|
|
Rudy
Moser’s Husker Probability Page UPDATED 11/16/11 |
|
The 2011 Nebraska Cornhusker Football Season
|
Team |
Current Rating |
Home field Adv. |
|
|
|
|
|
|
|
Nebraska |
84.21 |
2.41 |
|
|
|
Results |
|
|
|
|
|
|
Spread |
Probability |
|
Nebraska |
Opponent |
Difference vs. Spread |
|
Chattanooga |
64.03 |
H |
----- |
100.0% |
|
40 |
7 |
----- |
|
Fresno State |
63.13 |
H |
28 |
100.0% |
|
42 |
29 |
-15.0 |
|
Washington |
75.10 |
H |
16.5 |
100.0% |
|
51 |
38 |
-3.5 |
|
Wyoming |
66.12 |
A |
23.5 |
100.0% |
|
38 |
14 |
0.5 |
|
Wisconsin |
92.15 |
A |
-9 |
0.0% |
|
17 |
48 |
-22.0 |
|
Ohio State |
78.30 |
H |
11 |
100.0% |
|
34 |
27 |
-4.0 |
|
Minnesota |
61.83 |
A |
24.5 |
100.0% |
|
41 |
14 |
2.5 |
|
Michigan State |
82.92 |
H |
4 |
100.0% |
|
24 |
3 |
17 |
|
Northwestern |
72.80 |
H |
17.5 |
0.0% |
|
25 |
28 |
-20.5 |
|
Penn State |
81.20 |
A |
3.5 |
100.0% |
|
17 |
14 |
-0.5 |
|
Michigan |
88.16 |
A |
-6.36 |
32.5% |
|
|
|
|
|
Iowa |
74.92 |
H |
11.7 |
79.8% |
|
|
|
|
|
Big Ten (Assume Wisconsin?) |
92.15 |
|
-7.94 |
28.5% |
|
|
|
|
|
Bowl Game |
??? |
|
??? |
??? |
|
|
|
|
|
|
|
|
|
|
Average |
32.9 |
22.2 |
-5.1 |
|
|
|
|
|
|
Std. Dev. |
11.0 |
13.5 |
11.7 |
|
|
|
|
|
Median |
34.0 |
22.2 |
-3.8 |
|
|
|
|
|
|
|
||||
Expected Wins (Regular Season): 9.12
|
Record |
0-12 |
1-11 |
2-10 |
3-9 |
4-8 |
5-7 |
6-6 |
7-5 |
8-4 |
9-3 |
10-2 |
11-1 |
12-0 |
|
Probability |
----- |
----- |
----- |
----- |
----- |
----- |
----- |
----- |
13.62% |
60.45% |
25.93% |
----- |
----- |
Ratings are from Jeff Sagarin 11/13/11 “Predictor” column is used
Week 11 Update
This season I have been collecting data to see if the assumptions of the probability model are reasonable. The key assumption is that the result against the spread follows a Normal distribution. So far I have collected 11 weeks of data for spreads and the margin against the spread for FBS teams. For consistency I have used collected the spreads from Rivals.com’s Mike Huguenin weekly primer. The most recent one for week 11 is here. So far I have collected the results of 536 football games. After the season is over I should be able to come up with appropriate hypothesis tests to determine if the assumptions are appropriate.
Margin Against the Spread for the 2011 College Football Season
2011 - Big 10 Conference Probabilities
|
8-0 |
7-1 |
6-2 |
5-3 |
4-4 |
3-5 |
2-6 |
1-7 |
0-8 |
Expected Wins |
|
|
Legends |
||||||||||
|
Michigan State |
----- |
68.66% |
30.42% |
0.92% |
----- |
----- |
----- |
----- |
----- |
6.68 |
|
Michigan |
----- |
----- |
54.66% |
39.15% |
6.18% |
----- |
----- |
----- |
----- |
5.48 |
|
Nebraska |
----- |
----- |
25.93% |
60.45% |
13.62% |
----- |
----- |
----- |
----- |
5.12 |
|
Iowa |
----- |
----- |
----- |
11.29% |
53.56% |
35.16% |
----- |
----- |
----- |
3.76 |
|
Northwestern |
----- |
----- |
----- |
----- |
24.16% |
63.82% |
12.03% |
----- |
----- |
3.12 |
|
Minnesota |
----- |
----- |
----- |
----- |
----- |
3.36% |
30.06% |
66.58% |
----- |
1.37 |
|
Leaders |
8-0 |
7-1 |
6-2 |
5-3 |
4-4 |
3-5 |
2-6 |
1-7 |
0-8 |
|
|
Penn State |
----- |
8.74% |
50.92% |
40.34% |
----- |
----- |
----- |
----- |
----- |
5.68 |
|
Wisconsin |
----- |
----- |
69.30% |
27.90% |
2.81% |
----- |
----- |
----- |
----- |
5.66 |
|
Illinois |
----- |
----- |
13.24% |
70.21% |
16.55% |
----- |
----- |
----- |
----- |
4.97 |
|
Purdue |
----- |
----- |
----- |
32.23% |
52.77% |
15.00% |
----- |
----- |
----- |
4.17 |
|
Ohio State |
----- |
----- |
----- |
9.25% |
49.14% |
41.61% |
----- |
----- |
----- |
3.68 |
|
Indiana |
----- |
----- |
----- |
----- |
----- |
----- |
0.85% |
28.28% |
70.87% |
0.30 |
Big 10 Week 11
For the record, I still hate the “Legends” and “Leaders”
division names. I also hate it when
people abbreviate Big 10 by typing “B1G”.
I always want to pronounce “B1G” Bi-One-ng. I can’t even figure out the proper
pronunciation of “B1G”.
Methodology
Most of the credit for this page goes to Brad Carlin for giving me the idea. The probability of victory is based on the point spread between the teams. The logic makes fair sense, considering that bookkeepers want the distribution of betting to be even on both sides of the line. Therefore it is a good basis for our analysis. We will have to make some assumptions (albeit dodgy ones) in order to make our predictions:
1. The results of games are predictable.
2. Each game is independent. No one game affects the outcome of other games.
3. Games follow a normal distribution (a bell curve)
4. The mean and std. deviation are constant throughout the season.
5. Jeff Sagarin ratings are a reasonably good way to determine point spread for an upcoming game.
This can easily tell us the probability of any game. Each spread has a certain probability associated with it. All you would have to do then to find the probability of going 12-0 would be to multiply all the individual game percentages together. To find the other record possibilities it was necessary to create a calculator program to add up all the different one loss combinations, two loss combinations, three loss combinations, and so forth.
Email Rudy Moser: rudabega1@hotmail.com