Top 8 projection (final)

Name Song WNTS DI VF Not-safe Probability
C.J. Harris Soulshine 43 n/a 1 0.592
Dexter Roberts One Mississippi 62 n/a 6 0.477
Malaya Watson Ain’t No Way 65 n/a 7 0.456
Caleb Johnson Chain of Fools 67 n/a 12 0.368
Jessica Meuse Blue Eyed Lie 86 n/a 14 0.325
Alex Preston Fairy Tales 84 n/a 16 0.298
Sam Woolf Lego House 48 n/a 18 0.292
Jena Irene Rolling in the Deep 62 n/a 26 0.192


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 considered at risk for being 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 C.J., Dexter  and Malaya. However, anybody on the list being in the bottom 3 would not be shocking.

Updated 3:48PM EST.

Updated 11:00AM EST. No change in rankings.

Alex Preston and Jessica Meuse both received excellent marks for their original songs tonight. C.J. Harris had the lowest reviews for his song, Soulshine, which was already performed by Ben Briley this season. With low popularity, a prior trip to the bottom 3, and having been a wildcard to begin with, C.J.’s chances are grim.

Sam Woolf also got rather poor reviews for his song, Lego House, but remains popular. As the day goes on, this indicator has swung wildly for him, so watch this space as the numbers for him are likely to change.

My hope for Dialidol’s resurgence is dwindling, but feeling around for an alternative has yielded poor results. The MJsBigBlog poll, for instance, gives almost no information that Votefair doesn’t give. Below is plotted the Votefair share of vote versus the MJsBigBlog poll share of vote for favorite:

MJsVotefairAs you can see, these two are reasonably correlated. The two largest outliers are 1. a gross underestimate of Phil Phillips’ popularity and 2. a gross overestimate of the popularity of Jessica Sanchez by Votefair relative to MJs poll. But it’s quite uncorrelated with Dialidol, having R2 of only 0.276. This indicates that it’s a good substitute for Votefair, should that service stop existing, but as a substitute for Dialidol it simply won’t do.

Dialidol presented a different piece of information, information about vote-getting, which is quite a different thing than being popular or being a good singer. A random poll of tweets tonight showed that C.J. and Malaya had the least number of people tweeting that they had voted for them, but it’s anybody’s guess as to whether this has any predictive power. People on Twitter are a subset of the voting public, and there’s no a priori reason to assume that that subset is representative of the entire set.

Bookmark the permalink.

Comments are closed.

Comments are closed