The (Statistical) State of Utah’s Program – Part 3/4: Television Viewership

[krindor]

OK, so Utah has been very good on the field over the last decade… maybe even better than you thought.  But how about drawing a TV audience?  What sort of viewership has the team drawn

Truth be told, Utah’s not been at quite the same level there (as compared to on-field performance), so we really need fans to step up.  It’s not been bad either, but not great.  

An important caveat here is that rather than try to account for which channels and windows a team gets, I operate under the assumption that the channels/slots selected for them are also a reflection of their value.  Yeah, Tulsa could get a million viewers on Fox at noon… but Fox would never put them there since others could get so much more.  For more methodolody see the complicated notes at the bottom

So there’s a metric I used called viewership score.  Lots of sites will give you the average viewership (counting games without ratings as zeros…which isn’t really fair).  Others will just ignore those (which also isn’t fair since Utah probably doesn’t draw the same against Oregon St as they do against Oregon).  Viewership score strikes a balance between these and also prevents averages from being skewed too much by huge games against marquee opponents.

With that out of the way, here are the viewership scores of every team over the last 5 years (regular season games only).  Utah is #39 by this metric (#26 if postseason games are included; Utah probably gains more than any other team by including those)

1. Alabama 5.59
2. Ohio St 5.51
3. Michigan 4.47
4. Georgia 3.99
5. LSU 3.71
6. Notre Dame 3.35
7. Clemson 3.27
8. Auburn 3.26
9. Penn St 3.22
10. Oklahoma 3.05
11. Florida 2.85
12. Texas 2.67
13. Texas A&M 2.54
14. Tennessee 2.49
15. Wisconsin 2.36
16. Michigan St 2.33
17. Oregon 2.18
    SEC Average 2.10
18. Nebraska 2.03
19. USC 1.82
20. Ole Miss 1.72
21. Florida St 1.71
22. Arkansas 1.66
    B1G Average 1.64
23. Oklahoma St 1.61
24. Iowa 1.53
25. Indiana 1.49
26. Washington 1.48
27. Mississippi St 1.38
28. Miami 1.36
29. TCU 1.35
30. UCLA 1.31
31. South Carolina 1.30
32. West Virginia 1.27
33. Northwestern 1.25
34. Kentucky 1.23
35. Minnesota 1.21
36. Maryland 1.15
37. North Carolina 1.13
38. Stanford 1.10
39. Utah 1.10
40. Washington St 1.10
41. Iowa St 1.09
42. Purdue 1.05
43. Baylor 1.01
44. Kansas St 1.00
45. Louisville 0.99
46. Texas Tech 0.97
47. Virginia Tech 0.94
48. Missouri 0.94
49. Cincinnati 0.92
50. Arizona St 0.91
51. Pitt 0.91
52. Georgia Tech 0.89
53. Colorado 0.87
    PAC Average 0.87
54. BYU 0.87
55. Syracuse 0.86
    Big 12 Average 0.86
    ACC Average 0.80
56. UCF 0.79
57. NC State 0.76
58. Arizona 0.75
59. Boston College 0.70
60. Wake Forest 0.67
61. Virginia 0.66
62. Illinois 0.66
63. Cal 0.65
64. Oregon St 0.58
65. Kansas 0.56
66. Houston 0.56
67. Boise St 0.52
68. Memphis 0.47
69. USF 0.44
70. Duke 0.42
71. SMU 0.36
72. Vanderbilt 0.32
73. Rutgers 0.29
74. San Diego St 0.22
75. Fresno St 0.21
76. UTSA 0.05
77. UNLV 0.04

*Navy and Army are not included since the Army/Navy game and their lack of other scheduled games skews results so much

So, #39 definitely looks bad and it could certainly be better… But it’s not as bad as it looks.  If you look at only teams outside the (future) SEC/B1G, it puts Utah at #12

1. Notre Dame
2. Clemson
3. Oregon
4. Florida St
5. Oklahoma St
6. Washington
7. Miami
8. TCU
9. West Virginia
10. North Carolina
11. Stanford
12. Utah

It also puts Utah as #6 in the Western US (behind Oregon, USC, Washington, UCLA, and barely Stanford)

If Utah could rise up to TCU or West Virginia levels, it would make them a MUCH more appealling candidate

[krindor]

One of the really cool things is that now we’ve established metrics of success on two axes – winning football games and drawing TV viewers.  This allows us to graph every P5 program (and the G5s I looked at) on a chart with those two axes.

No picture this time because this chart works better if you can hover over it

(Note that the TV viewership score is scaled to put Alabama at 100% and everyone else as a percentage of that, but it’s the same numbers you saw here)

https://docs.google.com/spreadsheets/d/e/2PACX-1vQpjfzGL4aQZcNcAYRwLduNpS_rGQiNRFKbUkZAXLGzne9d2nBEZpuNdJTYwXFHRVewO2ODto88WJ_p/pubchart?oid=1287750268&format=interactive

Or zoomed in, to better see the main mass of teams at this link

https://docs.google.com/spreadsheets/d/e/2PACX-1vQpjfzGL4aQZcNcAYRwLduNpS_rGQiNRFKbUkZAXLGzne9d2nBEZpuNdJTYwXFHRVewO2ODto88WJ_p/pubchart?oid=1680908327&format=interactive

This will lead into the next and final installment, where we consider every team’s value as an expansion candidate – including Utah to the B1G (or Big 12) as well as which candidates the PAC might be looking at

[Yergensen]

Good stuff. You’re pretty good at this. You could quit your day job.

[wcrunner]

krindor

What does Vscore xx% mean?

[krindor]

VScore is just what I type when I get tired of writing viewership score over and over. And rather than a number, I like to scale it so the top program is 100% and everyone else is a factor of that. Just because viewership now is more than 10 years ago, but it’s all relative. So that’s why I use the V-Score %.

As to calculating viewership score, it’s really pretty simple.

Lots of people take the average viewership, but they generally either ignore games without TV ratings (not fair since those are generally weaker games) or assign them as zeros (also not fair). The other issue is that a huge game or two can skew the average a lot. So I take a few steps to mitigate those factors.

Step 1: Find the average viewership of all televised/rated games for a team (ignoring unrated ones for now).

Step 2: Find the median viewership of all televised/rated games for a team

Step 3: Take the harmonic mean of those values. The harmonic mean is 2*x1*x2/(x1+x2) and has the effect of more heavily weighting whichever value is lower. This way, if your average is way above the median, that effect will be minimized

Step 4: Find the average viewership of ALL games (now including unrated games as zeros)

Step 5: Take the harmonic mean of the values found in Step 3 and in Step 4.

This means that teams who aren’t on TV often will definitely be punished (which makes sense since we’re looking at how much exposure they get/drive) and is another step to minimize the impact of any outlier games.

Step 6 (if desired): Scale the score such that the top performer is 100%

[DrJazzy]

Fantastic analysis, thanks. Keep them coming!!