Spinning off of from KingJames post here I wanted to look at Fuente's performance against expected wins based on the Vegas Spread. By expected wins, I mean that a 14 pt favorite is expected win 85.1% of the time, a 7 pt favorite 70.3% of the time, and so on. Data sources and code are at the bottom, but betting line data was only available back to 2013 so I started from there.
Comparing the coaches overall:
| Coach | wins | losses | Expected Wins | win_perc | wins vs expectation |
|---|---|---|---|---|---|
| Beamer 2013 - 2015 | 19.0 | 16.0 | 21.21 | 0.54 | -2.21 |
| Fuente 2016-2019 | 32.0 | 17.0 | 31.17 | 0.65 | 0.83 |
Fuente actually exceeded Vegas expectations overall slightly in his first 4 years, but how does it look when we are favorites or dogs?
Comparing the coaches across types of games:
| Coach | spread_group | wins | losses | Expected Wins | win_perc | wins vs expectation |
|---|---|---|---|---|---|---|
| Beamer 2013 - 2015 | 1. 14+ Favorites | 4.0 | 2.0 | 5.72 | 0.67 | -1.72 |
| Beamer 2013 - 2015 | 2. 7-14 Favorites | 5.0 | 2.0 | 5.43 | 0.71 | -0.43 |
| Beamer 2013 - 2015 | 3. 0.5-6.5 Favorites | 5.0 | 8.0 | 7.84 | 0.38 | -2.84 |
| Beamer 2013 - 2015 | 5. 0.5-6.5 Dogs | 2.0 | 0.0 | 0.68 | 1.0 | 1.32 |
| Beamer 2013 - 2015 | 6. 7-14 Dogs | 3.0 | 2.0 | 1.34 | 0.6 | 1.66 |
| Beamer 2013 - 2015 | 7. 14+ Dogs | 0.0 | 2.0 | 0.19 | 0.0 | -0.19 |
| Fuente 2016-2019 | 1. 14+ Favorites | 13.0 | 3.0 | 15.4 | 0.81 | -2.40 |
| Fuente 2016-2019 | 2. 7-14 Favorites | 4.0 | 0.0 | 2.94 | 1.0 | 1.06 |
| Fuente 2016-2019 | 3. 0.5-6.5 Favorites | 7.0 | 5.0 | 7.28 | 0.58 | -0.28 |
| Fuente 2016-2019 | 5. 0.5-6.5 Dogs | 6.0 | 5.0 | 4.25 | 0.55 | 1.75 |
| Fuente 2016-2019 | 6. 7-14 Dogs | 1.0 | 3.0 | 1.1 | 0.25 | -0.10 |
| Fuente 2016-2019 | 7. 14+ Dogs | 1.0 | 1.0 | 0.21 | 0.5 | 0.79 |
Fuente lost 3 games in his first 4 seasons as 14+ favorites, but has done as well or better than expected in other games.
For some reference points, here is UVA:
| spread_group | wins | losses | Expected Wins | win_perc | wins vs expectation |
|---|---|---|---|---|---|
| 1. 14+ Favorites | 11.0 | 1.0 | 11.46 | 0.92 | -0.46000000000000085 |
| 2. 7-14 Favorites | 4.0 | 2.0 | 4.42 | 0.67 | -0.41999999999999993 |
| 3. 0.5-6.5 Favorites | 7.0 | 5.0 | 7.38 | 0.58 | -0.3799999999999999 |
| 5. 0.5-6.5 Dogs | 8.0 | 15.0 | 9.17 | 0.35 | -1.17 |
| 6. 7-14 Dogs | 5.0 | 13.0 | 4.12 | 0.28 | 0.8799999999999999 |
| 7. 14+ Dogs | 0.0 | 12.0 | 0.48 | 0.0 | -0.48 |
Clemson:
| spread_group | wins | losses | Expected Wins | win_perc | wins vs expectation |
|---|---|---|---|---|---|
| 1. 14+ Favorites | 48.0 | 2.0 | 48.91 | 0.96 | -0.9099999999999966 |
| 2. 7-14 Favorites | 14.0 | 0.0 | 10.99 | 1.0 | 3.01 |
| 3. 0.5-6.5 Favorites | 7.0 | 1.0 | 4.8 | 0.88 | 2.2 |
| 5. 0.5-6.5 Dogs | 0.0 | 2.0 | 0.76 | 0.0 | -0.76 |
| 7. 14+ Dogs | 0.0 | 1.0 | 0.0 | 0.0 | 0.0 |
| 7. 14+ Dogs | 0.0 | 12.0 | 0.48 | 0.0 | -0.48 |
Tennessee:
| spread_group | wins | losses | Expected Wins | win_perc | wins vs expectation |
|---|---|---|---|---|---|
| 1. 14+ Favorites | 21.0 | 2.0 | 22.18 | 0.91 | -1.1799999999999997 |
| 2. 7-14 Favorites | 8.0 | 2.0 | 7.63 | 0.8 | 0.3700000000000001 |
| 3. 0.5-6.5 Favorites | 4.0 | 5.0 | 5.27 | 0.44 | -1.2699999999999996 |
| 5. 0.5-6.5 Dogs | 6.0 | 9.0 | 6.0 | 0.4 | 0.0 |
| 6. 7-14 Dogs | 2.0 | 9.0 | 2.47 | 0.18 | -0.4700000000000002 |
| 7. 14+ Dogs | 1.0 | 14.0 | 0.72 | 0.07 | 0.28 |
Source: collegefootballdata.com
Code for those interested: https://github.com/hokiespider/win_probability
Comments
I did the best I could figure out formatting the tables but would love some help to make it better!
Set table style to "blog_statistics"
Thank you for the insightful analysis. I formatted the tables for you.
You're welcome. I thought this analysis would give some useful insight into game/team preparation and in-game coaching. By incorporating expected win percentage based on the spread it should (theoretically at least) normalize for talent levels, home-field advantage, etc. and isolate on game planning and in-game performance.
Would have been interesting to see Beamers whole career if the data was available
I will try to find more data this weekend to make this happen (or at least go back further than 2013. I'll also bring in this season's games (my code broke pulling in 2020 because the API returned unplayed 2020 games with null betting lines - I'm a beginner with Python/programming and couldn't get it working last night.)
One important statistic really jumps out at me. We got our butts kicked by LIBERTY. Very impressive data analysis, otherwise.
Yep and neither this year's Liberty Loss (17 pt favorites) or Wake (10.5 point favorites) are included so it will look worse when those are included.
Great analysis! Also love the GitHub link!
Related stat from Chris Fallica:
2018 and early 2019 were really rough.
2020: "hold my beer"