Thread by @rbkeeney

"He came into the league too soon"

That's something @noahmoreparties mentioned the other day when we were chatting about Gio Bernard. It's true for lots of players: Jahvid Best, Harvin, etc.

That got me to thinking: How are running back prospects changing over time?

And, spurred on by a simple question, your humble analyst sets off on a (partially and poorly animated) adventure...

So yeah, essentially you should prepare your timeline for another slowly unraveling twitter thread.

Can't stop with gganimate. Won't stop with gganimate.

Let's start a high level: If we split R 1-3 and R 4-7, we're not seeing a change in my model ratings.

That said, to get a 70+ rating, it usually requires either (a) a *generational* profile or (b) a very good profile w/ 1st round pick.

There are two primary narratives around a perceived drop in highly drafted RBs:
1/ This year's deep rookie class *might* have pushed some RBs down.
2/ Teams are devaluing the position more than ever #rbdm

I'm a bit skeptical of both claims have had a LARGE effect. A small one? Maybe.

My draft-independent model didn't put any prospect in this class above the 80th percentile. It's a bit apples-to-oranges, but a 70+ rating = 90th percentile.

More here:
breakoutfinder.com/keeneys-runnin…

There's a chance that this class is amazing and the two narratives I mentioned were *just* enough to knock Taylor, Swift, and Dobbins out of the 1st round. Or, suppose they are merely very good - and were drafted as such.

Of the two scenarios, I suspect the latter.

Okay, with that out of the way, I'll start focusing on how running back college production, efficiency, and usage is (and isn't) changing.

...tomorrow.

Before we jump in, I think it's worth revisiting the model when we neutralize draft capital.

Note: It's the same model - but I've just every prospect to a 3rd round pick. The r-squared to actual ppr per game is 0.42. Decent, but not nearly as good if we include draft capital.

To compare prospects, we no longer want to look at their breakout rating... it's better to look at where a player is in relation the mean (e.g. how many standard deviations they are from the mean)

That's what I've plotted here (trend line = ALL)

All things considered, from a high-level perspective, prospects haven't changed much since '03. Does that change when we look increasingly better and better prospects? No.

Alright, I think you'll agree that from a high level - looking at my model's outputs - the quality of running back prospects hasn't changed significantly since 2003.

That holds true for our draft-dependent AND draft-neutral reviews.

For fun, here's an animation.

Next, let's zoom in and see if individual metrics like yards per touch, MS rushing yards, weight, speed-scores are changing over time.

I'll be back later today with some more data.

What I'll look at the first:
- total scrimmage yards
- yards per touch
- yards per team pass attempt
- Share metrics (rushing, passing for yards and TDs)
- Speed score
- Weight

... I'll probably add in receptions and a few other things as well.

Total scrimmage yards are all over the place if you look at it by draft round.

Okay fine, you got me, this gif is just me having fun with box plots.

Total scrimmage yards: no significant change
Split by rounds 1-3 and 4-7 (this will be a theme)

It's all over the place.

The same is true for yards per touch, yards per team pass attempt, weight, and speed score.

Okay, a few more things to check... share-of-offense stats.

Best season share of rushing yards:
Outside of a few outliers, there's really nothing to suggest usage.

Animated by draft round...

And finally best season share of receiving yards:

...MS receiving yards animated...

Alright. That wraps it up.

In this simple exercise, I didn't uncover anything that leads me to think that the prospects from college are changing nor any evidence that the NFL is selecting them differently.

NFL usage might be a different story (the narrative of the RBBC growth, death of the workhorse, and increasing target shares, etc). I'll need to research that.

I'll remain skeptical about the latest RB narratives. I think it's often just a case of: exceptional prospects command exceptional usage and production.

And, we've had a lot of exceptional running back enter the league lately... but maybe not this year.

https://twitter.com/rbkeeney/status/1261001958816694272?s=20

@threadreaderapp unroll