Perhaps the BEST B1G Forum anywhere, here at College Football Fan Site, CFB51!!!

I'm not sure how else to explain what I'm saying.  Regression to the mean comes with greater sample size, not smaller.  

I always wonder if Emmitt had stayed for Spurrier's first year.  His carries might've been down, but his YPC might've gone up, thanks to a wide-open passing threat.  And despite the loss of carries, he'd have caught the ball a lot more and maybe helped his draft stock.  

It just would've been neat to see it.

Point you've been making: more carries = less ypc, because there's some "mean" for each player that they will regress to as sample size increases.

Point I've been making: ypc is dependent on system and usage.

Then you said:
I.e. you're admitting that system and usage plays a large part in ypc.

Ron Dayne had lower ypc because Wisconsin's OC and their opponents DC both knew that they were going to run him on first down, then second down, and probably again on 3rd down, because they'd be in 3rd-and-short more often than 3rd-and-long when you have a back like Dayne. Defenses were loading the box to stop Dayne but still couldn't do it, so it made sense offensively to just keep riding him all day long.

The point being that his ypc wouldn't go up BECAUSE his carries go down, but that his ypc would increase AND his carries would decrease BECAUSE the defenses had a wide-open passing attack to account for and couldn't key on Emmitt every down.

It's not regression to the mean. It's system and usage.

There's the team mean, though.  That's where the efficiency aspect comes in.  Every guy over the team mean should get more carries, every guy under it, fewer.

A few of you seem to insist it's an either/or thing when it's not.  It's both/all.  A RB and a team has a mean, depending on talent and system and play-calling.....his usage is mostly what I've been focusing on.

There's the team mean, though.  That's where the efficiency aspect comes in.  Every guy over the team mean should get more carries, every guy under it, fewer.

With additional carries, Jones' new carries 'type' will be McFadden-like usage.  If Jones' normal carries are of a type that allows his higher YPC, then adding McFadden-type usage will pull his YPC down.  

Okay, put it another way...
McFadden and Jones were both RBs at Arkansas at the same time. So we can remove coaching differences, system differences, etc. There's no truly obvious talent differential here either. McFadden was drafted higher at 4th overall vs Jones at 22nd, but they were both first-round NFL talents.
Do you believe that the coaching staff used them both interchangeably with regards to playcalling? Or did they call different plays based on whether they had McFadden or Jones on the field? both, at different times
 Did the coaches use them identically, or did they call plays suited to each player's strengths. Did they substitute players identically, or did they use one or the other more in certain down/distance, certain field positions, game/score/time situations, etc?
I am asserting that they were NOT used identically, and that usage is the reason why Jones had ~2 ypc advantage. Not that he was fresher  but he was, or not that the sample size wasn't big enough. I think Jones had ~2 ypc because the coaches used him in situations and playcalls that had a higher likelihood of being broken for long gains.  do you have evidence?
I submit that you're right, in one sense. If Jones had gotten more carries, and they were the playcalls that the coaching staff would have ordinarily given to McFadden, his ypc would probably have gone down. Possibly by more than "expected", as he wasn't suited to the playcalls McFadden got.
However, I submit that the corollary to this is that if McFadden got more carries, and they were the playcalls that the coaching staff would have ordinarily given to Jones, his ypc would probably have gone up. Possibly not by as much as "expected", as he wasn't as suited to the playcalls Jones got. But still up.  I'd need evidence of this, which is probably impossible, unless Jones was hurt for x-number of games and McFadden's carries increased by a noticeable number.  It's a "what if", so evidence isn't realistic, but if it's impossible, then it's unfalsifiable and we can stop wasting time.