Posts: 1,065
Threads: 80
Joined: Aug 2017
I attempted to visualize this by taking the same list to represent the teams by frequency of key contributor absences and including for each the win percentage that year, parenthetically listing the delta between that win percentage and the average win percentage of the four nearest seasons. I want to say up front that the shortcomings of this approach are quite evident to me. I realize team success/failure ebbs and flows based on numerous factors and that single year win percentages are subject to significant variance. I just want a way to visualize the relationship between injury absences and win/loss record in a way that made sense to me, so take this for what it's worth.
1. 2021 Washington - .333 (- .372)
2. 2017 UCLA - .462 (+ .074)
3. 2020 Washington State - .250 (- .375)
4. 2021 Arizona State - .615 (+ .057)
5. 2016 Arizona State - .417 (- .160)
6. 2021 Oregon - .714 (+ 0.33)
7. 2019 USC - .615 (+ .047)
8. 2019 Stanford - .333 (- .223)
9. 2020 Colorado - .667 (+ .271)
9. 2016 Oregon - .333 (- .371)
11. 2016 Utah - .692 (+ .032)
12. 2016 Arizona - .250 (- .308)
13. 2017 Arizona State - .538 (+ .028)
14. 2015 Oregon State - .167 (- .180)
15. 2021 Cal - .417 (- .083)
15. 2016 Oregon State - .333 (+ .125)
15. 2021 Stanford - .250 (- .328)
18. 2017 Oregon State - .083 (- .188)
18. 2016 UCLA - .333 (- .196)
20. 2015 Oregon - .692 (+ .032)
20. 2020 Oregon State - .286 (- .020)
20. 2014 Washington - .571 (- .089)
20. 2014 Washington State - .250 (- .260)
24. 2018 USC - .417 (- .322)
25. 2017 USC - .786 (+ .190)
26. 2019 Cal - .615 (+ .176)
26. 2018 Stanford - .692 (+ .092)
28. 2019 Colorado - .417 (- .012)
28. 2021 USC - .333 (- .311)
30. 2020 Cal - .250 (- .250)
31. 2017 Utah - .538 (- .184)
32. 2021 Arizona - .091 (- .153)
32. 2021 Colorado - .333 (- .119)
32. 2014 Oregon State - .417 (- .023)
32. 2019 Oregon State - .417 (+ .144)
36. 2015 USC - .571 (- .170)
37. 2018 Cal - .538 (+ .075)
37. 2017 Oregon - .538 (- .116)
39. 2018 Colorado - .417 (- .128)
39. 2017 Stanford - .643 (- .030)
39. 2020 USC - .833 (+ .284)
42. 2017 Washington - .769 (+ .084)
43. 2020 Utah - .600 (- .073)
44. 2015 Arizona - .538 (EVEN)
45. 2018 Utah - .643 (- .024)
46. 2017 Colorado - .417 (- .054)
47. 2017 Arizona - .538 (+ .150)
47. 2015 Colorado - .308 (- .112)
47. 2014 Utah - .692 (+ .112)
50. 2018 Oregon State - .167 (- .112)
50. 2015 Utah - .769 (+ .181)
50. 2019 Washington - .615 (- .013)
53. 2016 Stanford - .769 (+ .065)
54. 2018 Washington - .714 (- .036)
55. 2016 Cal - .417 (- .083)
55. 2018 UCLA - .250 (- .136)
57. 2020 UCLA - .429 (EVEN)
57. 2021 Utah - .714 (+ .062)
59. 2014 Cal - .417 (+ .070)
59. 2017 Cal - .417 (+ .132)
61. 2014 Arizona - .714 (+ .204)
62. 2020 Arizona - .000 (- .347)
63. 2019 UCLA - .333 (- .122)
64. 2016 Washington - .857 (+ .209)
65. 2019 Arizona State - .615 (+.057)
65. 2021 Washington State - .538 (- .090)
67. 2018 Arizona - .417 (+ .084)
68. 2014 Oregon - .867 (+ .161)
69. 2018 Arizona State - .538 (+ .060)
70. 2014 Colorado - .167 (- .206)
70. 2020 Washington - .750 (+ .135)
72. 2016 Colorado - .714 (+ .387)
73. 2014 Arizona State - .769 (+ .211)
73. 2015 Cal - .615 (+ .282)
73. 2015 Washington - .538 (- .184)
76. 2014 UCLA - .769 (+ .173)
76. 2015 UCLA - .615 (+ .027)
78. 2019 Oregon - .857 (+ .219)
78. 2020 Oregon - .571 (- .133)
80. 2015 Arizona State - .462 (- .153)
81. 2020 Stanford - .667 (+ .177)
82. 2018 Oregon - .692 (+ .105)
82. 2021 Oregon State - .538 (+ .305)
82. 2014 USC - .692 (+ .044)
85. 2019 Arizona - .333 (+ .023)
86. 2016 USC - .769 (+ .144)
86. 2015 Washington State - .692 (+ .182)
88. 2020 Arizona State - .500 (- .077)
89. 2017 Washington State - .692 (+ .038)
90. 2019 Utah - .786 (+ .156)
91. 2015 Stanford - .857 (+ .153)
92. 2014 Stanford - .615 (- .203)
92. 2021 UCLA - .667 (+ .303)
94. 2016 Washington State - .615 (- .012)
95. 2019 Washington State - .462 (- .189)
96. 2018 Washington State - .846 (+ .288)
Somebody better than me at statistics can perhaps let me know if this data shows anything interesting, but just eye balling it I can see that 61 percent of the teams in the bottom half of injury luck (the first 49 listed) had negative numbers (i.e., worst seasons than the program's average of that era) and 64 percent of the teams in the top half of injury luck (i.e., the second 47 listed) had positive numbers (i.e., better seasons than the program's average of that era). To me that suggests that a) injury luck is not determinative for a season but b) it's generally better to not have injuries than to have injuries. I realize these are exceedingly obvious observations, but I wanted to be empirical about this.
To further test this, I divided the teams into groups to see if things looked any different if I tried to look at it more in terms of teams with rotten, average, or good injury luck. What I found:
Rotten luck (teams 1-31): 61 percent had negative seasons; median was a - .089 effect
Average luck (teams 32-64): 55 percent had negative seasons, median was a - .023
Good luck (teams 65-96): 28 percent had negative seasons a + .084/+ /.105
This is interesting, if I do say so myself. The results seem so intuitive.
The median impact of rotten luck, negative 8.9 percent compared to a team's norm for that era, is very close to a one game difference ( one game is 8.3 percent of a 12 game season).
The median impact of average luck, negative 2.3 percent, is arguably negligible. That's about a quarter of one game.
The median impact of good luck, positive 8.4 percent or positive 10.1 percent, is in the ballpark of a one game difference. Perhaps others will have views on whether the positive impact of good luck being a bit bigger than the negative impact of rotten luck is significant or if it's just an issue of my flawed sample size/methodology/etc.
As a general, back of the envelope, Matt is not a statistician approximation, a team with rotten luck could generally be expected for that to cost them about one game. IMHO, this corroborates the gut instinct of most on this board, that poor injury luck in 2019 and 2021 explains part of those seasons but is a far cry from justifying those failures. Those teams were sub-par teams regardless of the injuries. Injuries may have made them 4 and 3 win seasons, but healthy they likely would have been 4-6 win seasons. Those weren't good teams irrespective of injuries.
Forecasting ahead to 2022, our four year program average is a .465 winning percentage. In other words, our current sea level as a program is an expectation of 5.58 wins. A 5-6 win regular season even if you give us the benefit of the feels-very-long-ago 2018 season. If we have rotten injury luck again, one might peg us for 4-5 wins. If we can actually benefit from some good injury luck, one might peg us for 6-7 wins. I actually don't think those win totals necessarily tell different things about the health of the program, but making a bowl would mean a lot to me and to the players so that's what I'll be hoping for this year. Six wins or bust!