This is just a quick post based on some numbers that Shawn Siegele asked about related to RBs and the defenses they face. All of the FPOP numbers that I post weekly are raw numbers, which is to say that they’re only adjusting fantasy points based on opportunity, and not based on opponent. Adjusting for opponent is certainly something that I’d like to look into further in the offseason as I think there are ways to get the FPOP numbers to also work with the Buy Low Machine.
My first stab at adjusting for opponent was to simply give each defense an average FPOP allowed, and then to plot RBs on a graph where the x-axis is the defense’s FPOP allowed per play (to RBs) and the y-axis is the RB’s FPOP per Opportunity. You can see from the graph below that there is definitely a linear trend. In order to highlight the players that fall furthest away from the trend line I made the players closest to the trend semi-transparent. I’ll try to walk you through some things in the graph that are a little easier for me to see.
- On an opponent adjusted basis Andre Brown has actually not been very good.
- Ray Rice and Bernard Pierce are still about equal amounts of bad.
- Ryan Mathews has been about average once you adjust for opponent. He’s right on the trend.
- LeSean McCoy and Jamaal Charles are still the two best high volume backes, even if both have played relatively easy schedules. The same is true to a lesser extend with DeMarco Murray. He hasn’t been as good as McCoy/Charles on a per attempt basis.
- Eddie Lacy has played an easy schedule which has resulted in a slightly positive FPOP/Opp number. But he’s not quite as good as would be predicted just based on opponents he’s faced.
- Adrian Peterson has faced a relatively easy rush schedule and has been slightly above average per attempt. His FPOP number looks really positive until you consider opponents.
- Montee Ball’s attempts have come against relatively weak rush defenses, which would suggest that he should be averaging .05 FPOP/Opp, and yet he’s been just slightly negative in FPOP. That’s on a small sample of runs though.
- Doug Martin faced a tough run schedule but then he also underperformed that schedule. Bobby Rainey has faced a slightly easier slate and then has gone on to overperform expectation.
- You’ll notice that the defense allowed FPOP/Opp is not centered on zero. That’s because in calculating FPOP I use the average of play results from 10 years of data. So when I display just this year’s results, it doesn’t center on zero (FPOP is net positive on the season, which may or may not be problematic – I’ll have to think about whether I should simply scale each season’s FPOP numbers – but that’s for another post).
Notes on the graph: FPOP/Opp is the y-axis, while the average of the defenses faced FPOP allowed is the x-axis.