Comparing Speight and O'Korn using 3 UFR-based QB metrics to control for Pass Pro & WR drops
Introduction
MGoBrethren,
Happy Day to all. Beat State. Beat them to an indistinguishable pulp.
With that said, I thought I'd share with all of you still sober (or black out or anxious as hell for anything else Michigan before the game) some of the analysis I've been working on since the Purdue game. I set out to review the UFR in an attempt to see whether or not "Red Zone Offense" is a thing. I ended up compiling a massive chart of all the UFRs plus a couple extra data points collected from rewatching the games. The RZ analysis is coming. But in the meantime, I wanted to share my attempt to create 3 new QB score metrics to best compare Speight and O'Korn on a level playing field.
-------------------------------------------
The QB Score metrics:
I have divided passing plays into two main categories:
- Standard passing downs (1st & >15, 2nd & >8, 3rd & >4)
- Standard running downs (1st & <15, 2nd & 8-, 3rd & 4-)
Then, I am filtering each category based on whether the charted pass protection is negative or possitive/NA (i.e. screen plays).
Finally, I am excluding plays where WRs dropped a catchable ball (CA/DO as charted in the UFR), or where the pass rush resulted in a charted PR.
This creates 3 types of plays:
1) Pure QBS: Passing plays on standard passing downs with non-negative pass protection scores, with WR drops on CA or DO throws excluded.
- This looks at a QB's ability to make plays on passing downs where the D is expecting pass, and the OL withstands the pass rush.
2) Explosive QBS: Passing plays on standard running downs with non-negative pass protection scores, with WR drops on CA or DO throws excluded.
- This looks at a QB's ability to make plays on theoretically unexpected passing downs, where the D may be caught off guard by the pass, and the OL provides time to deliver a throw.
3) Playmaking QBS: Passing plays on any down with negative pass protection scores, with plays killed due to PR excluded.
- This looks at a QB's ability to make plays when the pass protection fails, but not so much so that he can't still get the ball out.
I then look at each play as they fall into one of these three categories and create the following summary for each metric:
QBS | Metric | Speight (n=?) | O'Korn (n=?) |
Yards | ? | ? | |
Yards Per Play | ? | ? | |
DSR | %? | %? | |
Success Rate* | %? | %? | |
TD/INT | ?/? | ?/? |
*A quick note on success rate:
- Success rate is a very interesting metric in itself. See The Mathlete vs. Bill C. for full definition and discussion of its relevance.
- tl;dr: it has solid utility in analyzing Michigan's passing offense because the offense is average at best.
With the definitions laid out, we're onto the comparisons.
-------------------------------------------
1) Pure QBS:
Pure QBS | Metrics | Speight (n=30) | O'Korn (n=13) |
Yards | 275 | 124 | |
Yards Per Play | 9.17 | 9.54 | |
DSR | 66.67% | 83.33% | |
Success Rate | 50% | 62% | |
TD/INT | 1/1 | 0/1 |
-------------------------------------------
2) Explosive QBS:
Explosive QBS | Metrics | Speight (n=34) | O'Korn (n=12) |
Yards | 291 | 136 | |
Yards Per Play | 8.56 | 11.33 | |
DSR | 66.67% | 75.00% | |
Success Rate | 50% | 67% | |
TD/INT | 2/0 | 1/0 |
-------------------------------------------
3) Playmaking QBS:
Playmaking QBS | Metrics | Speight (n=17) | O'Korn (n=6) |
Yards | 43 | 22 | |
Yards Per Play | 2.53 | 3.67 | |
DSR | 57.14% | 60% | |
Success Rate | 24% | 33% | |
TD/INT | 0/0 | 0/0 |
-------------------------------------------
Analysis:
Go Blue, beat State. (I'm still working on breaking these down by O-form, play type--i.e. PA vs. dropback, WR vs. HB vs. TE target, route #s, pass blockers vs. pass rushers, etc. All that to come sometime soon, hopefully...)
October 7th, 2017 at 4:08 PM ^
October 8th, 2017 at 12:44 AM ^
guessing that the statistical comparison might not favor JOK anymore...
OT: nice logo. We just moved to the Columbia, TN area. Wolverines gotta stick together down here. Go Blue!
October 7th, 2017 at 4:18 PM ^
for tonight's game. I've had too much beer already to even try to figure this out. Not your fault, OP. I blame it on the late start time for the UM game.
October 7th, 2017 at 4:19 PM ^
in 1 game and 3 passing plays from florida, O'Korn performed better, controlling for poor pass protection and WR drops
October 7th, 2017 at 4:19 PM ^
October 7th, 2017 at 4:20 PM ^
October 7th, 2017 at 4:23 PM ^
October 7th, 2017 at 4:25 PM ^
October 7th, 2017 at 4:37 PM ^
i'm on my 4th mgobeer
October 7th, 2017 at 4:44 PM ^
wish I had finished it two days ago.
October 7th, 2017 at 5:22 PM ^
October 7th, 2017 at 5:23 PM ^
October 7th, 2017 at 5:23 PM ^
October 7th, 2017 at 5:43 PM ^
Late games. Pfff...
October 7th, 2017 at 6:20 PM ^
October 7th, 2017 at 7:43 PM ^
October 7th, 2017 at 7:49 PM ^
The eyeball test
October 7th, 2017 at 9:33 PM ^
is that O'Korn brought that extra edge last game b/c of the whole former coach thing.... Here's hoping it marks a permanent corner-turning...
October 8th, 2017 at 12:35 AM ^
October 8th, 2017 at 1:26 AM ^
well howd tht work out?