Top 10 prediction (updated)

Name Song WNTS DI VF Not-safe
Curtis Finch I Believe 24 0 2 0.407
Janelle Arthur Gone 43 0.544 2 0.383
Paul Jolley Amazed 47 0.563 3 0.367
Burnell Taylor Flying Without Wings 50 0 4 0.358
Devin Velez Temporary Home 46 0.563 6 0.336
Lazaro Arbos Breakaway 31 5.171 5 0.326
Amber Holcomb A Moment Like This 71 0.139 10 0.276
Kree Harrison Crying 64 0.515 14 0.244
Candice Glover I (Who Have Nothing) 89 0.417 14 0.226
Angela Miller I Surrender 68 3.266 41 0.078

(Updated: 9:33am on 3-14-2013 with final numbers.)


The methodology for the finals model is described here. The model is 87% accurate on ranking within a margin of error of +/- 3%.  Probabilities being what they are, somebody with a not-safe probability of just 0.25 will be in the bottom 3 one out of four times. Please do not comment that the numbers are wrong. They are probabilities, not certainties or even claims. Do not gamble based on these numbers.

Names in green are predicted safe. Names in red are predicted in the bottom 3. Names in yellow are undecided. Any person not in green is not considered safe by the model. The most probable bottom 3 is Curtis, Janelle, and Paul. However, anybody on the list (except Angie) being in the bottom 3 would not be shocking.

The contestants the model thinks are safe are notable in that they are nearly all women. Devin Velez Lazaro Arbos,, is the least of all his male cohorts. Indeed, it was a great night for women, and only so-so for men. My feeling during the liveblog was that Devin did enough, and the numbers bear that out [Update: Devin has been moved into uncertain territory]. Candice, Kree, Amber, and Angie are predicted safe even at the 95% confidence level.

Curtis Finch isn’t looking good. He was second least popular on Votefair, had a zero on Dialidol, and was also last on WNTS. He’s the official pick to go home, and I’m comfortable with that given my feeling and the numbers. However, people in his position were still safe 60% of the time. Janelle and Paul were people who I thought squeaked by in the semifinals, and so, again, I’m pretty comfortable with them as picks.

The model weighs WNTS and Votefair more heavily than Dialidol this early in the year. As time goes by, the situation reverses itself, and Dialidol becomes dominant. We are far from that point, and the low score that Amber got on Dialidol I don’t think should worry anyone too much.

The fact that the semifinals model was flawless this year gives no indication that the finals model will be as well. The finals model is always less accurate, as it works off a much smaller data set and smaller set of variables. I fully expect one or two of these calls to be wrong.

Bookmark the permalink.

Comments are closed.

Comments are closed