The NFL Draft - A Little Less Crapshoot, A Little More Blackjack

Nice work TJ. Rec’d.

I find it amusing that PFW states that a correlation does exist between draft quality and winning without actually checking to see if a statistical correlation exists. I guess they suspected it so much that they didn’t even do the math to see if it was true or not.

I was surprised to see 0 pro bowl percentage. However, when I checked, Palmer was a 2003 pick and OchoCinco was a 2001 pick, as was TJ Houseofvanilla.

It checks out.

Also, a note on the “still on the team percentage”. I look at a team like the packers. Just a few years ago, they were touted as one of the oldest teams in the NFL. Because of that, the younger players they drafted were basically pigeon holed into those positions because their older competition (hello Favre) had moved on. As more years go by, they’ll have more of an age variance, rather than just old guys and young guys.

I wonder what kind of impact would happen if overall team age was taken into account. Older teams trying to get younger would have a more immediate rookie holdover than a team with already young players, simply because the team is counting on the new guys they drafted to come in when the vets, inevitably, retire.

In the case of the Denver Broncos, if we look from last year to now, we’ve gotten more veteran players. We brought in free agents who were older and didn’t build through the draft as much as we did through free agency. So, there’s another factor to consider. Building through free agency vs building through the draft.

very interesting, thanks and rec’d.

Keep up the interesting work Bro!

I am going to cover this in a few weeks, but I can’t help but point one thing out.

100% of the Patriots 1st round selections from 2001-2008 started or start for the team. They have a horrible draft record in late rounds, but rounds 1 and 2 they hit pay dirt like no other team in the league.

So in a way, drafting does equal winning…it just depends on where you hit blackjack and in what round on a consistent basis. ;-)

Great post TJ! Does Ashley Lelie count as a “starter” in this chart? lol I’m glad we have McX in charge now….though I will be watching the Redskins from now on just to see if Shanny is able to break his old habits.

Blackjack is a good comparison to the draft. Every year the prospects are hyped so much that we all are guilty of thinking they are going to “save” the team. Aside from a few, this is not the case.
One thing I would like to note is even though NE has the lowest % of those still on the roster they have turned them into present draft picks this year.

He puts all of money in roulette on double-zero.

Thanks for the data TJ. A higher pro-bowl # would be nice to see, but winning teams get more attention for votes. If TEN would’ve won a few more games last year, there’s no way not one of their O-lineman wouldn’t have been snubbed. Sitting in the .500 range gets “mehs” for how fans perceive a players “talent”. I have a personal bias against pro-bowl player voting so I’ll just stop here.

signed,
Phaedrus

I loved this. It covers something I’ve been thinking about but hadn’t seen put into actual numbers. It was obvious to most of us that the Broncos had not drafted as well as they should have in the last few years of Shanny’s reign, focusing too little on that and too much on free agency. It’s also true that teams can get by without great success here (the Patriots, at least for their run; it may be coming back to haunt them now).

One question: How did you calculate average draft position? That seems an odd number to me. Do you take all their draft positions, round by round, and divide by the number of picks they had?

Anyway, great stuff TJ, thanks.

starter, probowler, allpro… all of these ideas leave so much out of consideration.

When I was doing an anlysis to find best locations for impact players at particular positions, I constrained myself to looking at two-deeps, but that left out the “comparison to peers” element, since somone might be on a two-deep despite not being truly qualified (such as if the GM wasn’t doing a good job of cycling in competition at each spot).

A useful metric would be a DVOA metric for individual snaps taken in context of all available snaps… that should translate from one player to the next…

Man you are it for interesting posts on the site. You sir have supplanted HT as my personal favorite staff guy. Feel honored. Rec’d all day, twice on Sundays my friend.

This is what I was trying to stumble towards with a fan post of 1st round QB picks a while back. The draft isn’t totally random, just hard to predict and a bit of a gamble.

Speaking of gambles, that might be the reason for the lack of correlation between wins and draft. If you use free agency to fill the needs on the team, you can gamble a bit more in the draft, going after players with higher risk/reward ratios. I fully believe that was what Shanahan was trying to do, but he was failing in both bringing in the best FAs as well as hitting on a 15 too many times.

One thing you might add to the average pick position is the number of total picks for the period. My suspicion is that the aggregation on quality might be hiding a trend in quantity. As an extreme example, think of Ditka trading his entire draft for Ricky Williams — if that event were in the time span studied it would certainly skew the numbers for the Saints.

And rec’d, of course.

I’d like to see a column that in some way quantifies how stable the front office/coaching staff has been over that period. Obviously, part of the reason that the Broncos have such a low percentage of players still on the roster is that they brought in new GM and HC, who purged many of Shanny’s players (some very deservedly, some others, maybe not as much). Some of the other teams with low numbers (Detroit, Cleveland, Tampa) are in the same boat, though as been mentioned, Cincy and NE, which I think have had had pretty stable front offices over that period, have pretty low numbers there, too. It’s hard to remember how many other teams change GMs and Coaches that often, but I wonder how strong the correlation is between front office changes (especially changing coach and GM at the same time) and the number of players from the draft still left on the team – it certainly seems to me that there’s a lot that goes into that number that may not have a lot to do with the actual drafts.

What about the number of picks that they had? If teams were trading draft picks for free agents wouldn’t that skew things? If I traded away most of my draft picks, and only picked a few ‘sure things’ out of the draft, then that would make the GM look great.

I think the main point is: there is way more to the game then the draft.

If free agency coul be brought in, as well as any other ways that teams add players to their rosters (undrafted pickups?), I think you could start seeing a strong correlation to winning… which doesn’t equal winning the super bowl,(where’s that in all this?).
Look at the Saint’s numbers, not exactly impressive.

Great work.

Ouch! The Chargers’ draft picks have a higher chance of becoming pro-bowlers than ours have of being starters. Ouch, ouch!!

The Jets draft well and these numbers are only through ‘08, didn’t they have a killer draft last year, too?

NE and IND are hard to put a finger on. As you mentioned, one draft can scew things. My initial thought was that these teams are so stacked with backups, being perrinially great, that they may have to flip players players more often than others. But if that were the case, I would think their pro-bowl percentage would be a little higher.

I’m gonna have to take some time after work to digest it all properly. :)

Awesome post tj, I always enjoy your writing and analysis. Writing this from my phone, so trying for brevity…

The only issue I have with your analysis is that you attempted to correlate winning percentage with draft statistics taken from the same period (2004-2008). But, if is a given that draft picks take years to develop, then shouldn’t drafting success correlate with wins in later years? In other words, instead of analysing the wins and draft data from the same years (2004-2008) shouldn’t the draft data from 2004-2008 correlate with wins from 2006-2010? This might explain the lack of correlation you found between drafting and winning- it might be simply delayed.

Anecdotally, of course. But it is certainly possible that NE’s success in 2004-2008 was a result of their drafting in 2001-2005, and that their recent decline was in partial response to the poor drafting of 2004-2008

I always felt like any draft that didn’t yield about 1.5 starters was a bust. I was very disappointed that Robert Ayers saw so little playing time last year. It would’ve been nice to pull two insta-starters in the first round. I guess I’m spoiled from how well Ryan Clady worked out.

darn subjective.

2 things:
Basically, SHanny sucked as GM and talent evaluator and therefore we are sucking now.
The system that New England runs obviosuly is a winner, and Belichick is a genius, which gives me hope that McD has a great system and is a better evaluator.

I don’t think we’ll have to worry about his draft abilities too much, because afterall it does take a “blend” of players to make it work. Mentors for students, now that is an awesome concept. Anyway, this is an area I never dreamed you would address, even though I have thought about it many times…Thanks dude, rec’d. Oh, and before I go, it will be interesting to me to watch Shannan’s drafts too, for the next few years.

Just read on Porkchops blog that Shanny’s 2006 draft class was the best draft in the decade. Interesting

I was just perusing your site and came upon this fanpost. Put up the graph on the Giants blog (Big Blue View), obviously making sure you were credited and linked this fanpost. Sick shizz, definitely rec’d