|
|
NFL Play by Play Angles
Nathan, the new manager of the Investor subscriber area is conducting major research projects in a number of handicapping approaches. Here he investigates how to use the play by play ratings to find situational angles...
|
Play by Play ratings: worth looking at?
There was a lot of hype around the TwoMinuteWarning play by play ratings when they first came out back in the late nineties, with the assumption that having the extra level of granular data would lead to spectacular results. Sadly the spread beating power of the PBP stats has often remained elusive. Yes there have been some good years for the PBP predicted scores, but as often it seems like followed by a bad or mediocre one.
In researching the use of these numbers I have to admit I almost gave up at one point. Trying to create a predicted score algorithm that is powerful across the different weeks of the season is tough going, at least using the season to date stats as I was. Maybe a "last five weeks" look can work better.
Eventually I stopped trying to find a one size fits all approach and went the other direction: searching for angles. Situations, patterns where the spread results were abnormal.
Sample Play By Play Stats matchup grid
| Arizona |
Rush
PSR |
Rush
Big |
Rush
Yds |
Rush
EFR |
Pass
PSR |
Pass
Big |
Pass
Yds |
Pass
EFR |
| ARI - Offense |
44% |
5% |
3.5 |
70 |
47% |
8% |
6.3 |
96 |
| NO - Defense |
-4% |
+0% |
-.1 |
-12 |
+0% |
+3% |
+1 |
+16 |
| ARI - Projection |
40% |
5% |
3.4 |
58 |
47% |
11% |
7.3 |
112 |
| New Orleans |
Rush
PSR |
Rush
Big |
Rush
Yds |
Rush
EFR |
Pass
PSR |
Pass
Big |
Pass
Yds |
Pass
EFR |
| NO - Offense |
45% |
8% |
3.5 |
73 |
52% |
7% |
6.3 |
106 |
| ARI - Defense |
-3% |
-1% |
-.4 |
-7 |
+0% |
+1% |
-.1 |
-3 |
| NO - Projection |
42% |
7% |
3.1 |
66 |
52% |
8% |
6.2 |
103 |
This is the basic layout used at TMW for a game specific play by play ratings view. There are also tables showing all the teams at once. There are four main ratings (play success rate, big gain%, adjusted yds/play, effectiveness rating) kept for rushing and passing. So eight total individual stats on offense, eight on defense.
My first run through looked at how teams have done against the spread when they had a stat that was above or below average. Note in cases where both teams had a qualifying stat the game was ignored.
PBP Angles Search using 2002-2007 NFL Seasons, weeks 4-17
| Type |
Pbp Stat |
Stat range |
Cover% |
Notes |
Stat range |
Cover% |
Notes |
| Off. |
Rush PSR |
50%+ |
52% |
65-43 at 54%+ |
<40% |
47% |
away teams 117-151, 43.6% |
| Off. |
Rush Big |
13%+ |
50% |
dogs 134-102 56% |
<8% |
50% |
|
| Off. |
Rush Yds |
4.5+ |
52% |
dogs 87-58 (60%)…59-28 3.5+ pts |
<3.5 |
47% |
|
| Off. |
Rush EFR |
85+ |
52% |
|
<65 |
47% |
|
| Off. |
Pass Psr |
50%+ |
51% |
|
<40% |
52% |
|
| Off. |
Pass Big |
10%+ |
50% |
dogs 75-54 (58%) |
<6% |
53% |
away dogs 57-38 |
| Off. |
Pass Yds |
6.5+ |
50% |
|
<5 |
53% |
|
| Off. |
Pass Efr |
110+ |
51% |
|
<80 |
54% |
|
| |
|
|
|
|
|
|
|
| Def. |
Rush PSR |
50%+ |
54% |
53%+ = 65-37 |
<40% |
52% |
|
| Def. |
Rush Big |
13%+ |
48% |
|
<8% |
50% |
|
| Def. |
Rush Yds |
4.5+ |
50% |
|
<3.5 |
51% |
|
| Def. |
Rush EFR |
85+ |
54% |
90+ away dogs 70-49 |
<65 |
50% |
33-53 <55, 0-7 pt faves 10-27 |
| Def. |
Pass Psr |
50%+ |
58% |
152-109 overall |
<40% |
51% |
|
| Def. |
Pass Big |
10%+ |
48% |
home faves 41-59 |
<6% |
44% |
dogs 35-49 |
| Def. |
Pass Yds |
6.5+ |
53% |
55% at 6.7+ |
<5 |
48% |
|
| Def. |
Pass Efr |
110+ |
54% |
7.5+ dogs 60%, 120+ 65-39 |
<80 |
46% |
<70 = 68-94 (42%), 18-32 dogs |
| |
|
|
|
|
|
|
|
| Proj. |
Rush PSR |
50%+ |
54% |
197-166, good with lines >3 pts |
<40% |
48% |
|
| Proj. |
Rush Big |
13%+ |
50% |
3.5+ dogs 87-65 (57%) |
<8% |
45% |
faves 62-87 (42%) |
| Proj. |
Rush Yds |
4.5+ |
50% |
3.5+ dogs 60-37 |
<3.5 |
47% |
faves<3.2 yds 74-99, 32-51 lay 3.5+ |
| Proj. |
Rush EFR |
85+ |
50% |
|
<65 |
49% |
|
| Proj. |
Pass Psr |
50%+ |
44% |
149-189 overall, but 62-61 54%+ |
<40% |
52% |
|
| Proj. |
Pass Big |
10%+ |
48% |
|
<6% |
51% |
66-37 3% or lower |
| Proj. |
Pass Yds |
6.5+ |
49% |
|
<5 |
53% |
<4 57% |
| Proj. |
Pass Efr |
110+ |
49% |
|
<80 |
53% |
<60 home 81-38 (68%) |
Some situations to intrigue. My list of angles is:
- Play 54%+ Offense Rush Psr teams (65-43)
- Play 4.5+ yds/rush offense underdogs getting more than a field goal (59-28)
- Play defensive Rush PSR 53%+ teams (65-37)
- Play defensive Pass PSR 50%+ teams (152-109)
- Play defensive Pass EFR 120+ teams (65-39)
- Play home teams projected under 60 Passing EFR (81-38)
Six angles. Only two have an extra filter. Combined records of 487-294 (62%) or roughly 80-50 per season.
Now I know there's no way these angles will hit 62% going forward. But it is reasonable to hope for a solid 55% in the coming season.
The other thing you can do is start doubling up on angles. For example when you have a team with angles 1 & 4 going (Rush Offense Psr 54%+ and Pass Defense Psr 50%+) it's been 25-9.
There will be a new "PBP Angles" weekly column in the
Investor Subscription for NFL 2008 so my research can be put to the test.
Find out more:
What the press says about TwoMinuteWarning
Information on subscribing
Frequently Asked Questions
|
