Are the old sayings true?

dwmyers

Well-Known Member
Messages
2,373
Reaction score
522
I've heard, from my late teens, that the best predictor of who will win a football game is a team's record. I've been curious about that, how to test that proposition. For now, a simple check is to compare teams before they play, count the number of wins each team has, find the difference, and if it is more than 0 (I assume 0 would predict a tie), then figure out if the "better" team won.

Anyway, I have 3.5 years of data so far. It's on the chart in the window below:

Code:
2007

WinDiff Predicted   Pct HomeWins VisWins  TotGames
0   :        0      0.0    12      11      23
1   :       15     62.5    15       9      24
2   :       10     66.7     9       6      15
3   :        5     62.5     5       3       8
6   :        1    100.0     0       1       1

2006

WinDiff Predicted   Pct HomeWins VisWins  TotGames
0   :        0      0.0    11      22      33
1   :       38     52.8    39      33      72
2   :       30     65.2    27      19      46
3   :       19     55.9    17      17      34
4   :       13     72.2    11       7      18
5   :        7     77.8     6       3       9
6   :        3     50.0     4       2       6
7   :        1     33.3     2       1       3
8   :        2    100.0     2       0       2
10  :        1    100.0     0       1       1


2005

WinDiff Predicted   Pct HomeWins VisWins  TotGames
0   :        0      0.0    26      20      46
1   :       32     57.1    37      19      56
2   :       26     55.3    30      17      47
3   :       24     77.4    16      15      31
4   :       14     87.5     6      10      16
5   :        5     62.5     5       3       8
6   :        7    100.0     1       6       7
7   :        7     87.5     5       3       8
8   :        4    100.0     3       1       4
10  :        0      0.0     1       0       1

2004

WinDiff Predicted   Pct HomeWins VisWins  TotGames
0   :        0      0.0    24      22      46
1   :       28     52.8    30      23      53
2   :       35     66.0    31      22      53
3   :       18     64.3    16      12      28
4   :       17     70.8    14      10      24
5   :        8     88.9     6       3       9
6   :        3     75.0     1       3       4
7   :        4     80.0     2       3       5
10  :        0      0.0     1       0       1
11  :        1    100.0     1       0       1


2006 first half

WinDiff Predicted   Pct HomeWins VisWins  TotGames
0   :        0      0.0     7      13      20
1   :       20     48.8    25      16      41
2   :       11     57.9    15       4      19
3   :        8     72.7     8       3      11
4   :        3     75.0     3       1       4
6   :        0      0.0     0       1       1

2006 second half

WinDiff Predicted   Pct HomeWins VisWins  TotGames
0   :        0      0.0     4       9      13
1   :       18     58.1    14      17      31
2   :       19     70.4    12      15      27
3   :       11     47.8     9      14      23
4   :       10     71.4     8       6      14
5   :        7     77.8     6       3       9
6   :        3     60.0     4       1       5
7   :        1     33.3     2       1       3
8   :        2    100.0     2       0       2
10  :        1    100.0     0       1       1

2005 first half 

WinDiff Predicted   Pct HomeWins VisWins  TotGames
0   :        0      0.0    20      13      33
1   :       15     50.0    20      10      30
2   :       13     61.9    14       7      21
3   :        8     88.9     5       4       9
4   :        3     75.0     2       2       4
6   :        1    100.0     0       1       1

2005 second half

WinDiff Predicted   Pct HomeWins VisWins  TotGames
0   :        0      0.0     6       7      13
1   :       17     65.4    17       9      26
2   :       13     50.0    16      10      26
3   :       16     72.7    11      11      22
4   :       11     91.7     4       8      12
5   :        5     62.5     5       3       8
6   :        6    100.0     1       5       6
7   :        7     87.5     5       3       8
8   :        4    100.0     3       1       4
10  :        0      0.0     1       0       1

2004 first half

WinDiff Predicted   Pct HomeWins VisWins  TotGames
0   :        0      0.0    12      12      24
1   :       19     59.4    18      14      32
2   :       21     67.7    16      15      31
3   :        5     83.3     4       2       6
4   :        1     20.0     5       0       5


2004 second half

WinDiff Predicted   Pct HomeWins VisWins  TotGames
0   :        0      0.0    12      10      22
1   :        9     42.9    12       9      21
2   :       14     63.6    15       7      22
3   :       13     59.1    12      10      22
4   :       16     84.2     9      10      19
5   :        8     88.9     6       3       9
6   :        3     75.0     1       3       4
7   :        4     80.0     2       3       5
10  :        0      0.0     1       0       1
11  :        1    100.0     1       0       1
Couple things. 2 or more wins is pretty significant, 1 win less so. I was thinking that 1 win didn't mean as much during the period of byes, but that after the 10th week or so, when byes are complete, that they would mean more. And if that's true, then the winning percentage for a 1 win advantage would increase in the second half of the season. That's true for 2006 and 2005, but not for 2004. Two out of three so far.

:D

David.
 
Top