A direct correlation in drafting a 1st round OT and immediate offensive improvement

[Oldfan]

....so I don't see the issue with posts made to counter the OP. Especially, since he used "DIRECT" and "IMMEDIATE" which is a pretty strong (and somewhat erroneous or overly general) couple of assertions that scream for refutation.

I realize now why you have a problem with the OP and I don't.

To me, the OP represents quick and dirty stats that are so one-sided on a small sample size that they would support a research project for an NFL team. The words "direct" and "immediate" were only descriptions of results to me.

You are taking the OP as a broad claim needing more proof.

[Oldfan]

I think your list proves the value of getting a good QB or RB more than simply drafting an OL. Just about each team you mentioned upgraded their QB and/or RB positions or had a young QB get more seasoning.

You listed three things -- my guess is that 28 of the 32 teams do at least one of the three in the offseason. How many spend high picks on a OT?

[JoeSkins]

Excellent leg work, but the sample is kind of small. Additionally, the kind of upgrade a rookie tackle would be over current personnel isn't quite as large as the teams listed got from their respective players.

Drafting an offensive lineman would definitely help, but not immediately.

[Veretax]

even if this is true its not always right to draft a OT

[Peregrine]

What I think this likely is an indicator of is not so much that if you draft an OT in the 1st round you will have success, but that these teams made it an emphasis to improve their offensive line as a whole, and they had success.

Likely they did not just draft one offensive linemen and say "we are done". The probably saw it as important, and then reaped the benefits of adding a couple of good linemen.

[Riggo-toni]

Two words: Robert Gallery.

Raiders passed up Phillip Rivers, Ben Roethlisberger, Larry Fitzgerald, KWII, Sean Taylor....

[Streater101]

Note that the teams with the most success took a tackle high in the 1st, while the teams with little to no success took theirs late. Thats why trading back would be foolish if we still have a shot at the big 4. Talent like Samuel, Clady, Otah, Long, Pace, etc... are very(very) hard to pick up outside of the top 15. Guards, halfbacks, WRs, are guys who can. You don't waste an opportunity like this.

[Streater101]

Two words: Robert Gallery.

Raiders passed up Phillip Rivers, Ben Roethlisberger, Larry Fitzgerald, KWII, Sean Taylor....

Good point, lets ignore the overwhelming evidence to the contrary to point out on LT who didn't pan out in the city where careers goto die.

[The Batman]

Stellar analysis. However, interesting how you choose not to note the QB changes at many of these offenses as well.

Let's take a look, shall we?

#1 Miami Dolphins - Jake Long (28st ranked offense ➞ 12th ranked offense) (16 starts) QB CHANGE - Cleo Lemon to Chad Pennington

#15 Kansas City Chiefs - Branden Albert (31 ➞ 24) (15 starts) QB CHANGE - Damon Huard to Tyler Thigpen

#19 Carolina Panthers - Jeff Otah (29 ➞ 10) (12 starts) QB CHANGE - Jake Delhomme/Matt Moore/Committee to Jake Delhomme (Delhomme started 3 games in 2007 due to injury, and 16 in 2008)

#21 Atlanta Falcons - Sam Baker (23 ➞ 6) (5 starts) QB CHANGE - Chris Redman/Joey Harrington to Matt Ryan

#3 Cleveland Browns - #3 OT Joe Thomas (31 ➞ 8) (16 starts) QB CHANGE - Charlie Frye to Derek Anderson (Anderson started 5 games in 2006 and 15 in 2007)

#5 Arizona Cardinals - #5 OT Levi Brown (18 ➞ 12) (11 starts) QB CHANGE - Matt Leinart to Kurt Warner (Warner started 6 games in 2006 and 14 in 2007)

#4 New York Jets - OT D'Brickashaw Ferguson (31 ➞ 25) (16 starts) QB CHANGE - Brooks Bollinger/Vinny Testaverde to Chad Pennington

So you see, while your analysis was very involved and thorough (and major props to you on that), it forgot to include a pretty significant confounding variable. Of the 9 teams you mentioned that showed improvement, 7 had significant turnover at the QB position as well. 78% QB change rate is certainly important at offensive improvement - almost enough to say that it is nearly a constant factor between the teams as well, and muddles the relationship between offensive line drafting and offensive improvement.

So there you have it, kids. Yet another example, unfortunately, of correlation not implying causation.

Maybe, if we draft offensive line high, AND have a shakeup at QB like 78% of these teams did, we can have a major offensive improvement too!

[Oldfan]

So there you have it, kids. Yet another example, unfortunately, of correlation not implying causation.

I don't think you've shown that the OT factor wasn't a cause in the improvement. In fact, you imply correctly that it very well might be in saying:

Maybe, if we draft offensive line high, AND have a shakeup at QB like 78% of these teams did, we can have a major offensive improvement too!

Logically, I would expect the most offensive improvement from a positive change in the (1) offensive scheme, (2) quarterback, (3) blindside tackle -- in that order.

[The Batman]

I don't think you've shown that the OT factor wasn't a cause in the improvement. In fact, you imply correctly that it very well might be in saying:

Logically, I would expect the most offensive improvement from a positive change in the (1) offensive scheme, (2) quarterback, (3) blindside tackle -- in that order.

You're right. If you reread my post, I never claimed to be trying to show that OT drafting WASN'T an improvement. Simply that the OP's claim of correlation (which certainly is valid) doesn't imply that it is the sole CAUSE of the improvement. Hence, correlation does not imply causation.

[Oldfan]

You're right. If you reread my post, I never claimed to be trying to show that OT drafting WASN'T an improvement. Simply that the OP's claim of correlation (which certainly is valid) doesn't imply that it is the sole CAUSE of the improvement. Hence, correlation does not imply causation.

Maybe I'll learn something here.

I never understood the phrase "correlation doesn't imply causation" to be restricted to sole causes. In my mind, a factor that contributes to the effect is still a cause.

I see the principle applying clearly in this example: There is a direct correlation between kneel downs and winning football games, but it's obvious that having your QB kneel down more often wouldn't be a good strategy in winning more games.

A less obvious example is the high correlation between turnovers and wins. Turnovers are the effects of superior play, just as points on the scoreboard are the effects of superior play. I've seen a couple of studies that put the value of the average turnover at four points. However, coaches seeing that high correlation, and thinking of the evil turnover as a cause, can go overboard in trying to prevent turnovers at the expense of offensive production.

I don't think the OP intended to establish the OT as a sole cause and rule out other factors making their contributions to the effects.

[Laron Burgundy]

1) How did you logically rule out the possibility that the stats support both the QB and OT factors? That other factors can be argued to contribute in every situation is a given. We didn't need your stats to prove that.

I'm not saying that a new OT doesn't help, but the fact that almost every team had a significant change at qb or rb certainly leaves me to believe that the OP's original logic is deeply flawed. If the OT isn't the only change to a team's offense, then you can't correlate that team's offensive success to that OT's drafting. Certainly that OT may have helped the new qb play better or the new rb run better, but that is impossible to know for sure. Seeing as how our qb and rb are both healthy and unlikely to be replaced, I wouldn't count on significant offensive improvement specifically because of any OT we may draft. If we're going to make any big jumps in production it will be because Portis won't get dinged up as much or our young receivers will contribute more.

2) There is far more subjectivity built into your nine-team sample than the OP has in his 12-team sample. You decided which QB was new, which healthy, and which RB or QB to count where multiple players were involved.

My sample showed that defensive tackles help offenses. Defensive tackles don't play on offense. It was supposed to be absurd. It is no less objective than his sample. I took teams from the same 3 drafts and they all showed offensive improvement the year they drafted defensive tackles.

That's as much of a direct correlation as drafting an OT. 3 out of his 12 teams showed no significant change or got significantly worse. 1 out of mine showed no significant change. That's 9 teams for him compared to 8 teams for me, and mind you that means it's a much higher probability that if we draft a defensive tackle that our offense will improve. You have a 75% of improving if you draft an OT and an 89% chance of improving if you draft a DT. That's cold hard facts and voodoo math. Numbers never lie.

As far as me deciding on new qb's and healthiness and all that jazz. Look it up yourself. If a different qb threw the ball more, that's a new qb. That's not subjective.

Brooks Bollinger-> Pennington UPGRADE

Leinart-> Warner UPGRADE

Charlie Frye-> Derek Anderson UPGRADE

Harrington-> Matt Ryan UPGRADE

Huard-> Thigpen UPGRADE

Cleo Lemon-> Pennington UPGRADE

Do you want to argue that the new qbs weren't better than the old ones? The only argument I can see is Huard to Thigpen. Huard had a great 2006 but the past two years he's shown why in a 12 year career he has never started for a whole year. If a good player missed half a season or more, that's a significant loss. When focal points of an offense like Ronnie Brown, Larry Johnson, or Andre Johnson or starting qb's like Delhomme and Schaub get hurt and miss significant time, that's going to put a damper on the offense. I didn't discount people who were dinged up but still played, I only counted those who were hurt enough to miss significant play time. That is why I noted Denver as an exception, because Cutler wasn't healthy but he still played and because their new rb's weren't exactly significant upgrades to the year before.

[The Batman]

Maybe I'll learn something here.

I never understood the phrase "correlation doesn't imply causation" to be restricted to sole causes. In my mind, a factor that contributes to the effect is still a cause.

I see the principle applying clearly in this example: There is a direct correlation between kneel downs and winning football games, but it's obvious that having your QB kneel down more often wouldn't be a good strategy in winning more games.

A less obvious example is the high correlation between turnovers and wins. Turnovers are the effects of superior play, just as points on the scoreboard are the effects of superior play. I've seen a couple of studies that put the value of the average turnover at four points. However, coaches seeing that high correlation, and thinking of the evil turnover as a cause, can go overboard in trying to prevent turnovers at the expense of offensive production.

I don't think the OP intended to establish the OT as a sole cause and rule out other factors making their contributions to the effects.

To me, the OP was implying that the offensive improvement of these teams was due to their drafting OT early in the first round. Perhaps I misunderstood his intent.

However, from a statistical research perspective, the terms correlation and causation are two different things. "Causation" implies only one cause. There can be correlations between many things - QB switches, OT drafting, and offensive improvement, in this example. But the correlation between OT drafting and offensive improvement is NOT a causational relationship. It cannot be argued that drafting OT CAUSES offensive improvement, as there are too many confounding variables involved (i.e. QB changes, new RBs, other changes on the line, new offensive schemes, easier schedules, etc.) So while there may be a strong correlation between the two (and there is), it does not even begin to mean that because you draft an OT, your offense will improve, which was my point.

[Oldfan]

I'm not saying that a new OT doesn't help, but the fact that almost every team had a significant change at qb or rb certainly leaves me to believe that the OP's original logic is deeply flawed. If the OT isn't the only change to a team's offense, then you can't correlate that team's offensive success to that OT's drafting.

I think the problem here is that you and Batman think that you have disproven a claim that the OP never made.

He showed that there was a direct correlation. He never claimed that the the OT factor was the sole cause of the improvement effects. In fact, he said this:

Post #1 --- Obviously, this correlation doesn't take into account all of the other offseason management/coaching developments that occurred with each franchise. However, the one constant remains: they all drafted an offensive tackle with their first round selection.

By showing ANOTHER correlation (the QB factor), you do nothing to challenge the data offered in the OP since the QB and the OT factors can, and probably do, interact in causing the improvement.

[Laron Burgundy]

I think the problem here is that you and Batman think that you have disproven a claim that the OP never made.

He showed that there was a direct correlation. He never claimed that the the OT factor was the sole cause of the improvement effects.

The problem with saying there is a direct correlation but it may not be the sole cause is that drafting a 1st round OT may not be a cause at all. The point of this thread was to prove that drafting an OT helped the offense immediately but that is not proven here.

The NFL is evolving, and more and more teams are seeing the benefit of drafting an offensive tackle in the first round of the draft. Though certainly not the "sexy" pick, this method seems to indicate a very consistent level of immediate improvement to the offense.

Correlation does not mean causation. This paragraph from the OP would imply that that the result of drafting an OT in the first round is immediate offensive improvement. There's a correlation, sure, but there is a correlation that drafting a DT immediately improves the offense too. Neither one implies causation. Therefor we can't even claim that drafting an OT was a cause for any of the improvements.

To note, I'm not arguing that we shouldn't draft an OT; I just don't like faulty logic.

[Oldfan]

Batman: However, from a statistical research perspective, the terms correlation and causation are two different things.

I know that.

"Causation" implies only one cause.

This doesn't make sense to me. If it were true, 98% of all statistics in team sports would be useless since there is rarely a sole cause for the effects observed.

It cannot be argued that drafting OT CAUSES offensive improvement, as there are too many confounding variables involved.

It cannot be argued that it is a sole cause, but no one was doing that. It can be argued that the data makes it likely to be an interacting cause.

Let's assume that a long-term study shows that:

= the new QB factor has a .75 correlation

= the new OT factor has a .65 correlation

Neither is the sole cause... neither correlation trumps the other.. this stat would tell us that the QB is the more important factor of the two

So while there may be a strong correlation between the two (and there is), it does not even begin to mean that because you draft an OT, your offense will improve, which was my point.

You're mistaken. The data confirm what we know logically, the drafting of the OT will likely improve the offense, but by how much on its own, we don't know.

[[[ghost]]]

The only time I've seen him play was this year in the Cotton Bowl, but he played very well.

Also on a side note, he has an interesting story. He was homeless and a family basically just adopted him off the streets.

He had a superb game at the Cotton Bowl. He's been playing stout all year.

The problem is, he supposedly has mental limitations. It could have been a joke, but I HEARD that his coaches use Dolls to teach him different blocking techniques and such.

Eugene Monroe wouldn't be a bad option either. I know there are a lot of Virginia Fans, so if they could shed some light, that'd be great,.

[mcarey032]

BTW, Sam Baker could have been our pick had we kept our first round pick that we sent to ATL.

[IHOPSkins]

Andre Smith, OT, Alabama

Michael Oher, OT, Ole Miss

Eugene Monroe, OT, Virginia

Jason Smith, OT, Baylor

Eben Britton, OT, Arizona

If these guys are gone by #13......

No disrespect Oldskool......

Does everybody agree (for now) this group is our top tier choices at OT for pick 13?

If there are 5 prospects at this level, that makes me very happy

[Rufus T Firefly]

Oldfan, the problem is that ALL the teams that drafted a T and got better had other upgrades, in particular at QB and RB, thru acquistion, promotion or health (Denver with a couple more games from Cutler, more important was his maturation).

In fact, the couple of teams that didn't upgrade elsewhere (Detorit and San Fran) got worse. Atlanta got better with a new QB and RB and their drafted T not playing much.

The OP's argument was that drafting a OT meant immediate improvement. Looking at the examples, that could seemingly be amended to "draft a T and get much better at QB and you'll improve. In fact, get the QB and skip the T and you'll still improve."

[Rufus T Firefly]

];6057456']He had a superb game at the Cotton Bowl. He's been playing stout all year.

The problem is' date=' he supposedly has mental limitations. It could have been a joke, but I HEARD that his coaches use Dolls to teach him different blocking techniques and such.[/quote']

That's not a joke. It's talked about in the book. Ketchup bottles, furniture, etc.

And, after 4 years of that kind of personal tutoring, the knock on him is still that he doesn't understand his assignments or have good technique. It's a pretty big red flag.

[RabidFan]

Sign Jordon Gross first then draft a stud OG or another Tackle....samuels is key but getting old and if heyer can become a stud OG or Center....we be stylin

[Laron Burgundy]

That's not a joke. It's talked about in the book. Ketchup bottles, furniture, etc.

And, after 4 years of that kind of personal tutoring, the knock on him is still that he doesn't understand his assignments or have good technique. It's a pretty big red flag.

The fact that he plays so well despite those mental limitations says something. It's not like he's playing in the MAC, he's playing in the SEC and playing well. I've seen him speak on videos and he doesn't seem like the dumbest person ever. He's no Ms. teen south carolina.

[Oldfan]

RTF: Oldfan, the problem is that ALL the teams that drafted a T and got better had other upgrades...

That isn't a problem. The only valid criticism of the OP stats is that the sample size is small. That other factors contributed is a given. The correlation for each factor is unaffected by the existence of other factors.

Let's assume that 26 factors can contribute to offensive improvement (that's probably a low number) and we had a long-term statistical study on this. We'll label those factors A-Z. Logically, I would expect the correlations to turn out something like this:

A) New and better offensive scheme 0.78

New and better quarterback 0.70

C) New and better blindside tackle 0.60

Imagine D thru Z with descending correlations