DWTS 11 Stat Class: Who has the Edge? Part 1
Yay! My turn. 😀
Heidi and Courtney (Part I and Part II) turned in their feelings of how the power ranking would work for the season. While John wrote about odds/competitiveness factor as well as number of twitter followers for each celeb.
If you were following last season, I tried to predict scores for the stars before they actually started dancing. Well here I am again, using a more “advanced” statistical analysis to try to predict where each celeb will ultimately end up.
These calculations are based on three factors:
– Age Group: Taking the average of all the placements of all past celebrities and breaking them down by age group (teens, 20s, 30s, 40s, etc), it came to pass that those in their teens through 30s tend to do better than those in their 40s on up. As in on average three placements better. That being said, the celebs (on average) place lower in accordance with their age… with one notable exception: those in their 60s tend to place higher than those in their 50s.
– Occupation: Again taking the average of all the placements of all past celebrities and breaking them down by occupation. I know there are a lot of celebs that work in multiple mediums, however, I am focusing on the medium that they are best know for. David Hasselhoff is known for Baywatch thus he’s known as an actor. Brandy is best known for her music thus she’s in the singer category. And yes Reality stars have their own category and those like Bristol Palin are found in the “other” category.
– Professional Dancer: And yes, I even went so far as to calculate the average placement of all the professional partners through all the seasons. I tried to normalize the result as much as possible, but obviously as much as I try to make this a scientific calculation, it is only half the battle.
After averaging all three of the above factors out I come up with a fairly decent representation of how the celebs could be placed in competition. Granted things will change with every passing season, but we’ll see how things go.
In Part II of the DWTS Stat Class: Who has the Edge? I will present the above three factors in action and give a bit of an analysis of each factor as well as my thoughts, and then follow that with a final averaged placement.
So for your reading enjoyment… I will give you the ranking of the pros (including Alec, Edyta, and Julianne for nostalgia’s sake) in order of ranking from highest to lowest.
The professional dancers (dancers participating on Season 11 are in bold):
1.) Derek Hough – 6 Seasons
2.) Julianne Hough – 5 Seasons
3.) Cheryl Burke – 9 Seasons
4.) Lacey Schwimmer – 3 Seasons
5.) Kym Johnson – 7 Seasons
6.) Chelsie Hightower – 3 Seasons
7.) Dmitry Chaplin – 2 Seasons
8.) Karina Smirnoff – 7 Seasons
9.) Louis van Amstel – 6 Seasons
10.) Mark Ballas – 6 Seasons
11.) Tony Dovolani – 9 Seasons
12.) Corky Ballas – 1 Season
13.) Maksim Chmerkovskiy – 8 Seasons
14.) Anna Trebunskaya – 5 Seasons
15.) Edyta Sliwinska – 10 Seasons
16.) Alec Mazo – 5 Seasons
17.) Ashly Costa (nee DelGrosso) – 4 Seasons
18.) Jonathan Roberts – 7 Seasons
Way cool Mairianya. I love this!!! I can’t wait to see how your analysis turns out in part 2.
Whoa, haven’t been on here in a while! While I def agree with these pro rankings, what exactly did you guys base them on?
I’ve seen the math, Hotaru, and it holds up. Of course there is built in bias, but Marianya normalized it as best she could. When you have a pro like Dmitry with two seasons, one of which is a 2nd place, going against someone who’s been around a lot longer and has a several crappy partners (read: Edyta), of course the comparison is going to be biased in Dmitry’s favor. The bias is reduced when comparing pros that have a closer number of seasons in the competition.
I’m surprised you put Maks is so low.
@Vogue – I think the reason Maks is so low is due to a few early elims with bigger casts – Misty withdrew early, Denise was sent home the 2nd week of her season, & Debi was one of the earlier elims on the season with the biggest cast. It definitely skews the results, and there’s unfortunately not a good way to normalize it. It kinda screws Mark too – being the first slim last season and the 3rd elim in season 7 w/Kim likely dragged down his average quite a bit…and he’s won twice! And it kinda works in the opposite way for Dmitry & Lacey – he’s only been on the show 2 seasons, one of which he had a very early elim and one where he got 2nd place (which also happened to be the season with the biggest cast), so it makes him look really good; Lacey made it to the finale on the 2nd largest season, and decently far on the largest season, but got sent home relatively early in a smaller season. Sometimes I wish we had a professional statistician to consult on these things – it’d be so much easier to give airtight stats!
Courtney – No kidding… I’ve taken stats classes but dang… I tried normalizing in one way but it skews the results more, Louis was higher up the ladder than in this one…
In this version, I set all seasons to a standard scale (which is the average number of contestants from all seasons… 11.5) so where ever they placed in their season it’s standardized to that 11.5 scale.
That’s the best I could do with what limited knowledge of stats I have… ah well.
Thanks for explaining Courtney. I see what you mean on Maks now.
Fascinating post truly!!
There is NO way to do airtight stats on a situation that has so much flux and variability in it. I doubt seriously that a statistician could do it. I’m a data person and I could find all sorts of bias that just can’t be eliminated. We just do the best we can. 🙂
Like a lot of areas of science, you just have to put all the qualifiers in place – acknowledge that some pros consistently get crappy partners – Edyta and Jonathan. Then there are some seasons that just have a lower overall capability of celebs, versus a season that had a lot of very capable celebs. Then there’s the fanbase issue.
Another way I thought about it was to count how many couples each pro beat in each season. For instance, Kym beat 15 couples to win season 9 while Alex only beat 5 couples for his season 1 win. Same theme of more credit for bigger seasons but with absolute numbers instead of relative (scaled) ones. Similar order too, but interestingly not identical. Obviously I’m a numbers nerd and absolutely love this sort of thing – probably should have gone into insurance or accounting instead of my chosen field. 😀
Derek (beat an average of 9.3 couples/season competing)
Lacey
Julianne & Dmitry
Cheryl
Kym & Chelsie
Mark
Karina
Louis & Corky
Tony
Maks
Anna
Edyta
Alec
Jonathon
Ashly (beat an average of 2.3 couples/season competing)
Both of your ranking method are interesting 🙂
I like this. Very Interesting. Of course it doesn’t show the whole pic…still.