Create Account

Or log in with Facebook


Log in

Or log in with Facebook

How Much Should The Packers Spend On Each Position

By Category

How Much Should The Packers Spend On Each Position

“Do what you can, with what you have, where you are” – Theodore Roosevelt


In the NFL, doing what you can with what you have means taking full advantage of the salary cap.  The NFL has one of the strictest salary caps in professional sports with both a salary cap and salary floor and as a result league parity is predominant as most teams hover around .500 given enough games.  Teams who are good year in and year out, notably the Patriots, Ravens and Packers, while vastly different in scheme, personnel and coaching, all are fantastic at managing their salary cap. 

With that in mind, what’s the best way to make use of the $143.28 million that each teams has at its disposal?  Is it to pay whatever it takes to secure a franchise quarterback or is it to field a top-flight defense to stymie franchise quarterbacks?  Should teams pay more for running backs or wide receivers?  Should offense or defense take up more resources?  How much should teams really pay kickers? (Just kidding, again I don’t care about kickers)

To answer these questions, I gathered each teams salary cap allocation for each position and correlated that with wins in the 2014 season.  On a technical note, I will be using a Pearson’s r analysis with p-value < .05 as the standard for significance (as a quick primer, a p-value of less than .05 means there is a 95% chance of the correlation being statistically significant)

Keep in mind while there is a salary floor, no teams spends 100% of its salary cap in any given season since it needs to keep a cash pool for injury replacements and in-season acquisitions, while most teams utilize most of their salary cap, some teams in certain season will spend right at the salary floor while in rebuilding mode (Reggie McKenzie has done this for a couple years with the Oakland Raiders to bring them out of salary cap hell).

Not even close

Safety (p = .8227) – Despite the rising importance of safeties, apparently it doesn’t make much sense to pay premium money for safeties as it’s essentially a coin flip whether or not they make a team better.  It is possible that free agent safeties (who typically command the biggest contracts) tend to fare poorly moving to a new team; Dashon Goldston and Jarius Byrd both pop to mind in safeties who commanded big money in free agency and didn’t produce anything on their new teams. 

Running Back (p = .8152) – Everyone in the league knows that paying running backs is a bad idea; not only do runners have typically short careers but so many talented running backs enter the league every year that it’s relatively easy to find a franchise running back even in the later rounds of the draft or in free agency.  I will say this number is probably slightly inflated due to Adrian Peterson (the highest paid running back) being placed on paid administrative leave for a domestic abuse allegation while obviously contributing 0 wins.  

Wide receiver (p = .6580) and tight end (p = .7026) – The shiny hood ornament theory seems to apply to pass catchers in terms of salary allocation.  Pass catchers are likely too dependent on their quarterbacks to really make much sense to pay on their own.  To put it another way, teams would rather prefer to overpay the quarterback and underpay the wide receiver than the other way round (Unless you are the Lions where you overpay for both quarterback and wide receiver).  

Close but no cigar

Offensive Line (p = .1249) and defensive line (p = .2641) – Linemen (especially offensive linemen) are considered some of the safest investments in the NFL and it appears as if the NFL has a decent grasp of their true value.  Another thing to consider is that linemen don’t have many stats to rely on when it comes to negotiations which means teams likely have more leverage when signing players.  While Aaron Rodgers can say he’s thrown for so many amount of touchdowns, with so many amount of yards while taking only so many sacks and interceptions and rushing for so many yards, Bryan Bulaga can only say that he’s given up so many sacks on his side. 

Quarterback (p = .1786) – Ironically the most important position in football does not have a correlation between salary and number of wins.  This might go back to the catch-22 of a superstar quarterback argument, which states that the quarterback market is so overinflated at the moment that almost no quarterback can actually perform to his contract.  Again, Jay Cutler is a prime example of an average quarterback being paid exorbitant amounts of money simply because he’s a quarterback. 

Actually statistically significant

Linebackers (r = .3899 p = .0274) – Paying linebackers more seems to result in more wins.  Case in point, Justin Houston just signed a 6-year $101 million contract.  One interesting point is that linebacker salaries differ drastically based on whether a team is playing in a 4-3 or 3-4 alignment, while I haven’t broken down the numbers it would be interesting to see if paying 3-4 linebackers (who typically make more than their 4-3 counterparts) is a better use of money than paying 4-3 linebackers. 

Cornerbacks (r = .4051 p = .0164) – Leading all positions is cornerback; apparently good cornerbacks cost a lot of money, but it’s money well spent as cornerback salary is most correlated to wins compared to any other position group.  It’s possible that cornerback salaries are most market equitable since cornerback is another position of low career lifespan, like running back, but unlike running back, cornerbacks pull in great salaries but teams seem to have an easier time cutting their losses when their cornerbacks begin to decline (case in point: Charles Woodson).       


Offense (r = -.0428 p = .8160) versus defense (r = .2757 p = .1267)

I don’t really know how to explain this and keep in mind neither p value points to significance but it is interesting defense is a lot closer to being statistically relevant and also paying more on offense might lead to less wins (again the correlation between wins and offensive spending is not statistically significant so take the previous comment with a grain of salt).  Nevertheless, with the eclipsing of offense over everything else in the NFL it is interesting that there isn’t a bigger correlation. 

So what does this all mean?  To be honest, interpretation is up to the observer and I will admit that using only one year will result in a lot of variability (In my defense I was not able to get spending break downs for every year).  I will say on a personal level this does seem to follow traditional thought that running backs and pass catchers are terrible wastes of money while it’s unexpected that linebackers and cornerbacks make the most prudent investments. 


  • Like Like
  • 0 points

Fan friendly comments only: off Comments (31) This filter will hide comments which have ratio of 5 to 1 down-vote to up-vote.

ebongreen's picture

I think this is a promising line of inquiry, but I also think this dataset is far too limited. I'd love to see whether these trends hold over multiple seasons, and any conclusions you reach would be potentially much stronger if, say, paying cornerbacks well correlated with wins over ten or twenty seasons instead of just one.

hobbes's picture

I agree, but as I said, I wasn't able to get salary cap data for multiple years so that definitely puts a hamper on any longitudinal analysis. Furthermore, I don't think going back 10 or 20 years would be of much benefit since the NFL is a very different business now; for instance, I bet running back salary would be a lot more important than it is now.

Idiot Fan's picture

I agree that 10 or 20 years probably wouldn't help, but even having 3 or 5 would be really helpful in this case. With one year of data, you can definitely get skewed based on who wins and loses that year. Win totals for teams can change dramatically from one year to the next, even with most of the same players.

But, as you keep telling us, you don't have the data. Too bad though, this is a very interesting analysis!

Bearmeat's picture

As usual - very intriguing Hobbes.

However, I would agree with the above poster. If data from, say, back to 2005 could be reached, I'd bet we'd have a much clearer picture.

hobbes's picture

I'm sure it's out there somewhere, but I wasn't able to find an easy source. One other caveat is that NFL contracts are so ludicrously complicated and obtuse, different sources will report different numbers (i.e. the huge mess that Donovan McNabb's Redskin's contract was), so I didn't want to combine cap numbers from various sources.

Since &#039;61's picture

Nice effort Thomas. However you and others are correct in that you need more than one year worth of data. Given the current rules which favor the passing game I think that you would see over time an improvement in the value of paying for offense especially QBs and OL. Also, probably the value of CBs will probably remain high because they have the primary responsibility for stopping an opponents passing attack. This year is a good start and you can go forward from here. Thanks, Since '61

hobbes's picture

It is interesting that while the current rules favor the passing game, which would lead to a shift in importance to QBs, OLs and CBs, you would think that the biggest gain in terms of correlation would be in fact at wide receiver and tight end (to some extent, certainly now with the "big wideout" tight end). Furthermore, you would think that safeties would pretty much follow in lock step with cornerbacks in terms of correlation, I've done previous analysis that shows that safeties are in fact more important to the defense than corners and yet for this year at least, paying big bucks for a safety was definitely a bad idea.

Since &#039;61's picture

WRs and TEs have proven to be replaceable without much impact on the passing game. For example, Rodgers has lost Driver, Jennings, J. Jones and TE Finley over recent seasons and the passing game hasn't missed a beat. Except for TE, the current WR corps may be the best Rodgers has ever had. Going back further Favre played with virtual revolving doors at WR during many seasons and he had a HOF career. However, the QB can't be effective without the OL and QBs have completion %'s higher than ever. As for safeties, again consider the turnover at the position. The Packers most experienced safety is Burnett with 5 years in the league. If you track this over 3-5 seasons it will be interesting to see if the longer term results change significantly or not. Thanks, Since '61

hobbes's picture

I wonder if this is because teams have started to use different flavors of pass catchers (deep threat/possession/red zone/slot/etc) thus making more and more players NFL-caliber, hence more available players on the market. As for QB/OL correlation, Rodgers has had epically bad offensive line play over the years but it hasn't effected his game significantly one way or the other. Andrew Luck has become one of the best quarterbacks with one some of the worst offensive lines to boot.

Thegreatreynoldo's picture

Very interesting, Thomas. As for the limited data, this is something that you can put in your personal database and do again for the next few years to get more data. The floor for paying QBs is so inflated that it skews the results. Demand is high, supply is really small. Any QB who might, just might, develop gets paid like a god. WR salaries have surprised me. We've had a bumper crop of WRs in the last few drafts - I suspect their price will drop over time. The fact is that there are a lot of adequate guys who can play RB. Since RBs are the most NFL ready to play at a high level right away, you get good results while they're on their rookie contracts. Very few of them get a 3rd contract, and most get a 2nd contract hedged with outs for the team, so cost is deflated by the rookie wage scale.

My guess is that difference makers, pretty much regardless of position, get paid and influence wins. An interesting exercise would be to compare for significance on winning the 2nd highest quintile at each position to the 4th highest quintile with the aim of determining at which positions a team can get away with below average player and still win. You would have to remove all players on their rookie contracts. Another way would be to use PFF rating rather than salaries.

hobbes's picture

I think we're already seeing a drop in wide receiver salaries. Both Thomas and Bryant tried to use Calvin's Johnson's contract as a leverage point but both got deals with Mike Wallace as the reference point instead. I think Johnson and Fitzgerald were probably the straw that broke the camel's back so to say, no matter how good a wide receiver is, it just doesn't make much economical sense to pay them that much, especially when you are a dumpster fire at quarterback.

Steve Cheez's picture

Love the effort, but the math is a bit over my head, Professor Hobbes. I think I get that the p values are the inverse of the correlation (i.e., p of .05 indicates a 95% correlation. But then you thres in r values on a couple of them and that's where you lost me.

But I guess even with my ignorance of the math, I can say I'm really happy that Mike Sherman isn't running our cap any more.

Steve Cheez's picture

And I guess spelling the word "threw" and remembering to close my parentheses are also beyond me...

4thand1's picture

What gets me is the edit feature was beyond you.....................jk

Steve Cheez's picture


hobbes's picture

Sorry for the confusion. Essentially the r value determines how related two variables are, in this case number of wins versus the salary cap of a position. An r value of 0 would indicate that there is no correlation at all (meaning there's no relationship between how much you spend in a position and how many wins you get), an r value of 1 would mean there is total correlation (i.e. the more you spend at a position the more you win) and a r value of -1 would mean there is an inverse correlation (i.e. the more you spend at a position the less you win). Numbers in between -1 and 1 indicate that there is a correlation but it isn't perfect (which is expected)

The p value is to determine significance of the r value; essentially how much faith you can put into the r value being true. One analogy might be to consider flipping two quarters and seeing if getting heads on the first coin correlated to getting heads on the second. Obviously the results of flipping the first coin has no bearing on the results of the second but if I were to flip both quarters 10 times, the r value might indicate that there was a correlation simply by chance. Now if I were to flip both coins a million times then it's highly unlikely that I would find a correlation and that's in essence what the p-value is showing. In other words while the r value in flipping 10 coins might show correlation, the p value would should we can't trust the result.

As a side note actually every position has an r-value but I didn't include it in positions that weren't statistically significant (p<.05) since in a sense they don't mean anything.

Idiot Fan's picture

One helpful way to think of the p-value (IMO) is that it is the probability of obtaining the results you are seeing by chance alone. So, for example, if you calculate that two numbers have a correlation of .5, that could be because they really have a correlation of .5, or it could be that randomness in the data just makes it seem like they have a correlation. By doing the test that produces the p-value, you are calculating the probability that there is NO true underlying correlation and the result is due to randomness. This is why smaller p-values are preferred when trying to show that there IS in fact a true underlying relationship. By choosing a cutoff value of 0.05, you're saying that you are comfortable concluding that anything with a 5% or less chance of being due to randomness does indicate a true relationship. There's no universally accepted threshold value, though 5% is the most common.

hobbes's picture

Is there any reason why p<.05 was deemed "good enough"? Is it just convention or is there some underlying stats (how this would work I don't know) that justify 95% confidence being statistically significant?

Jay Hodgson's picture

+/- 2 standard deviations under the normal curve.

Idiot Fan's picture

Correct, although that's still a matter of personal preference and/or requirements of the model. I've had models where it was important to use 1% and others where 10% worked just fine. Most people, in my experience, default to 5% and then make adjustments if necessary.

Idiot Fan's picture

The phrase "statistically significant" is only meaningful after you've chosen your standard, such as p<0.05. It is based on that standard that you are able to measure whether something is "statistically significant" or not.

Jay Hodgson's picture

Does this call for another bro hug?

hobbes's picture

At this point, I think "statistically significant" by convention implies p<.05 I know I'm very careful to only use the phrase "statistically significant" when I'v hit that CI.

MarkinMadison's picture

I hated statistics. P values make my brain hurt. What I take from this is that impact LBs and CBs matter. The ILB position has been neglected in Green Bay, and it shows in the rushing D numbers. The CB position is a bit of a calculated risk by TT this year, one that he could have reduced by overspending on one of the two corners lost in FA this off-season. We'll see how it plays out.

hobbes's picture

I wouldn't say that the article is discussing impact per say, more how much bang for your buck do you get by spending at certain positions. Last year at least, it made sense to pay LBs and CBs. In terms of raw impact, I think everyone agrees that quarterbacks reign supreme but at this point in time quarterback salaries are crazy cause teams are desperate to find "the guy" that they are willing to throw money at just about anyone.

Idiot Fan's picture

I'm curious to know what the teams were that won a lot of games and spent on LBs. Is it possible, with Hawk's and Jones's salaries that we were one of them?

Thegreatreynoldo's picture

Damn, Idiot, that's hilarious! And it is probably true since the data is not broken down by 4-3 or 3-4, or by ILB or OLB. So, with CM3 (until recently the highest paid LB) and Peppers (8th highest), and Neal, Hawk, Brad Jones all being paid decent bucks, we might well have been.

Dummy me, Mike Reuter wrote an article on this on 6/18/15. GB ranked 2nd in LB pay in 2014 and is poised to rank 2nd again in 2015. So our high payments to LBs is indeed being used to suggest that there is a correlation between paying linebackers well and winning. In truth, at OLB, GB is richly paying the starters and Neal, one of the main backups, is getting really good back-up money. Here is the link.

Portland Mark's picture

Good news from Minnesota! Running back spending = .8152 and Vikings just gave Peterson $25 million for this year. Way to go Vikings!

hobbes's picture

From what I understand, the Vikings actually didn't give up too much with this "renegotiation". Peterson's big fuss was that the remaining years of his contract were essentially un-guaranteed, meaning while his actual salary is high, there was no protection from getting cut by the way of prorated bonuses, injury guarantees etc. Of course, from the Vikings perspective, why give Peterson a better deal when they took all the risk at the beginning of the contract plus the fact that Peterson is still the highest paid RB in the game. What I think the Vikings did was guarantee essentially this years salary, which in effect was already guaranteed due to every statement they made about keeping Peterson this year, but they still can cut Peterson the year afterward without any penalty.

While this doesn't make the Vikings any more dumb for signing Peterson to that outrageous contract, it shows that they are still willing to be the same amount of dumb for him.

DrealynWilliams's picture

**Straight face**

Soooo, I was at the barbershop earlier this morning and my barber goes, "Oh yea! Congratulations! I heard y'all picked up Reggie Wayne."

.... I Immediately replied," Eff no! Where did you hear that?! We're too deep."

I never checked CHTV, PackersMix, BleacherReport & NFL so fast!


murphy's picture

I heard that they cut Trish to make room for Wayne. He was a big trouble maker anyway.

Log in to comment, upload your game day photos and more!

Not a member yet? Join free.

If you have already commented on Cheesehead TV in the past, we've created an account for you. Just verify your email, set a password and you're golden.

Or log in with Facebook

Packers Tickets, Ticket King


"A school without football is in danger of deteriorating into a medieval study hall. "
"I firmly believe that any man’s finest hour, the greatest fulfillment of all that he holds dear, is that moment when he has worked his heart out in a good cause and lies exhausted on the field of battle – victorious."
"The Bears still suck!"