I’ve been watching lots of episodes of “Drop the Mic” on Youtube lately with @klam227 and I think I noticed a pattern on how the winners are selected. It got to a point where I will guess who the winner is before the episode really gets underway and I would be right 8/9 times.

If you haven’t yet found the simple pleasure of watching Drop the Mic, here’s a summary from IMDb:

Hosted by Method Man and Hailey Baldwin, each week Drop the Mic pits four stars from the worlds of entertainment, music, sports and pop culture to face off against each other in a rap battle royale packed with lyrically creative genius and “Did they really just say that?” moments. At the end of each head-to-head battle, there can only be one winner.

And here’s one of the episodes I particularly liked:

Now, granted I have not watched every single episode, but of the ones I did watch I noticed two factors which determined the winner.

My simple-in-head-algorithm for determining the winner is:

1. If one contestant is female and the other is male, the female wins!
2. If both are male, or both are female, or if they are a group, the one standing to the left of the screen at the end wins!

😄

Now I want to actually see how good this algorithm is!

First off, scraping the episode data from IMDb:

library(rvest)
library(tidyverse)
library(knitr)
library(kableExtra)

# Scrape data from IMDb
data.dropmic <- read_html("http://www.imdb.com/title/tt6396094/episodes?season=1&ref_=tt_eps_sn_1") %>%
        html_nodes("strong a") %>%
        html_text() %>%
        as.data.frame()

# Title column
names(data.dropmic) <- "Episode"

# Display results
kable(data.dropmic, "html") %>%
        kable_styling(bootstrap_options = c("striped", "hover"))

(The pretty table styles are courtesy of @haozhu233 and his excellent kableExtra package!)

In the above table, there are some non-episode data in the last few rows. Let’s filter that out with the following code:

# Filter
data.clean <- data.dropmic[grepl(" vs", data.dropmic$Episode),] %>%
        data.frame(Episode = .)

# Display results
kable(data.clean, "html") %>%
        kable_styling(bootstrap_options = c("striped", "hover"))

Excellent. Now just one more minor step to split each episode up into “battles”, since each episode contains two battles:

# Split into battles
data.clean <- str_split(data.clean$Episode, "&| and ") %>%
        unlist() %>%
        data.frame(Battles = .)

# Display results
kable(data.clean, "html") %>%
        kable_styling(bootstrap_options = c("striped", "hover"))

Perfect. Now we can add the two features of this algorithm and watch every episode on the list in order to record the winner!

Looks like my algorithm scored 85%!

And the prediction for the last two episodes (which haven’t yet been posted on Youtube as of Jan. 8 2018) are:

1. Danielle Fishel will beat Jonathan Lipnicki (female)
2. Whoever is standing on the left side of the screen (looks like Shania)

Well, that was certainly an enjoyable afternoon! 😄