Emmitt Smith on all-time ranking: My stats speak for themselves

[midswat]

Emmitt had the most rushing yards.

Most rushing TD's.

Once held the single season rushing TD's record.

Three Super Bowl rings.

One Super Bowl MVP.

One NFL MVP.

Longevity.

Epic moments in critical games.

Total. Package.

 

[Genghis Khan]

Which RB do you think is misrepresented by the YPC stat? Which YPC stat do you think is misleading?

Definitely YPC can be skewed by runs on 3rd & 20, team always losing/winning or big plays (80yd run vs 30yd run rarely differ in skill but show up stat-wise), but I wouldn't call it useless. These aren't yardage stats from RBs with 100 touches. They range from seasons with 250-375 rushes.

Take Emmitt Smith, who we know scored a lot of <3yd-run TDs. Say he scored 10 TDs-per-year of <3yds more than other RBs. Let's eliminate these total yards while eliminating 50 rushes. His Top 5 stat of 4.59 YPC increases to 4.64 YPC. His Top 7 stat of 4.47YPC increases to 4.52YPC.

I'd agree with you about +/-5% in the middle spectrum that you can't call a 4.7 YPC player automatically better than a 4.5 YPC player, but this stat does confirm perceptions that AP, Barry, Chris Johnson were more electrifying big play threats over Eddie George, Jerome Bettis.

Nearly useless. Not entirely. I do think if there's a low outlier it is telling, like George or Bettis.

But take a case like Randle. Or Hambrick when he was backup. So many things can skew YPC that context for that stat is extremely important for a running back. Volume matters too because outliers affect it less as volume increases.

Ultimately it comes down to the fact that I value consistency from a running back much more than big plays, but big plays skew YPC.

Give me 20 carries of exactly 4 yards each all day over 15 carries of 2 yards each plus 5 carries of 10 yards each.

The reason that the 4 YPC threshold matters is that if you consistently get 4 or more yards every carry, you can get a first down theoretically without passing (obviously no one gets it every single time but we're talking theory).

The problem with YPC is it is terrible at measuring consistency, and consistency is extremely important for a running back.

 

[data]

Nearly useless. Not entirely. I do think if there's a low outlier it is telling, like George or Bettis.

But take a case like Randle. Or Hambrick when he was backup. So many things can skew YPC that context for that stat is extremely important for a running back. Volume matters too because outliers affect it less as volume increases.

Ultimately it comes down to the fact that I value consistency from a running back much more than big plays, but big plays skew YPC.

Give me 20 carries of exactly 4 yards each all day over 15 carries of 2 yards each plus 5 carries of 10 yards each.

The reason that the 4 YPC threshold matters is that if you consistently get 4 or more yards every carry, you can get a first down theoretically without passing (obviously no one gets it every single time but we're talking theory).

The problem with YPC is it is terrible at measuring consistency, and consistency is extremely important for a running back.

I agree with you that consistency is important. I'd love to include a stat that considers range (also mean/median). What percentage of runs were over 4 yards, especially when down & distance was 4+ yards?

However, other than maybe Barry Sanders, do you think this would significantly affect other RBs on that list? Jim Brown, AP and OJ had the highest YPC. Do you think a intelligent YPC stat would reveal any of these guys to be frauds?

Regarding Randle (4.5 career YPC) and Hambrick (4.1 YPC) debunking the above list, c'mon man...You think anyone is confusing Randle/Hambrick to the great RBs on the above list? Randle averaged 60 rushes-per-season and Hambrick had ~600 career rushes. Conversely, the stats above incorporate 1,5000-2,500 rushes per player! Which RB on the above list is a Randle/Hambrick-type?

I think we agree on this, though. Overall, I don't think the YPC stats within 6% of each other mean much for comparing RBs. 4.50 YPC vs 4.30 YPC isn't a good indicator.

However, YPC stats is useful to claim AP's 4.99 YPC that he's a better RB than Curtis Martin's 4.12 YPC, despite their average-rushing yards-per-season to be almost exact (3% difference). I think we already knew this and this stat confirms our perception.

Sounds like you think YPC is 1/5 useful. I think it's 3/5, especially since the above list eliminates the Randle/Hambrick-types.
-----------
Comparing RBs of different generations is a whole different potato. It'd also apply to yards-per-season, but I'd like to see how YPC varies per generation. Very similar to dollar inflation. Adjusted for 'inflation', what would Jim Brown's 5.0+ YPC be in today's game? I'm assuming it would be higher since today's game is more spread open and gaining 1-3 yards is easier now than the 1960s. How many RBs in 1960 were averaging 4.5+ YPC versus 2010s?

 

[Cowboysrock55]

But take a case like Randle. Or Hambrick when he was backup. So many things can skew YPC that context for that stat is extremely important for a running back. Volume matters too because outliers affect it less as volume increases.

If you think in terms of statistics it is your sample size. If a runningback is carrying the ball 20 + times a game and still maintaining a YPC over over 5 yards, that's extremely impressive. The large sample size means that a few big runs aren't skewing the statistic at all.

So ultimately, yeah I'd say YPC does mean something when you're looking at a list of guys who carried the ball more than any other RBs in the NFL.