Writers of Pro Football Prospectus 2008

21 Jul 2010

The Impact Of First-Round OL

Florida Danny consistently does fantastic work at Niners Nation. His current look into the effects of drafting offensive linemen in the first round is ongoing, and worth a read.

Posted by: Bill Barnwell on 21 Jul 2010

12 comments, Last at 25 Jul 2010, 6:53pm by buckturgidson

Comments

1
by JimZipCode :: Thu, 07/22/2010 - 3:05am

References you guys enough, anyway.
(First?)

2
by Mr Shush :: Thu, 07/22/2010 - 7:14am

It's a very interesting discussion, but the sample size issues are a huge problem: on the one hand, even lumping every OL position in together doesn't really give as big a sample as you would like; on the other, it's far from clear that it's reasonable to expect players at different OL positions to have the same sorts of effects. Surely a left tackle ought to be expected to have more impact on pass blocking than a right tackle, who in turn should have more impact than a guard? And since left tackles are more valuable than guards, we should expect guards drafted at a given spot to be better guards than left tackles drafted there are left tackles, which means that the above scale ought to work in reverse for run-blocking: guards most, left tackles least. And of course what's important here is their NFL position, not their college position . . . and that get's complicated because of course for many players it changes.

We also need to remember that we are not comparing teams that drafted offensive linemen to teams that didn't draft anyone (Redskins notwithstanding). The number of first round picks exercised by the teams we're considering will be only very slightly higher than the league average over the period. So, for one thing, if teams that draft offensive linemen perform at roughly the same level going forwards as teams that don't, that's probably an indication that the market is pretty efficient when it comes to deciding when to draft OLs. Even if this weren't the case, we should always remember that talent identification is vastly more important than strategy. By far the most important question about Iupati and Davis is the one we won't know the answer to for at least a couple of years: how good are they?

3
by tuluse :: Thu, 07/22/2010 - 9:20am

which means that the above scale ought to work in reverse for run-blocking: guards most, left tackles least.

Wait, what?

You made a huge leap of logic to get there.

8
by Mr Shush :: Fri, 07/23/2010 - 8:02am

I made an assumption, namely that the drafting process is more-or-less efficient with regard to where players at those positions are drafted. In other words, that the overall marginal value of a left tackle drafted at position n is on average the same as the overall marginal value of a guard drafted at position n (and I could have been clearer about the fact that I was talking specifically about players drafted at the same spot). I think we can all agree that a higher proportion of the left tackle's marginal value will be the result of his play in the passing game than is the case for the guard's. Discounting special teams value, it therefore follows logically that the guard's marginal value in the run game is higher.

That doesn't mean that guard play is more important to the run game than left tackle play, only that a guard will have more impact on the run game than a left tackle of equivalent value. Left tackles are more valuable than guards, overall, and may well therefore have more impact on the run game than guards, overall.

4
by ChicagoRaider :: Thu, 07/22/2010 - 10:27am

It would be interesting to note the value of the replaced players. Were they league-average? This has implications when you consider that all choices are made against various alternative courses.

Maybe other positions have better first round rewards, or better high first round rewards. Or for the window you are interested in you might be better off getting a league-average free agent.

5
by piratefreedom (not verified) :: Thu, 07/22/2010 - 3:21pm

if you draft a first round OL you didn't draft another position.
It would be more interesting to me to see an analysis that was position wide.

what positions are over drafted and which are under drafted regarding increase in wins over the course of X years, especially when attempting to include the usual regression to mean.

I would expect WRs and QBs to be over drafted because of all the busts and expecting a low impact from even a good WR.
RB would peak early but again it is a "glamor" position that might over attract teams desperate to sell season tickets.

OL would likely be under drafted by these ticket desperate teams because casual fans rarely know their names.

However the Jags would be an example of a team that drafted linemen recently when desperate to sell tickets so without a large scale analysis those theories are just talk.

6
by BigDerf :: Thu, 07/22/2010 - 5:42pm

I just feel like that was a lot of statistical analysis to tell me that drafting O-Linemen high can be good... assuming the linemen themselves are actually good.

7
by tuluse :: Thu, 07/22/2010 - 5:47pm

Draft analysis usually comes down to "pick good players."

If there was a magic formula, teams would be using it.

9
by Bobman :: Fri, 07/23/2010 - 5:18pm

Bill Barnwell,

Your comment after the article, about the world ending in 2012 ALSO left out one key ingredient: SWAGGER.

If you draft a 1st round OL with it, he will make a 2012 season happen, simply by force of his swagger. Example: Recently, when I was feeling really confident but was stopped at a red light, I concentrated really hard and it turned green. Took like, I don't know, 45 seconds. Now THAT is the power of swagger, my friend. NO sitting at red lights for guys like me. In fact, it's pretty sunny and hot in Seattle right now. I'll turn my swagger up to eleven to cool things down and in about six hours, it'll be dark and cool, with maybe a light breeze off the lake.

Swagger, baby.

10
by buckturgidson (not verified) :: Sun, 07/25/2010 - 10:33am

There was a lot of B.S. in there . " a team is awarded 6 points for crossing the opponent's goal line in possesion of the ball " . Really ? wow ! . I can't believe I wasted as much time as I did on this B.S. This guy should use about a third the words he does .

11
by jimbohead :: Sun, 07/25/2010 - 2:43pm

Perhaps he was a little verbose, but the point he was making is sound. Statistical models are best when focused on describing how things actually work in practice. That is to say, running multiple linear regression after multiple linear regression to find factors associated with winning is less valuable than describing a functional model of how something should contribute to winning, then testing the theory. Since the site he's posting on is a fansite, rather than a stats site like FO, it makes sense that he gets wordy in explanations, rather than assume everyone will understand SEM analysis from a summary of the model.

Context my friend. Learn it.

12
by buckturgidson (not verified) :: Sun, 07/25/2010 - 6:53pm

I don't care what the context is , there's a lot of B.S. in there .

Post new comment

The content of this field is kept private and will not be shown publicly.
  • Web page addresses and e-mail addresses turn into links automatically.
  • Allowed HTML tags: <a> <em> <strong> <cite> <code> <ul> <ol> <li> <dl> <dt> <dd>
  • Lines and paragraphs break automatically.

More information about formatting options

CAPTCHA
To skip this, please log in.