the contestant with the third-lowest rated performance (the "Sesame Street Effect").
The Idol Guy coined the term Idolmetrics to refer to the scientific study of the factors that go into winning and losing on American Idol. Over the course of the next few months, visitors to our site and TIG's will be encountering that word quite a lot. We plan to spend much of Season Eight in the pursuit of Idolmetrics: that is, forming hypotheses, isolating variables and running experiments in the hopes of shedding some light on the shadowy, arcane world of Idol success and failure.
If we're successful enough, we might someday even be able to attain the Holy Grail of AI: an accurate, predictive formula for long-term survival on the show...and, not incidentally, a blueprint that studious contestants of the future could use to plot a course to the crown. If you plan to be one of them, just promise us free tickets to the Finale, 'k?
As always, we welcome the input and talents of our readers in this endeavor. If you would like to run an Idolmetric experiment of your own, let's talk. Send us your hypothesis and a description of an experiment to prove or disprove it. If we think it's worthwhile and attainable, and you need some data from us, we'll run the appropriate queries into the WNTS.com database and send you the results in a spreadsheet. Then, write up your conclusions and we'll post them in our Library. Yeah, it's a lot like Chemistry lab in high school, except it's fun, you won't be graded, and there's no risk of an experiment blowing up in your face. Er, at least not literally.
Please understand that seven years, 112 episodes, and 983 performances isn't really enough on which to draw firm conclusions. But, that's okay. It's still a wealth of evidence – more than enough, we believe, to identify likely trends and perhaps to put a few urban legends to rest along the way. Besides, if we waited until we had truly sufficient data, we'd all be old, gray, and feeble, and the only Idol-related job we could perform would be as a judge. Where's the fun in that?
- The WNTS.com Team