Top 7 prediction (final)

Name Song WNTS DialIdol VoteFair Not-safe Probability
Burnell Taylor You Give Love A Bad Name 28 0 1 0.725
Lazaro Arbos We Are the Champions 20 1.065 3 0.658
Janelle Arthur You May Be Right 55 0.527 10 0.491
Candice Glover (I Can’t Get No) Satisfaction 65 0.885 18 0.340
Amber Holcomb What About Love 67 2.224 16 0.335
Kree Harrison Piece of My Heart 52 1.465 20 0.313
Angie Miller Bring Me To Life 71 4.208 32 0.138

Final update 2:35 PM. No change in ranking of the bottom 3, but Candice dropped a little relative to Amber. They’re basically neck-and-neck.

Updated 9:49 AM. Final update probably around 2pm since I have a job. There were no changes in the ranking from the update.

Important!

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 the most likely 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 Burnell, Lazaro, and Janelle. However, anybody on the list being in the bottom 3 would not be shocking.

The Idol Analytics model was sort of devised to be in line with how I think Idol works. But I’m beginning to think maybe it’s a little too representative of that this year.

In any case, there are no surprises in this listing. Janelle has been considered the weakest girl, though she never has been in the Bottom 3, so the model could (of course) be wrong. All predictions are beyond the margin of error at 87% confidence. Burnell and Lazaro are predicted not-safe at the 95% confidence range.

Dialidol looks to be having a lot of volatility. I watched Candice go from #1 to #4 in a span of about 20 minutes tonight. I don’t really know how to account for that (since, I confess, the Dialidol weighing algorithm is a bit opaque to me).

I never have put a lot of stock into the idea that certain songs or even artists determine the disposition of a contestant, but for what it’s worth I’ve compared the song selections tonight to the song choice indices I’ve developed here and here. Candice and Kree made choices that were above average (both for song and artist), and Amber made bad to mediocre picks. Janelle and Lazaro picked below average artists to cover, though their actual songs were not common enough to rank. No data is available for the other picks.

WhatNotToSing has an editorial up arguing that the judges’ save has to be used tomorrow … or it doesn’t. Their reasoning goes that the judges will never go a season without using the save, and since next week is the last possible round to do so, anybody next week will know that nothing counts. This is not convincing at all to me, though your opinion may differ. I think it’s easily possible the save will not be used this year.

Bookmark the permalink.
  • Louis

    Really like your work but this week means you need to add a new metric – the Producer KILL FACTOR IN THE CURRENT FORMAT. The show no longer draws an adequate audience against the comparables, externally verifiable vote counts are down and a total dud is in the top three.