Part III -- Preseason Offense/Defense
A prior analysis here was based around a team's won-lost record in the previous year and how that effected preseason results. This time out we'll do something similar, but focus on the a team's rankings on offense and defense from the prior year and see if we can find some areas of historical value.
The first decision is what to use for ranking each team on offense and defense. Conventionally people have focused on yards, but we are electing to use points scored and points allowed, as although these have some level of distortion (eg special teams and defensive points scored), the general standings will better reflect perhaps the mindset of the team. For instance a team that scored a lot of points while ranking only average in yards probably doesn't view their offense as in need of that much improvement, whereas a team that rated high in yards but lower in points probably does think that they need to be able to score more effectively.
Using the total points for and against then, we rank each team to the rest of the league and consider a top 10 ranking to be "GOOD" a rank of 11 to 20 to be "AVERAGE" and a rank of 21 and higher to be "POOR". As we have two numbers for each team (offense and defense) there are many ways to look at this -- the following table reflects some of the areas of promise for predictive purposes:
Home
Team |
Away
Team |
Situation |
Won
(vs Spr) |
Lost
(vs Spr) |
Home
Win% |
| Poor offense |
Ave/Poor offense |
Home favored by 3+ |
16 |
32 |
33 % |
| Average offense |
Ave/Poor offense |
Week 1 |
3 |
16 |
16 % |
| Good offense |
Ave/Poor offense |
Week 2 |
5 |
15 |
25 % |
| Good defense |
Any |
|
42 |
65 |
39 % |
| Good Defense |
Poor Defense |
Home favored by 3+ |
8 |
19 |
29 % |
| Average Defense |
Good Defense |
|
28 |
13 |
68 % |
| Ave/Poor Offense |
Poor Defense |
|
38 |
63 |
38 % |
| Average Offense |
Good/Ave Defense |
Home favored by 3+ |
23 |
14 |
62 % |
| Good Offense |
Ave/Poor Defense |
Weeks 0-2 |
10 |
25 |
29 % |
| Poor Offense |
Ave/Poor Defense |
Weeks 3+ |
15 |
35 |
30 % |
| Good Defense |
Ave/Poor Offense |
|
28 |
51 |
35 % |
Still, these angles seem limited since they are at best considering only half of the story (eg ignoring either a team's offense or defense). Another way to look at it is to rate a team from 1 to 3 (good is 1, average is 2, poor is 3) and add the offense and defense together.
This way you end up with a team ranked from 2 (Best, meaning top 10 in both points scored and points allowed) to 6 (worst, meaning ranking 21+ in both areas). Running some queries on this, with the expectation that the two extremes (the "2" teams and "6" teams) will most likely have the unusual results, here's what we find:
Home
Rtg |
Away
Rtg |
Won
(vs Spread) |
Lost
(vs Spread) |
Home
Win% |
| 2 |
3+ |
15 |
28 |
35 % |
| 2 to 5 |
6 |
18 |
41 |
31 % |
The above seem to have a better chance to continue the pattern than some angles -- it makes sense that the worst teams from the prior year will probably try a little harder in preseason to win, and on the road such a team can play without the often negative home crowd experience (plus they will usually be underdogs getting a field goal or more). Likewise the top NFL teams of the previous season will have little need to show their stuff in the exhibitions and so can be fodder for other teams, particularly when the top squad is at home and more often than not, favored.
All well and good then, we've discovered a handful of useful angles that in conjunction with some other handicapping tools might plausibly give us the edge in the preseason.
| 
