Offensive Success Rate (OSR) Stats

[IdiotKicker]

New item up on Inside the Pylon, as we've partnered with Mike Backherms to display his Offensive Success Rate (OSR) data. Long and short of it is that it measures how often an offense picks up the yardage needed to continue moving the chains. So it measures how often teams pick up 50% of the yards needed on 1st down, 50% of the yards needed on 2nd down, and 100% of the yards needed on 3rd and 4th down. Helps to give context to how different offenses operate and where they are at their best.
 
OSR Stats

 

[Super Nomario]

For those who don't know, "success rate" is one of the foundations of DVOA.

 

[singaporesoxfan]

Any reason why OSR uses 50%/50% for first and second down while DVOA uses 45% and 60% respectively?

 

[soxfan121]

Any reason why OSR uses 50%/50% for first and second down while DVOA uses 45% and 60% respectively?

 
I shall get you an answer. 

 

[Bellhorn]

Also, has any research actually established that this is a worthwhile approach to measuring offensive success in the first place?  One of Brian Burke's articles from a few years ago showed that probability of achieving a first down was more or less linear with respect to yards required - if this is true, it means that the whole notion of "staying on schedule" for a first down is of dubious relevance.
 
Article is here, though the key graph seems to have disappeared.

 

[EricFeczko]

Also, has any research actually established that this is a worthwhile approach to measuring offensive success in the first place?  One of Brian Burke's articles from a few years ago showed that probability of achieving a first down was more or less linear with respect to yards required - if this is true, it means that the whole notion of "staying on schedule" for a first down is of dubious relevance.
 
Article is here, though the key graph seems to have disappeared.

Success rate is useful for itnerpreting the running game. There's less of a relationship between success rate and running efficiency metrics (e.g. YPC), and success rate has a stronger association with certain forms of WPA in win probability models.
This makes sense when you think about it; normally, the running game is used to accomplish a specific purpose as opposed to just racking up yardage. For example, situtations like 3rd and 3 or 4th and 1, or 1st and goal from inside the 5 are situations where running the ball successfully does not require getting a lot of yardage.

 

[IdiotKicker]

Also, has any research actually established that this is a worthwhile approach to measuring offensive success in the first place?  One of Brian Burke's articles from a few years ago showed that probability of achieving a first down was more or less linear with respect to yards required - if this is true, it means that the whole notion of "staying on schedule" for a first down is of dubious relevance.
 
Article is here, though the key graph seems to have disappeared.

This is pretty much what OSR measures. If you get 0 yards on first down, can you pick up 5 yards on second down to make things manageable for third/fourth down. If you get 6 yards on first down, you only need to pick up 2 yards to maintain the same odds of a first down. They're slightly different in terms of the exact yardage they demand, but it's the same principle.