DWTS Season 20 Assessment of Historical Strength

The following is a historically-based analysis of the competitive strength of Dancing with the Stars Season 20 celebrities.  This is not meant as a prediction, but as a simple statistical assessment for anyone interested in statistics.  I am posting just the comparison groups and overall results.  The analysis is similar to what I have posted for past seasons, with changes noted.

First, here's the Excel spreadsheet showing the historical comparison group for each Season 20 celebrity. The group headings explain how the groups were defined.  Red text indicates discretionary inclusion of celebrities who do not fit the group definition, but are otherwise comparable.  Red can also indicate guesstimated data, wherever necessary.  The groups are listed from strongest (highest average historical results) to weakest (lowest average historical results):



Now here is a summary of the the Age-Job historical averages, clustered by comparable historical strength, from strongest to weakest.

Usually, I would present height and pro data separately, but I'll skip to the final assessment.  The height and pro data are included here.  Red text indicates pro averages based on fewer than three official results (bearing in mind that Emma's 2nd partner withdrew).  These newer pro averages are less meaningful than more established pro averages.  *Technical note: The factor weighting has changed to de-emphasize height, as explained below.

Personal comment:  I'll repeat my usual point that this historical analysis is not meant to be a prediction.  Nevertheless, these results look similar to many predictions I've seen.  The one pre-season surprise for me was realizing that at 39, Emma's "young" partner RedFoo is neither historically young, nor relatively young among Season 20 males.  Also, I think the relative results between Riker and Chris could depend heavily on their still unproven partners.

-----------------------------------------------------------------------------------------------------------------

*Please note that height was also included in the definition of some comparison groups.  For example, I limited young female athletes to those within 5 inches of Nastia's height.  This fine-tuning of comparison groups becomes increasingly possible as more data becomes available across more seasons.  Consequently the weighting of height was de-emphasized from past season analyses.  The new weighting is:  (1.75*Age-JobAvg + ProAvg + 0.25*HtAvg) / 3.