2015 Combine Observations

[Telperion]

For perspective, last year Aaron Donald ran the 40 only 7 hundredths of a second slower than Abdullah and was 80lbs heavier.

There's no way to spin his 40 time in a good light, and if he tries to put on any weight to add power, he's going to get even slower. It would have to be a perfect situation (maybe Dallas) to work.

Abdullah plays fast, but doesn't have elite straight line speed. His 40 time confirms that. But we already knew that going in.

He plays fast in college. That's a big difference.

[Cameron Giles]

For perspective, last year Aaron Donald ran the 40 only 7 hundredths of a second slower than Abdullah and was 80lbs heavier.

There's no way to spin his 40 time in a good light, and if he tries to put on any weight to add power, he's going to get even slower. It would have to be a perfect situation (maybe Dallas) to work.

Abdullah plays fast, but doesn't have elite straight line speed. His 40 time confirms that. But we already knew that going in.

He plays fast in college. That's a big difference.

The college talent level isn't why he plays fast. He's actually a pretty good football player.

[OhCruelestRanter]

OK. So how does Gio fit in here, or shady? The both ran 4.5, same as Gordon who is significantly bigger. Shady has had a great deal of success, while Gio has had some moderate success. Just two off the top of my head.

Right. So, that tells you that Gordon has a better speed-size profile than two guys, one of whom has had tremendous success, and one who's been very up and down. This tells you that Gordon's size and speed shouldn't be a concern, and that the ability he showed on his tape is more likely to translate to the NFL.

Abdullah is about their size and ran a 4.6 plus, so how do you co-relate all of this?

Abdullah has worse measurables than two guys who you could say out-performed their speed scores. There are a decent number of guys in McCoy and Gio's range who have performed well in the NFL, but when you get down to Abdullah's range, you're really throwing darts.

Gordon is a fast as Gio and Shady, but has 10 pounds or so on them, so is that good? Abdullah is slower but the same weight so that's bad? That would make sense.

Yes, and yes.

I'm just saying Abdullah is about short space quickness, that's his game, so the shuttle simply shows you what the tape is showing you. He's not a burner.

That's true, and it's also true that Abdullah does fit the profile of a RB who can outperform his speed score, but his lack of speed for his size is a very large obstacle. His game might be short space quickness, but he'd need to be incredible quick to overcome that lack of explosive speed.

I have been watching a lot of Gordon, and everybody has been saying he can't be used as a receiver, but he caught the ball extremely well, but I haven't heard that mentioned by anyone else. Surely that's a good sign, to see if you have him as a possible player to draft, no?

Absolutely. I think most people actually realize that's due to Wisconsin's offense.

How do you feel about Duke Johnson being outrun by Gordon? How do you feel about Yeldon posting a similar time to Abdullah? Is a 4.6 for Yeldon at nearly 230 disappointing/average?

I'm quite worried about Duke Johnson, especially with the news that many teams view him as a complimentary back. Of course, he could always land in a perfect situation and become a PPR dynamo, but the problem is that now you're hoping for a perfect situation as opposed to just hoping he doesn't end up in a bad situation. As for Yeldon, he was fine. 4.6 for a back of his size meets that Jeremy Hill profile somebody mentioned a little while back. Of course, I'm not suggesting Hill as a comparable. Hill has a great pad level when he runs and Yeldon looks like Forrest Gump to me when he runs. Still, there seem to be a lot more successful backs in Yeldon's mold than in Duke Johnson's.

[Telperion]

The college talent level isn't why he plays fast. He's actually a pretty good football player.

I get that he's talented, but he's slower than NFL linebackers and on par with some defensive lineman, but he's the same size as defensive backs. That's just not a winning combination.

[FantasyFreak]

Not sure if maybe they had a problem with the sync, but this is an overlap by NFL.com on the top 3 receivers running the 40, and Cooper actually looks to have beaten White by a nose.

http://alabama.247sports.com/Bolt/Did-A ... e-35794229

[FantasyFreak]

Another question to all you guys who play PPR. The guys who were disappointed with all the RB's, and have Gurley as their top pick, doesn't that worry you, at all? I mean, are you guys wearing rose coloured glasses to the point where you think Gurley would have been the only guy not to "disappoint"? Surely coming in you would have though he would of put up better than his below average 17 reps in the bench press, so how would the rest of his numbers been compared to expectation? I don't play PPR, but if I did, I would be looking to take Cooper or White for sure with the 1.01, based on tape/combine and injury history. Cooper is my personal preference. Standard scoring I get taking the RB, but after this weekend, I think Either of those WR's would be a better option than Gurley in a PPR setting.

[AS3Dynasty]

For perspective, last year Aaron Donald ran the 40 only 7 hundredths of a second slower than Abdullah and was 80lbs heavier.

There's no way to spin his 40 time in a good light, and if he tries to put on any weight to add power, he's going to get even slower. It would have to be a perfect situation (maybe Dallas) to work.

Abdullah plays fast, but doesn't have elite straight line speed. His 40 time confirms that. But we already knew that going in. It's lifted by the fact that he has really good change of direction speed and burst/explosion.

I don't think I've noticed anybody mention the fact that Abdullah had the best Vert, Broad, 3-cone and 20-yard of any of the RB's, and was only bested in the 60-yard shuttle by Gordon and Ajayi. Everybody seems to be stuck on his 40 time. We knew he wasn't a burner, but people WANTED him to be because of his size.

[Csl312]

FO's Speed Score has a .45 correlation with NFL RB success. That's a much better number than you'll get from any other piece of combine data, and it comes from an intuitive study that's repeatable and testable. Moreso, when the formula misses, it usually misses on a specific type of RB (the complete back who lands in a great offense) meaning that you know when to be less concerned about a poor score. I rarely put too much stock in statisical analysis of college prospects because there's so much damn noise, but speed score is gold. If anybody could show me a way of evaluating college players that correlates better than .45, I'd be absolutely thrilled. Thrilled.

And yes, this was a bad year for RBs at the combine. Gordon's score wasn't bad, so it won't hurt his stock, but I think a lot of us wanted him to get under 4.5 flat. Karlos Williams obviously killed it, and Ajayi didn't really let me down. For what was seen as a really talented RB group, I don't see how anybody can spin the combine as anything other than a disappointment.

I'm not a stats expert but I would consider a 0.45 correlation coefficient to be pretty poor. Doesn't this just tell us that there is no number that is going to let us predict NFL success? I agree with what contrarion and gorisih are saying, being a agile/fast/whatever is not going to mean you are a good football player. So-called workout warriors pop up every year, and every year people are taken in by the 'potential' but they don't usually pan out. The combine is more useful as a way to confirm that there is some baseline level of athletic competence IMO, not really a good way to separate out prospects unless there are large disparities.

On the subject of 40 times and the earlier discussion about the 10m split - I just don't see how this is relevant to a real game situation. A RB is never getting the ball when just coming out of starting blocks so the 10m split is mostly useless IMO. If you wanted a more useful measure then maybe these guys should get a 2-3 step head start before they cross the starting line of their 40m dash. And honestly I don't put much stock in what is happening at the end of a 40 either. If a guy is accelerating or holding steady (slowing would be bad). NFL players rarely have to run much more than 40 yards at a time, and if an RB is doing so, there is probably nobody catching him at that point anyway.

[dlf_jules]

For dynasty drafters, a metric's predictive power in isolation doesn't matter. That's because we already have a very good predictor: NFL draft position. For me to care about a metric, I need to be shown that it has predictive power even if we control for NFL draft position. And almost none of the metrics thrown around here and elsewhere even purport to have such predictive power.

[Cameron Giles]

For perspective, last year Aaron Donald ran the 40 only 7 hundredths of a second slower than Abdullah and was 80lbs heavier.

There's no way to spin his 40 time in a good light, and if he tries to put on any weight to add power, he's going to get even slower. It would have to be a perfect situation (maybe Dallas) to work.

Abdullah plays fast, but doesn't have elite straight line speed. His 40 time confirms that. But we already knew that going in. It's lifted by the fact that he has really good change of direction speed and burst/explosion.

I don't think I've noticed anybody mention the fact that Abdullah had the best Vert, Broad, 3-cone and 20-yard of any of the RB's, and was only bested in the 60-yard shuttle by Gordon and Ajayi. Everybody seems to be stuck on his 40 time. We knew he wasn't a burner, but people WANTED him to be because of his size.

Yep, overall he had a pretty good combine. He looks like a real life 2nd rounder. He seems to have a high floor. At the very least I'm expecting a high end COP back.

[TommyL31]

For dynasty drafters, a metric's predictive power in isolation doesn't matter. That's because we already have a very good predictor: NFL draft position. For me to care about a metric, I need to be shown that it has predictive power even if we control for NFL draft position. And almost none of the metrics thrown around here and elsewhere even purport to have such predictive power.

Completely agree. I saw the Bill Barnwell article posted a few pages back but that was not only 7 years or so old but it also wasn't very verbose about its methods and seemed to focus on one draft class. If there are any that are a bit more explanatory about their research methods and a bit more current I would be very interested.

In addition to NFL draft position I would also think some sort of adjustment for college production/offensive share (maybe even adjusted for strength of schedule) would help adjust for the workout warriors.

Other than the Barnwell article the only things I've been able to find so far are:
2013 article focusing on WR combine predictors. Makes the case that 10 yard split and vertical jump are better predictors than speed score for WRs and also that bench press, amusingly enough, has some predictive power for WR success. The writer was pretty surprised at this (as I would be too) and suggested that since speed is more natural than bench pressing it may serve as some proxy for discipline and work ethic.

2012 article that goes through all the positions but fairly quickly

One of my favorites so far is this one published yesterday: Combine survivor effect basically stating that there's essentially a huge selection bias of people invited to the combine.

[Telperion]

On Ameer Abdullah, RB's since 2006 with weight between 195 & 215 & 40 time > 4.59:

- Ka'deem Carey
- Montee Ball
- Stepfan Taylor
- Graig Cooper
- Shaun Draughn
- Mark Ingram
- Evan Royster
- Jacquizz Rodgers
- Ryan Williams
- Lance Dunbar

Those are just the recognizable names, there are even more I didn't even remember. If you want to ignore his 40 time and go by his other metrics and "eye test" be my guest, but his size / speed profile has not succeeded in the NFL in at minimum a decade, as far as I went back. There are also some high profile busts and career plodders in that list.

[Cameron Giles]

On Ameer Abdullah, RB's since 2006 with weight between 195 & 215 & 40 time > 4.59:

- Ka'deem Carey
- Montee Ball
- Stepfan Taylor
- Graig Cooper
- Shaun Draughn
- Mark Ingram
- Evan Royster
- Jacquizz Rodgers
- Ryan Williams
- Lance Dunbar

Those are just the recognizable names, there are even more I didn't even remember. If you want to ignore his 40 time and go by his other metrics and "eye test" be my guest, but his size / speed profile has not succeeded in the NFL in at minimum a decade, as far as I went back. There are also some high profile busts and career plodders in that list.

What about the other drills that he did superb in?

[Telperion]

That's what comes down to owner preference. Do you feel that the other drills outweigh the 40 and he's the exception, or do you go with history. For me, I'm going to draft a proven physical profile over a failed one every time.

Getting into the "well this one's the exception" mindset is a trap and you'll lose more often than you win.

[Jason3123]

For dynasty drafters, a metric's predictive power in isolation doesn't matter. That's because we already have a very good predictor: NFL draft position. For me to care about a metric, I need to be shown that it has predictive power even if we control for NFL draft position. And almost none of the metrics thrown around here and elsewhere even purport to have such predictive power.

Completely agree. I saw the Bill Barnwell article posted a few pages back but that was not only 7 years or so old but it also wasn't very verbose about its methods and seemed to focus on one draft class. If there are any that are a bit more explanatory about their research methods and a bit more current I would be very interested.

In addition to NFL draft position I would also think some sort of adjustment for college production/offensive share (maybe even adjusted for strength of schedule) would help adjust for the workout warriors.

Other than the Barnwell article the only things I've been able to find so far are:
2013 article focusing on WR combine predictors. Makes the case that 10 yard split and vertical jump are better predictors than speed score for WRs and also that bench press, amusingly enough, has some predictive power for WR success. The writer was pretty surprised at this (as I would be too) and suggested that since speed is more natural than bench pressing it may serve as some proxy for discipline and work ethic.

2012 article that goes through all the positions but fairly quickly

One of my favorites so far is this one published yesterday: Combine survivor effect basically stating that there's essentially a huge selection bias of people invited to the combine.

I also found more relevant articles pertaining to Barnwells/Football Outsiders speed score, and they pointed out big time steals of DeMarco Murray, Knile Davis, however for the 2014 class they pointed out Jerick McKinnon, Damien Williams, Tyler Gaffney, and George Atkinson. Hard to put a ton of stock into the latter 3 do to draft position/getting cut/practice squad/etc...

It might be more relevant for some of the bigger backs taken in rounds 2-4. Where you see a guy like Bell, DeMarco, Knile get drafted in the 2nd or 3rd rounds, you might be better off taken them early in a rookie draft over some other backs who go higher.