Yards Above Replacement – Week 1
If you’ve read any of my stuff on this site, you’ve undoubtedly heard about Yards Above Replacement (YAR).
YAR is a statistic that I developed because I realized that with the arrival of spread offenses in college football, the value that a QB brings to the table running the ball is significant and needs to be quantified.
This past weekend (9/3/2022) is a great example. Kentucky’s Will Levis threw for 303 yards and averaged 9.5 yards per attempt. Florida’s Anthony Richardson threw for 168 yards and averaged 7.0 yards per attempt. Levis had the better day, right?
Well, the thing missing from that analysis is that Levis ran the ball 7 times for -18 yards (4 sacks) while Richardson ran 11 times for 106 yards. What that means is that on a per-touch basis, Richardson averaged 7.8 yards per play while Levis averaged 7.3
But more than that, the average QB pass in college football is somewhere around 7.5-8.0 yards per attempt while the average QB run in college football is somewhere around 2.5-3.0 yards. That means that we should weight Levis’ and Richardson’s averages to reflect that if we’re looking for a comparison.
That’s how you end up with Richardson having a YAR of 1.57 against Utah with Levis having a YAR of 0.34 for Kentucky. And also why I heavily lean Florida this week against the Wildcats.
Those are just numbers, but they’re normalized to zero, meaning that a WAR of 0 is considered an average QB. For frame of reference, here’s a list of some QBs you might know from last season and where they stood in regards to YAR.
- YAR = 0, average QB (Levis Lewis, Louisiana, -0.17)
- YAR = -1, terrible (Will Rogers, Mississippi State, -1.28)
- YAR = 1, very good (Matt Corral, Ole Miss, 1.05)
- YAR = 2+, Heisman candidate (Stetson Bennett, Georgia, 2.32)
The following is the top-5 SEC QBs by YAR during this week and then a running tab of where all of the starting SEC QBs stand (minimum 25 combined passing and rushing attempts per game).
Week 1:
Overall:
PMB-BTR
I guess I will never understand YAR. My takes aways:
1. a positive YAR is good
2. The bigger the number of the YAR, the better
3. Therefore the better the QB when comparing to other QBs. How does this add to my understanding in a way different than total yard passing and running?
If running QBs is the metric being sought, why doesn’t rushing stats alone suffice?
Why do sacks matter at all anymore?
Will Miles
So I wrote a primer on this last week, but the point of YAR is to help bring into one number all of the efficiency stats that matter for a QB. Zero is the baseline and anything below is below replacement (i.e. you’d be better off with an average QB vs. what you have) and anything above is above replacement (you’d rather have that guy than an average QB). But the problem is that especially in college, QBs put up numbers in very different ways.
For example, QB1 throws for 350 yards on 30 attempts but runs for -17 yards on 10 carries. His YAR is 1.63 (very good). However, if he throws for 350 yards on 45 attempts and has the same running stats, his YAR is -0.98 because it took him way more throws to get to 350 yards. If he throws for 350 yards on 45 attempts but runs for 126 yards on 10 carries, his YAR is 1.63 because the running portion of the game has differentiated him amongst his peers.
YAR isn’t the be-all/end-all. You can look at a whole host of numbers to draw conclusions, and you really should if you want to draw good conclusions. But just total yardage doesn’t tell you anything about efficiency, and efficiency over the course of a season correlates to points.