This post contains spoilers for episode 6 of season 8 of RuPaul’s Drag Race.
This week’s mini challenge was the namesake for this blog series: The Library is Open
In this mini challenge, the Queens must put on their reading glasses and read the other Queens. The goal is to make the Queens, and especially RuPaul, laugh while throwing some good natured shade. While most Queens are funny, Bob the Drag Queen wins the challenge hands down.
For this week’s maxi challenge, the queens must create runway looks inspired by characters from one of RuPaul’s favorite books: the Wizard of Oz. Joining the queens is the cast of Little Women L.A., who had previously been assigned a character from the Wizard of Oz and paired with a queen. The queen must create matching runway looks for themselves and their partner, making over their partner as their drag daughter.
Naomi Smalls is assigned Jasmine Arteaga Sorge and the Scarecrow. Her high fashion look, and a change in silhouette from past runway looks, earns her the win for this challenge.
Derrick Barry is assigned Terra Jole and the Tin Woodsman, but comes down the runway looking like the Tin Woodsman’s “vibrator.” Her look earns her a lipsync for her life.
Robbie Turner is assigned Christy McGinity Gibel and the Cowardly Lion. Unfortunately, Kim Chi set the bar for this look pretty high in the very first episode of this season, and Robbie doesn’t meet the bar, joining Derrick in her lipsync.
After an intense lipsync for their lives, Robbie Turner must sashay away.
The Algorithms
Before I get to the algorithms’ performances, I have a quick update. Previously, for the neural network, I had been scaling the entire data set at once, then dividing it into the training, testing, and season 8 subsets. This meant, however, that the values of the testing and training sets would change each week as I added new data for season 8. To stabilize the training and testing of the neural network, I now split the data into different subsets and then scale each subset separately. This has produced a lower Kendall’s Tau for the neural network based on its season 7 predictions: 0.666. I’ve updated the calculation of the weighted average score for season 8 predictions using this new tau.
I’m also now calculating a Kendall’s Tau based on the season 8 predictions. Before calculating this tau, I remove queens still in the running (so they don’t have a place yet) and then calculate tau based on those queens who have gone home. You can find each algorithm’s tau at the bottom of the table below.
Last week, the algorithms predicted that Naomi Smalls was going home this week, but she won the challenge. Robbie Turner wasn’t expected to go home for at least a couple more episodes, so her elimination this week is unexpected as far as the algorithms are concerned. Feeding in the new data from this week, the algorithms have updated their predictions. Naomi Smalls is now expected to make it into the top 3, along with Bob the Drag Queen and Kim Chi (ignoring Acid Betty who has already gone home). Chi Chi DeVayne is now predicted to go home next. With only six queens left in the season, the competition is heating up!
Actual | Support Vector Machines | Gaussian Naive Bayes | Random Forest Classifier | Random Forest Regressor | Neural Network | Average Predicted Score | |
---|---|---|---|---|---|---|---|
Tau | 0.54 | 0.519 | 0.571 | 0.5 | 0.624 | ||
Bob the Drag Queen | 3 | 1 | 1 | 1 | 4 | 1.890 | |
Acid Betty | 8 | 1 | 1 | 1 | 8 | 1 | 2.873 |
Naomi Smalls | 2 | 1 | 6 | 4 | 1 | 3.052 | |
Kim Chi | 12 | 1 | 4 | 2 | 4 | 3.583 | |
Thorgy Thor | 9 | 1 | 6 | 4 | 1 | 3.616 | |
Robbie Turner | 7 | 3 | 9 | 3 | 2 | 6 | 4.513 |
Derrick Barry | 3 | 6 | 6 | 4 | 6 | 5.223 | |
Chi Chi DeVayne | 3 | 9 | 5 | 4 | 6 | 5.515 | |
Naysha Lopez | 9 | 3 | 7 | 9 | 10 | 9 | 8.433 |
Dax ExclamationPoint | 11 | 3 | 11 | 10 | 10 | 9 | 9.369 |
Laila McQueen | 11 | 11 | 11 | 10 | 9 | 9 | 9.746 |
Cynthia Lee Fontaine | 10 | 9 | 7 | 12 | 12 | 9 | 10.153 |
Further Reading:
- The first episode of The Library is Open
- The code and data for generating the predictions