re: Remember when Fournette said the NFL was "really easy"

quote:

---

You keep saying that, but the data actually shows he has 4.6 ypc, while having the second most carries

---

quote:

---

You aren't "looking deeper into the data."

You are actually removing data

---

quote:

---

13 of his runs were for negative yards, 9 were for no gain, 20 were for 1 yard, 17 were for 2 yards, and 17 were for 3 yards. So 10% of his runs have been for negative yards, 16.9% of his runs 0 or fewer yards, 30% of his runs have been for 1 or fewer yards, 45.3% of his runs have been for 2 or fewer yards, and 58.5% of his runs have been for 3 or fewer yards.

Only 34 of his runs have been for 5 or more yards (26.2%), he's had fewer 5+ yards carries than 1 or fewer yards carries (39), and his 21 carries of 7 or more yards is one fewer than his 22 carries of 0 or fewer yards.

---

quote:

---

I could so that, but a quick perusal of the top rushers, indicates that of the top 20 besides Fournette, the longest 3 anywhere near 70 yards (Hunt at 69, Blount at 68, and Hyde at 61) with 3 more rushers with a long in the 50s. 8 of those have more 20+ yard carries than Fournette.

In the top 50, 2 other rushers have a long of 75. And 12 have more 20+ yard carries.

---

quote:

---

But there have been 4777 total rushes thus far or 4647 non-Fournette rushes.

There have been only 4 rushes of 75+ yards, and Fournette has 2 of them (50%), 100 the rate as the rest of the league.

But it drops from a there as he only has 2 of the 18 rushes of 50+ yards (11.1%), 2 of the 25 of rushes of 40+ (8%), 3 of the 47 of the 30+ yards (6.4%), and 3 of the 113 of (2.7%).

Given how rare a 75+ yard rush is, and he has 2 of them, if we would expect that to be indicative of a home run hitter, I would think that he would have more than 1 of his other 128 rushes (2nd most when minus 2 carries and 3rd is 123), to be more than 20 yards.

As it stands, that's a slightly lower percentage than the league average of runs of 20+ yards, he just happens to have 2 extremely long but extremely rare runs.

---

quote:

---

And that would make sense if all but 3 of his 130 carries were 18 yards or less.

He actually breaks long runs at a rate slightly lower than the rest of the league, he just happens to have 2 extremely long runs, on back carries.

---

quote:

---

So, if the 4646 non-Fournette carries, 494 are if 10 or more yards, which is 10.6% of all non-Fournette carries.

Of Fournette's 131 carries, 13 are of 19 or more yards, which is 9.9% of Fournette carries.

So the percentages are pretty comparable (1 rush less for Fournette)

---

quote:

---

Statistically, if we were modeling the data for a study, we would usually institute some sort of resampling mechanism (bootstrap, jackknife, etc.), for this very reason. In fact we could do it for all players, to find an expected value and the bias of the expected value of just typical YPC.

---

quote:

---

Damn. I feel bad for you man. You've spent so much time researching the statistics to back up your opinion on Fournette and posting about it in this thread and no one cares.

Must suck.

---

quote:

---

But that's the thing, even with 130 carries, there is so much variability (standard error of 0.923 YPC) that 95% CI around that 4.6 is +/- 1.8 YPC. Take Bell for instance and his 134 carries, the standard error around his YPC is only 0.44 YPC, or a 95% CI of 0.88 about +/- 0.86. I even did the 100,000 resamples of the data, with 100 carries for each sample, taking the mean of each resample and subtracting that from the observed YPC. Bell's median difference of all 100,000 samples from his YPC was 0.01447 yards lower, whereas Fournette's was 0.125 yards lower with a standard deviation of 1.05 for all 100,000 samples. In other words, despite having 130 carries, Fournette's data are so variable, with such a positive skew, that it's it's not a reliable measure of YPC at this point, and all data suggest it's an overestimate. Which is all I've been trying to say.

---