DWTS Season 22 - Historical Strength

This analysis is similar to those in previous posts, but more data-driven.  Previously, comparison groups for each star have been hand-tailored around general rules for age and job comparisons.  (For anyone interested, here's a belated pre-Season 21 analysis that includes height in the selection of Age-Job comparison groups.)  As we approach 20+ seasons of usable data, fewer adjustments are needed to compensate for insufficient data.

We'll start with the Season 22 comparison groups for each star.  The "job" categories are borrowed from the defunct "Cast Next Season's Stars" game that used to appear on ABC's website.  The age range is the star's age +/- 5 years.  (Note that a one-year age difference results in slightly different comparison groups for Antonio and Von.)

Historical results were converted to a scale from 1-12, representing 1st place through 12th place in a standard-size DWTS field.  Withdrawals and All-Star results are excluded from this analysis.
The comparison groups are listed from best/highest/strongest average to worst/lowest/weakest average:



Notes:  Age ranges were broadened for Paige, Geraldo, and Marla due to insufficient data.  Wanya's age range was broadened because Master P / Billy Ray Cyrus / RedFoo doesn't seem like a fair comparison for anyone.  (But I thought the same for RedFoo with respect to P and Cyrus, who turned out to be fair comparisons after all.)  Job categories were broadened for Geraldo due to insufficient data, and for Nyle, due to insufficient data and inadequate comparison.

Before proceeding, here's a tabular summary of the group averages from the above spreadsheet, arranged in ascending clusters of comparable strength:

Next is a summary of Age-Job, Height, and Pro averages for each Season 22 star.  The count is the number of comparison data used to calculate the group average.  For Age-Job, these match the spreadsheet above.  Averages based on expanded Age-Job definitions (as discussed above) or limited pro partners are highlighted in red.

Finally, here are the overall weighted averages*.  They are arranged from strongest to weakest average, in clusters of comparable strength.  Wildcards are highlighted in red, indicating that the corresponding averages are based on just one or two data points, making them less reliable than other averages. Also marked as wildcards are Keo, whose first 2 or 3 seasons were probably in lieu of troupe, and Nyle, who is difficult to compare to previous stars within the limit of this analysis.



*Technical Note:  The factors were weighted as follows: 50% Age-Job average, 30% Pro average, 20% Height average.  These weights were based on correlation analysis between weighted historical averages and actual historical results.

Personal Comments:

First, a reminder that these are not predictions.  Inevitably, some stars will beat their historical averages and some will fall short.  Considering that all 12 stars fall within a range of 4.5 to 7.9, i.e. within 4th-5th place to 8th place, this truly is inevitable.

For me, the most interesting part of these analyses are the potential surprises.  Young actresses in Mischa's age range have done well in the past.  Her historical average looks surprisingly high, while actresses in Jodie's age range have averaged a bit lower, and she is further dragged down by Keo's poor results.  Offhand, I would guess that Jodie and Keo are a better bet for beating their historical 8th place than Mischa and Artem are for matching their historical 5th place, but it may be a closer battle than many expect.  Another good bet to beat their average may be Wanya and Lindsay, who seem popular in pre-season fan discussions.

Two final comments:  Although Paige's historical strength isn't necessarily a surprise, her partner Mark has his own history of surprise eliminations--well known to DWTS fans--that seems worth mentioning.  (Hopefully they'll avoid that fate.)  Finally, whatever Nyle achieves as DWTS' first fully deaf star will be unprecedented.  At some point, I may include a "human interest" factor in these analyses.  Without such numbers, I'll simply acknowledge a general consensus among fans that "human interest" can generate votes leading to better-than-expected results.