You almost always see things as either a great stat or awful. I would challenge you to not be so binary in your analysis and be open to the idea that a stat can be useful without being a perfect indicator
I think you'd be hard-pressed to find a statistic in football that I feel is great. I think we're too used to baseball and basketball, where we have massive amounts of repeatable events, and therefore terrific statistics. Between the small number of plays per game, the small number of games per season, and the immense number of interdependent variables on the field for each play, football isn't a sport that lends itself to summary statistics very well.
For what it's worth, I think yards per carry can be useful in the same way that wins can be useful for evaluating a pitcher. Sure, mention them, but they are an awful crutch to try and support yourself if you're going to make any evaluation of the player themselves.
Your example falls flat for me because I never claimed it to be anything other than "not an awful" stat
I think that by bringing it up, and by claiming that yards per carry was the reason for praise, you did infer that the stat is significant.
. You picked out some guys that most wouldn't consider good that had objectively good seasons
This is a point where there's a massive disconnect.
I don't think those RB's were good in years they had high yards per carry.
The stat is one that's situational and speaks more about the team around a back than the back himself.
Whether it's by using examples of bad players having good YPC, good players having bad YPC, or individual players having wildly differing YPC, those examples are used to illustrate the lack of correlation between YPC and RB ability.
and I don't see how that proves that it's an awful stat. Was Doug Martin not good in his rookie season? Was Lamar Miller not a good runner during his peak?
My claim was never: YPC is the end all be all and it is a definitive ranking for the best RBs in the NFL, year in and year out.
My claim is that is is a heavily correlated stat for RB play over large samples.
I still disagree with that claim. I think the correlation is closer to Wins for a pitcher in baseball than it is for an actually predictive baseball stat like xFIP.
Some guys have anomalous games. Some guys have anomalous years
Agreed. If we have scales and one side is "individual performance" while the other is "imperfect statistic" then which side do you feel carries more weight as to the variance of yards per carry?
A lot of people on this board pretend they watch every game for every team so they know everything. They don't. The stats are a short hand, objective, and quantified way to compare players. YPC isn't the only thing that goes into a RB, but I can tell you this much, there is no good RB that has ever been below 3.
You are taking a heavily correlated with talent variable, pointing to some anomalies and then saying "SEE! I told you it was flawed!
I agree with most of this except for the value you're placing in yards per carry. Below 3 certainly is atrocious, but Gurley, who was arguably the most talented RB in football, averaged 3.2 yards per carry under Jeff Fisher before winning Offensive Player of the Year and finishing second in MVP voting the following season.
So, putting all that aside, if we were to talk about Najee's yards per carry this year as opposed to last year, my impression would be "Wow, how bad is that Pittsburgh offense!" I don't feel the stat says much about Najee Harris himself.