Stat Analysis
Advanced analytics on player and team performance

Speed Score 2019

by Bryan Knowles

The NFL Scouting combine can be a place where draft dreams are made. A great combine performance can vault someone up the draft order, giving a prospect more opportunities, more snaps and, frankly, more money than they would have received from their tape alone. It can also, however, be a place where dreams are crushed; where an apparent lack of athletic ability or poor performance on the day can send someone's draft stock spiraling.

Last year's headliner among the running backs was Saquon Barkley, who wowed us with one of the greatest combine performances we had even seen, and the fourth-best Speed Score since they introduced electronic timing back in 1999. We knew there was no Barkley in this year's class, and the consensus best running back of the year in Josh Jacobs didn't perform at the combine due to a groin injury. Still, we had a notably slow set of times: for the first time since 2014, no running back managed to top a 115 Speed Score at the combine, and a couple of decently touted prospects may have done serious damage to their draft hopes.

Created by Bill Barnwell and introduced in Pro Football Prospectus and ESPN Insider back in 2008, Speed Score is one of Football Outsiders' metrics for evaluating running back prospects. It's built on the simple idea that smaller backs tend to run faster than larger backs, so we should be more impressed by a 4.5-second 40-yard dash from a 220-pound back than the same clock reading from a 170-pound back. As such, Speed Score incorporates a back's official time in the 40-yard dash with his weight to produce a measure of his speed given his size using this formula:

(Weight * 200)/(40 time^4)

The average running back who makes it to the NFL will have a Speed Score around 100, with most running back prospects falling between 85 and 110.

Speed Score measures speed in the context of strength and power. It doesn't measure agility, receiving ability, or any of the other aspects related to the position. It does not claim that a larger player with a higher 40 time is somehow faster than a smaller player with a lower 40 time thanks to the power of exponentiation. Speed Score is useful because it's beneficial for a running back to be both fast and large.

Speed Score has a higher correlation with yards, carries, and DYAR than 40-yard times alone, making it a better way to contextualize the performances at the Underwear Olympics and a better tool for finding valuable players later in the draft. It's also a part of our BackCAST projections, which combine these numbers with college production and will come out later this offseason.

Let's start with the Speed Score table, and then discuss the notable names and numbers on it.

2019 Speed Scores
Player School Weight 40 Time Speed Score
Ryquell Armstead Temple 220 4.45 112.2
Mike Weber Ohio St. 211 4.47 105.7
Justice Hill Oklahoma St. 198 4.4 105.7
Jordan Scarlett Florida 208 4.47 104.2
Miles Sanders Penn St. 211 4.49 103.8
Qadree Ollison Pittsburgh 228 4.58 103.6
Darrell Henderson Memphis 208 4.49 102.4
Alex Barnes Kansas St. 226 4.59 101.8
Karan Higdon Michigan 206 4.49 101.4
Tony Pollard Memphis 210 4.52 100.6
Travis Homer Miami 201 4.48 99.8
Trayveon Williams Texas A&M 206 4.51 99.6
Damien Harris Alabama 216 4.57 99.0
Dexter Williams Notre Dame 212 4.57 97.2
David Montgomery Iowa St. 222 4.63 96.6
Benny Snell Kentucky 224 4.66 95.0
Myles Gaskin Washington 205 4.58 93.2
Alexander Mattison Boise St. 221 4.67 92.9
James Williams Washington St. 197 4.58 89.5
Devin Singletary Florida Atlantic 203 4.66 86.1
Alec Ingold Wisconsin 242 4.89 84.6
Nick Brossette LSU 209 4.72 84.2
Elijah Holyfield Georgia 217 4.78 83.1

Did not run: Rodney Anderson, Josh Jacobs, Bryce Love, Jalin Moore, L.J. Scott

The Big Names

Two of the top four running backs in the class did not run: Jacobs and Bryce Love were both nursing minor injuries, and decided to skip the running section anyway. That left Florida Atlantic's Devin Singletary and Iowa State's David Montgomery as the biggest names running. They had very similar 40 times, with just three hundredths of a second separating them. However, Speed Score shows a bigger discrepancy.

Montgomery weighed in at 222 pounds and has a nice, compact frame -- he has a more prototypical NFL size. He's certainly not considered a particularly explosive back, but a 4.63 40 at that weight isn't a terrible result -- a speed score of 96.6 is below average, but not terribly so. While he won't be beating people in the NFL with pure speed, he can create yardage for himself with his vision and instincts, and he has enough speed and mass to power through hits. He was considered a solid Day 2 back coming in, and didn't do much to hurt his stock.

Singletary was only a tick slower at 4.66, but he only weighed in at 203 pounds! There are small running backs who had success in 2018 in the NFL -- Christian McCaffrey weighed in at 202 and had nearly 1,100 yards last season, and Marlon Mack and Tevin Coleman also had solid seasons at small weights. But McCaffrey ran a 4.48 at his combine, Mack had a 4.50 and even Coleman managed to squeak out a 4.59. Singletary's Speed Score clocks in at just 86.1, and that has to be a red flag for evaluators. An undersized back like Singletary needs to be elusive and quick in order to be productive, because he's not going to run people over. A 4.66 is not the sort of speed you hope to see out of a back that size, and that drops him below the dreaded Ahmad Bradshaw Line.

The Ahmad Bradshaw Line

Speed Score doesn't guarantee anything, of course, but the higher your score, the better career you generally have. You can see the production of backs fall off as Speed Score drops off in this table:

There are exceptions to every rule. Arian Foster had a Speed Score of 94.2, but still topped 5,000 yards in his first five years -- he's that "1.4 percent" on the chart, the only running back with a Speed Score under 100.0 who managed to average a thousand yards a season for five years. So, no, a low Speed Score doesn't mean you're guaranteed to be a bust. But…

Ahmad Bradshaw holds the record for worst Speed Score for a productive running back. Back in 2007, Bradshaw ran a 4.61 40 at 198 pounds, for a Speed Score of 87.7. Only one player below that score even managed 1,000 yards in their first five seasons, and most of them washed out of the league after a year or two. That's the low-water mark; an athletic point of no return. If Singletary turns into a successful NFL back, he will be bucking all modern statistical trends.

Singletary did not have the lowest Speed Score of the year, however. Wisconsin's Alex Ingold is a fullback, and thus his Speed Score isn't particularly relevant. LSU's Nick Brossette is a seventh-round prospect who may have bumped himself down into UDFA territory with his combine performance. Then, we get to Georgia's Elijah Holyfield.

Holyfield managed a 4.78 40-yard dash on his first attempt. To put into context, Da'Shawn Hand, defensive tackle for the Detroit Lions, managed a 4.83 at the combine last year. When you have trouble outrunning defensive tackles, you're not going to have a great 40 time. You do get two runs, however, and on his second run, Holyfield could only manage a 4.81. That's a fine time for an Ironhead Heyward or a Le'Ron McClain, basically a bowling ball you crash into the line. But Holyfield isn't that type of player -- he weighed in at just 217 pounds, which is more Joe Mixon or Adrian Peterson size. That's a Speed Score of just 83.1; Elijah's dad Evander had knockouts that happened quicker than Elijah ran the 40.

Before the combine, there was talk of Holyfield sliding up in to the second day thanks to his stout frame and aggressive rushing style, but he just looked like he was running through molasses out there. He has a lot riding on his Pro Day now to show that this was some kind of fluke; it was one of the worst 25 Speed Scores we've ever recorded, including fullbacks.

Standout Performers

Not everyone had a terrible performance in the 40, mind you. There may not have been a Barkley or a Leonard Fournette, but one player did get into the 110s: Temple's Ryquell Armstead, who opened the day with a 111.2 Speed Score. Armstead is a late-round flyer without the lengthy college track record to support him -- he didn't break into Temple's starting lineup until he was a senior, and struggled with injuries. Still, his 4.45-second 40 was the second-fastest time of the day, and doing it at 220 pounds should at least cause teams to give his physical running style another look, and probably will make sure he gets his named called at some point during the 2019 draft.

A 111.2 score is the lowest high-water mark we've seen in the electronic timing era, beating out Jerious Norwood's combine-leading 112.1 back in 2006. This is not a top-heavy class of running backs.

Most of the other backs over 100.0 are also lightly regarded prospects and mid- to late-round flyers. Mike Weber and Justice Hill are ranked around 190th on the current Scouts Inc draft board, though that might go up after Speed Scores over 105.0. Darrell Henderson of Memphis briefly topped the Speed Score rankings with a hand-timed 4.37; the electronic timing knocked him down by more than a tenth of a second, however, turning a great score into just an above-average one.

Of the highly regarded backs, perhaps Penn State's Miles Sanders had the best day. His Speed Score of 103.8 isn't anything to scoff at, and he had a great performance in some of the other drills as well. He was docked some in scouting reports for average acceleration, but he showed that he can turn on the jets when need be, putting up a very good 4.49 40 and showing off some change-of-direction skills in the following drills. For the second year in a row, your combine winner at running back might just be a Nittany Lion, even if Sanders pales in comparison to Barkley's day a year ago.

Comments

22 comments, Last at 07 Mar 2019, 10:50pm

1 Re: Speed Score 2019

BackCAST seems to be all over the place as far as predictions go. What sort of correlation are you looking for to present a stat as meaningful?

3 Re: Speed Score 2019

It looks like D.K. Metcalf just put up a speed score of 129.7...that would give him a better score than any RB on record...by a significant margin.

5 Re: Speed Score 2019

This is true, and it was a tremendous time...but other research has shown that while the 40-yard dash is actually a statistically significant drill for nearly every position, it is in fact ~not~ correlated significantly with success for wide receivers!

http://harvardsportsanalysis.org/2015/02/the-combine-actually-matters-part-2/

That being said, wow, what a 40 from Metcalf...and then what a ~terrible~ performance in the agility drills. Metcalf's Combine spider graph is one of the most bizarre I've ever seen.

https://twitter.com/WhatGoingDowney/status/1102005695522705412

13 Re: Speed Score 2019

'while the 40-yard dash is actually a statistically significant drill for nearly every position, it is in fact ~not~ correlated significantly with success for wide receivers!'

Probably because WR is the default landing spot for speedsters with minimal football background. So we get a small number of extremely fast guys who aren't particularly adept at getting open. (For every Bob Hayes there are five speedsters who cannot catch a ball.)

I don't think there are other positions where speed alone will get a person a tryout. (Well, maybe RB, but a fast guy seems more likely to break through at RB with minimal coaching compared to WR.)

4 Re: Speed Score 2019

Ahmad Bradshaw is interesting because he would have easily topped 5000 yards in the his first five seasons but for the fact that the Giants only made him the lead back in his third season. At the moment, I'm blanking on whoever got the carries ahead of him. It also seems like Eli only realized Bradshaw could catch in the season three as well.
It would be interesting to see how the 2018 top ten or twenty performed in their first seasons. The Chubb-Michel comparison is an interesting one. It looks like Backcast was right in preferring Chubb, although they had similar seasons. One of the reasons I thought the Pats took Michel was that he was a somewhat better pass catcher - but he only had 11 targets on the season.

6 Re: Speed Score 2019

Brandon Jacobs and Derrick Ward were getting carries ahead of Bradshaw for the first two years. Jacobs had one of the best Speed Scores ever at 123.5 (a 4.56 40 at 267 pounds is insane!), so it makes sense that he would be getting those carries...but in hindsight, keeping Bradshaw on the bench in favor of Ward makes significantly less sense.

As for 2018's top ten:

1. Saquon Barkley (124.3) -- went 2nd overall to the Giants; 1,307 rushing yards (1st among rookies) and over 2,000 yards from scrimmage.
2. Kalen Ballage (115.2) -- drafted in the 4th round by the Dolphins; finished with 191 yards on 36 carries (16th). Only had double-digit carries in two games: a 123-yard day against Minnesota and a 47-yard day against Buffalo, both in November. We'll see if he can break out of the Ballage/Drake/Bolden committee in 2019.
3. Rashaad Penny (111.2) -- drafted 27th overall by the Seahawks; had 419 yards on 85 carries (9th). Couldn't beat out Chris Carson as a rookie.
4. Derrius Guice (110.2) -- drafted in the 2nd round by Washington. Tore his ACL, missed the season.
5. Bo Scarbrough (109.2) -- drafted in the 7th round by the Cowboys; spent time on the practice squad and is now a Seahawk.
6. Nick Chubb (108.8) -- drafted in the 2nd round by the Browns; finished with 996 yards (3rd), taking over the starting job when Carlos Hyde was traded.
7. Royce Freeman (107.8) -- drafted in the 3rd round by the Broncos; finished with 521 yards (7th). Beat out Devontae Booker, but the emergence of Phillip Lindsay limited his opportunities.
8. Nyheim Himes (107.6) -- drafted in the 4th round by the Colts; finished with 314 yards on 85 carries (12th). Added 425 yards on 63 receptions as Indy's primary ~receiving~ back, though.
9. Ryan Nall (105.5) -- UDFA, currently on the Bears' practice squad.
10. Chris Warren (103.3) -- UDFA, was stashed on the Raiders' injured reserve list in 2018 with a knee injury.

Of the missing rookies from the top 10:

Phillip Lindsay (2nd), Gus Edwards (5th), Kerryon Johnson (6th) and Josh Adams (8th) did not run at the combine, and thus do not have Speed Scores.
Sony Michel (4th) had a Speed Score of 100.7, 11th-best in the class.
Jordan Wilkins (10th) had a Speed Score of 87.8, 25th-best in the class

Michel's Speed Score is fine; 2018 just was a really deep class. Wilkins had just 336 rushing yards, so I wouldn't call him a huge rookie miss just yet.

15 Re: Speed Score 2019

Hard to judge Wilkins by his rookie year as he was initially kind of fourth on the depth chart, was forced to be the starter due to Marlon Mack's injury, and as noted, Hines got most of the backfield reception reps. He stepped in and did a decent job. My guess is he will be what he was, and what his SS projected--an adequate backup who may well be replaced in a couple years by a 5th round rookie. Colts ran more and better in 2018 than most of the past decade, so if Wilkins's rookie year was in 2016 or 2017, he'd likely have had even fewer yards.

Hines's 63 receptions for a rookie RB look pretty impressive--I suspect every year there's just one or two rookie RBs who might have more, and he was not the starter.

16 Re: Speed Score 2019

Thanks for info, Bryan. Really helps.
Michel is decent, for sure. Hits the hole hard, powers through arm tackles, durable in games, accelerates quickly and makes a decisive first cut. On the negative side, he's not super elusive, doesn't have breakaway speed and takes some odd routes in the open field, which could be poor vision or rookie error. Then there's the lack of targets and the knee concerns. Michel also doesn't escape from too many Belichick taking him in the first is probably the oddest thing although I guess the Pats needed a bell cow and didn't think anyone would be left by the end of round two. Knowing Belichick (not that I do), pass blocking and ball security were probably the biggest selling points.

19 Re: Speed Score 2019

A significant part of Michel's low targets is simply that James White plays on passing downs, and hes both really good in those situations, and has a significant track record of not making game altering mistakes , which Belichick values highly.

9 Re: Speed Score 2019

If only Montez Sweat were a running back...he's 260lbs and ran a 4.41 40! Speed score of 137.4.

While I'm calculating, here are some other top speed scores of non-RBs from the weekend:
129.7-DK Metcalf (WR) 228lbs, 4.33s
125.9-Rashan Gary (DL) 277lbs, 4.58s
124.1-Devin White (LB) 237lbs, 4.42s
121.5-Devin Bush (LB) 234lbs, 4.43s
121.4-Noah Fant (TE) 249lbs, 4.5s
120.9-Justin Hollins (Edge) 248lbs, 4.5s

10 Re: Speed Score 2019

There's been some really athletic performances, for sure.

I do need to clarify that we have not seen the same level of correlation between Speed Score and success at other positions. It does not exist for receivers (possibly because there are more types of successful receiver than there are types of successful running back), and we have not looked at it's correlation with defensive players at this point.

17 Re: Speed Score 2019

That's an impressive time for Metcalf. His body type and speed reminds me of Megatron. However his college resume is pretty sparse. Also a pretty amazing sprint by Rashan Gary who I think is going to be a very good pro. Edit: "A 111.2 score is the lowest high-water mark we've seen in the electronic timing era, beating out Jerious Norwood's combine-leading 112.1 back in 2006" the 111.2 s/b 112.2

18 Re: Speed Score 2019

Looked back at Dontarie Poe's score. 112.5? I figured a 350lb dude running a sub 5.0 had to be off the charts.

Guess not.

22 Re: Speed Score 2019

By my calc, he's actually at 121, which would rank a lot higher among RBs.

If you take a Z-Speed Score (how many standard deviations above average each measurement is, then adding those together) -- where you use position-specific averages & StdDvs, he actually rates out higher than any RB. (Which I think lines up with how freakish that 40 + that weight is).