After the debacle that was my use of Twitter to predict the outcome of American Idol, I decided to attempt redemption in the form of another online tool.  After doing some research, I realized that many fans of reality television were using grids made on Wikipedia to calculate statistics used for elimination predictions.  Statistics can be created for any show that uses a high-low-win system or a straight callout order system.  Donning my nerd cap, I delved into the world of reality television and Wikipedia articles to see if I could discover their true usability.

Callout Order Methodology

For the first two shows (America's Next Top Model and Flavor of Love), I calculated the stats by taking each contestant's weekly callout order number and averaging all of those numbers together.  For example, if a particular contestant was called 5^th, she received a ‘5' for that week.  In theory, the contestant with the highest (worst) callout average should be the next contestant eliminated.  Since each of these shows is down to the Final Three, I also went ahead and predicted the winner, also based solely on statistical data.

Interestingly, current Top Model finalist Anya has what could be considered the best callout average in the history of the show.  This blog calculates stats for all the cycles, and with a current callout average of 2.8, Anya is statistically the best contestant to ever appear on Top Model.  Fellow finalists Whitney and Fatima fare far worse, with callout averages of 4.3 and 4.5, respectively.  While the lower averages are very close, I am going to stick to my theory and predict Fatima being the next eliminated model, with a definite Anya victory.

For the next study, I took a popular VH1 dating show, Flavor of Love, which is now is its third season.  There are three remaining contestants competing for the hand of rapper Flavor Flav.  Two of them have been on the show since the first episode, while the other finalist was added midway through in a surprise episode (i.e. to lengthen the highest rated show on VH1).  Contestant ‘Thing 2' has an average of 4.25, while Contestant ‘Sinceer' has an average of 5.5.  It almost seems unfair to compare them to Contestant ‘Black' who most recently joined the show, but her callout average is 2.2, which is impressive.  While technically Black should win the program based on my theory, given her unique situation in the house as a replacement contestant, I am going to predict Thing 2, having a higher callout average than the other lady who has been with the show from the start.  I'll predict Sinceer for the next elimination, as she has the worst callout average of the three contestants.

High-Low-Win Methodology

For the other two shows in my study, I had to adjust my methodology to fit with the format of the shows.  Instead of an ordered callout system, Top Chef and Step It Up and Dance arrange their contestants in one of four groups: Safe, High, Low, and Winner.  Safe contestants are neither the judges' favorite or least favorite that week, so I assigned a 0 score.  High means that they were favored by the judges, and therefore received a 1, while a Low ranking received a -1.  The overall Winner of the week received a score of 2.  Using this methodology, I am attempting to predict who will be the next eliminated chef or dancer.

Top Chef still has seven remaining contenders for the title.  Their high-low-win averages are as follows.

1. Stephanie: .89
2. Richard: .67
3. Andrew: .56
4. Dale: .56
5. Antonia: .44
6. Lisa: .00
7. Spike: -.22

It is also interesting to note that Stephanie is the sole contestant that has never been simply considered ‘Safe' and therefore received no 0 scores.  She has always been either High or Low.  In addition, there has never been a female winner of the programme, so I am going to predict that she takes the title.  Also based solely on the statistical analysis, I am predicting that Spike is the next eliminated chef.  He is the only chef with more Lows than Highs, as well as being the only chef with no Wins.

I have never actually watched an episode of this Elizabeth Berkley-hosted reality television series, but it is very similar to Top Chef in format.  There are currently five aspiring dancers remaining, with the stats as follows:

1. Cody: .83
2. Janelle: .50
3. Miguel: .33
4. Mochi: .00
5. Nick: -.17

The statistical analysis of this show gives a very clear view of who stands on top, and who is at the bottom, as of the sixth episode.  From looking at the overall grid, this show seems to have a habit of following what the statistics would predict.  Being that he is the only dancer with a negative overall average, I am going to predict Nick is the next contestant sent packing.

I have no clear way of knowing if these statistics will accurately predict the outcome of the next episode, but numerous fans of the shows seem to think so.  Statistical analysis pops up frequently in the forums, and as shown above by the Top Model fansite, some fans take it to a very detailed level.  In fact, this is one of the few uses of Wikipedia's grid system that I see implemented.

[This blog post is considered ‘Safe' this week.]