We know what field position is and we know it's importance, but we may not really know exactly how important. How much of a difference does 10 yards make? How much does a kickoff touchback affect a game versus a return to the 28 yard line? How much difference does downing a punt at the 5 yard line vs letting it go for a touchback make? Most people would answer that it makes some difference, give some pointed examples of how it affects scoring, and then shrug their shoulders. This is an attempt to quantitatively analyze the NFL data from 2011 and actually give numerical answers to the above (and more) questions.
The full blog post which includes all the charts and graphs mentioned is here : http://www.raiders1.ninjagoro.com/?p=946
The importance placed on field position is related to scoring. If superior field position did not yield more scoring, few people would concern themselves over it; but it does. So the first task is to construct a model where scoring is a function of starting field position. That is to say, build a mathematical model that shows how many points were scored when a drive starts at various places on the field.
Here's the model that represents the NFL average.
This gives us is an idea of how effective defenses are in general and how that related to field position.
The farther to the right you go, the easier it is to score, so the Points per drive (PPD) increases. If you start a drive at the 90 yard line, the model shows that, on average, a defense will surrender 4.23 points per drive. That means at least a Field goal and often a Touchdown. But then, if you start a drive at the 60 yard line, the PPD drops to 3.44. And if you go all the way back to the 20 yard line, we have a 1.53 PPD scoring rate.
The shape is essentially as expected, but we see an interesting plateau from about 60 yards to 86 yards. This is a leveling-off and then dipping of the scoring output and then a sharp rise after that. At 60 yards, the ppd is 3.44 it flattens out to from 67 to 86 (about 3.8 ppd)
This is an interesting phenomenon and makes sense when you think about it. Consider the range from 83 to 60, which is the opponent's 17 yard line to 40 yard line. What the chart shows us is that if a drive starts between the 60 to the 83, there is little difference in scoring results. 10 or 20 yards in this range slightly increases the chance of scoring a TD, but not significantly (over time).
Going from the 60 to the left, the graph trends downward. But we also see that around the 35 yard line, the graph starts to flatten again, leveling off to about the 15 yard line and then dropping off again.. This also makes sense. As we get farther away, it gets correspondingly more difficult to score even a field goal and as we get to around 35 yard line, the benefit of a few yards to scoring is less crucial than beyond 35 yards.
For example, at the 45 yard line, we have a PPD of 2.37. 10 more yards to the 55 gives us a PPD of 3.07, a net increase of .7 point per drive (an extra TD every 10 drives), a benefit for that offense, a detriment to that defense.
Now, at the 20 yard line, the PPD is 1.54. Add 10 yards to the 30 and the PPD increases to only 1.62, a net gain of .08 points per drive.
So 10 yards of starting field position is VERY valuable at the 45 yard line and less so at the 20 yard line.
You can look closely at the different regions, combine it with football situations and you have an interesting profile of how field position plays into scoring.
But there's another important piece: the number of drives. The graph above tells us that we want to have more drives start with better field position. It tells us that farther to the left and the defenses perform better and farther to the right the offenses perform better. We also see where it lulls. We can also deduce that on offense, we'd like all our drives to start at beyond the 90 yard line and on defense, we'd like all the drives to start beyond the 10 yard line. That much was obvious before looking at the graph! But now, we can have a look at where the drives actually did start.
First of all, we notice a steep hill at 20 yards that then tails off to the right. There are a couple of different regions with differing "steepnesses" (slopes). At 20 yards, we see the maximum of 47 drives. At 15, it drops to 28 drives and at 6, it is a mere 16 drives (which averages to 1 drive per game). 36 drives began at the 25 and out at the 75 yard line, a mere 3.8 drives began.
Touchbacks occur at the 20. Kickoffs end up in the endzone most of the time (this past year, especially) so we'd expect that large number. Punting shifts the field position to the left and we see how effective teams were at pinning opponents inside the 20. The tail to the right is composed of turnovers, long returns, blocked punts/kicks, and missed field goals.
So what we glean from this graph is that the offense wants it shifted to the right and the defense wants it shifted to the left. Obviously. But this graph just gives a visual representation of how much this is occurring.
So we see that the trend is that as the scoring rate increases, the scoring decreases; this is because the number of drives drops sharply beyond 20 yards. What is interesting is that even with the substantial increase in scoring rate as the graph moves to the right, the majority of the points scored over the course of a year occurs around the 20 yard line. 73 points at the 20 yard line. 56 at the 25 yard line and 25 points at the 54 yard line.
Imagine being able to shift the drive curve consistently one way or the other, diminishing the lower performing regions while pushing more drives into the higher performing regions. The result is a noticeable scoring change.
From a defensive perspective, the goal is to get the opponent to start from within their own 30. Beyond that, is gravy but not a significant change. From a punting perspective, if you are punting from the 70 (your own 30), then there's about the same value in a 40 yard net punt as a 45 or 50 yard net punt. So if you are Shane Lechler (for example) and can boom a punt that is fair caught at the 30 yard line (40 yards), that's probably a better play than punting it to the 20 (50 yards) and allowing a return that has a chance to go beyond the 40. There's not much difference in giving up a ball from the 80 to 65. If you are going to fumble or throw an interception, the 60 yard line is about the same as the 80 yard line. But fumbling inside the 15 is a killer.
From 65 to 35 is the huge slope change. There's a huge benefit to the defense for each yard gained in this range.