Friday, August 6, 2010

QB Value: Kyle Orton vs. Peyton Manning

When trying to asses the value of any particular player in a salary cap world, it is vital to compare their on field performance to the resources that they use up (Cap Space). A great player can only demonstrate their greatness if a team is able to provide other valuable players around them, so any player who does not perform up to their salary (or more specifically their cap number), is hurting their team. For example, if a QB has a cap number of $8 million, but their performance is equivalent to that of a $4 million QB, that QB has wasted $4 million of their teams money that could have been spent on other players and positions to bolster the entire team.

The problem of course is evaluating the performance of one player in a game that has 22 players on the field and data collected for only 3 or 4, then translating that on field performance into a relevant salary. For starters, lets look at QBs. Of course, due to data limitations, this is truly an evaluation of the passing game, not just the QB, but again, this is a start. For all players, we need to measure both the quality and the quantity of their performance. We need to differentiate between the QB that has thrown for 4000 yards on 500 attempts and the one that had thrown for 4000 yards on nearly 600 attempts. Clearly the more efficient QB is more valuable, but we additionally need to recognize that a QB who starts 16 games has contributed more than a QB who has started 2 games, regardless of how well the 2 game started played.

Combining efficiency (Yds per attempt), accuracy (completion percentage) and total yards, produces a familiar ranking of the top QBs of 2009: Brees, Favre, Schaub, Manning and Roethlisberger. Rivers and Rodgers are also high up on the list. The data suggests that, when compared to their peers, these guys all played well enough to "earn" over $11 million. Problem is, many of these guys made well over $14 or $15 million.

Dividing a player's earned salary by their actual salary (or cap value), we get the percentage of the salary that they actually earned. Players who have earned more than 100% of their salaries are adding more value to their team than they are costing the team. Sorting the top 20 QBs by percentage of earned salary provides a slightly different ranking of QBs. Top of the list is now Kyle Orton whose 62% completion percentage, 3800+ yards and 6.8 yards per attempt suggest that Orton provided the Broncos with over $8 million in value, while costing them only $1.1 million which means that Orton earned his salary 7X over.

2009 Regular Season
Rank Player   Cmp%   Yds   AY/A   Earned Salary 2009 Cap Charge % Salary Earned
1 Kyle Orton  62% 3802 6.81  $8,038,507  $1,095,000 734%
2 Tony Romo  63% 4483 8.36  $10,340,410  $4,588,479 225%
3 Jason Campbell  65% 3618 6.59  $8,417,210  $3,895,280 216%
4 Joe Flacco  63% 3613 7  $8,442,026  $4,881,760 173%
5 Aaron Rodgers  65% 4434 8.72  $11,013,690  $9,652,500 114%
6 Drew Brees  71% 4388 8.9  $12,090,021  $10,660,400 113%
7 Matt Schaub  68% 4770 8.02  $11,541,761  $10,250,000 113%
8 Philip Rivers  65% 4254 9.07  $11,158,450  $11,541,630 97%
9 Brett Favre  68% 4202 8.56  $11,597,205  $12,000,000 97%
10 Ben Roethlisberger  67% 4328 8.51  $11,344,517  $13,221,560 86%
11 Kurt Warner  66% 3753 7.1  $9,620,553  $11,504,680 84%
12 David Garrard  61% 3597 6.68  $7,162,938  $9,000,000 80%
13 Tom Brady  66% 4398 7.74  $10,681,780  $14,627,280 73%
14 Eli Manning  62% 4021 7.72  $9,269,839  $13,066,668 71%
15 Peyton Manning  69% 4500 7.78  $11,358,447  $21,205,718 54%
16 Donovan McNabb  60% 3553 8  $8,181,514  $16,773,950 49%
17 Jay Cutler  61% 3666 5.47  $5,442,645  $11,534,999 47%
18 Matt Hasselbeck  60% 3029 5.34  $4,225,949  $9,456,240 45%
19 Matt Ryan  58% 2916 6.04  $4,289,558  $9,907,280 43%
20 Carson Palmer  61% 3094 6.29  $5,820,264  $14,300,000 41%

Compare Orton's value to that of Peyton Manning. Manning put up great numbers as usual with 69% completion rate, 4500 yards and 7.78 yards per attempt for an earned salary of  $11.4 million: 42% higher than Orton's earned salary. But Peyton had a cap cost last year of $21.2 million, meaning that he cost the Colts about $10 million more than he gave them (Percentage of Salary Earned was 54%).

For the team, that means that, if they had paid Peyton his earned salary, they would have had an additional $10 mil to sign players. Imagine adding Ray Lewis (cap charge of $5mil for the Ravens last season) to the team that lost the Super Bowl and tell me that having him and another $5 mil player would not have made a difference.

Of course Manning signed a long term contract and I am not suggesting that the Colts made a mistake in locking him up as they did. Manning is one of the greats of all time and is playing at a very high level. It is also likely that everything that he brings to the team is not captured in this quickie statistic of earned salary. What I am trying to demonstrate however, is the power and importance of not overpaying players in a salary cap world. Every dollar spent on one player is a dollar not spent on another player, so that signing a Sam Bradford to an enormous contract, can kill a team if he does not live up to it while signing a guy like Kyle Orton can provide huge flexibility to the organization, even if he does not provide the top level of QB play.

Note that of our top five QB seasons from last year on a pure performance basis, none of them make the top 5 of Percentage of Salary Earned with Brees being the best coming in at 6th with an earned salary $1.4 mil above his cap charge, followed immediately by Schaub with an earned salary $1.3mil above his cap charge. This demonstrates the danger of large long term contracts for NFL teams, as well as the ease with which teams can overpay at the QB position.

Performance data is provided by Pro-Football-Reference and salary data is from USA Today.


  1. Counterpoints:
    Franchise QBs bring in huge amounts of revenue.

    The Annual Harris poll of NFL Team popularity loves pro-bowl/celebrity QBs. Favre leaving the Jets dropped them from 9th to 20th!

    Also (history changing defenses notwithstanding), the consensus is that you need a franchise QB to win a superbowl.

  2. Mark Rypien
    Trent Dilfer
    Brad Johnson
    Tom Brady (the first time at least)
    Jeff Hostetler
    Doug Williams (debatable)

    These were not franchise quarterbacks, so I'm not sure about that consensus.

  3. I still think you need to add or quantify post season appearances and wins and as well as a huge bonus for SB appearance and or win!!!

  4. This kind of analysis would mesh nicely with PFR's Approximate Value system (or something like it), which gives value to every player on a team on a seasonal level.

  5. Nice work but one problem I see is your formula makes every a/ypa equal when it seems most likely that each additional yard of a/ypa means much more. I am not sure if I am explaining it clearing so I will try it this way a QB that has a a/ypa of 6.5 and makes league average while the next guy has a a/ypa of 7.5 but makes 10 million a season you system will reward the first guy and hit the second guy when the differences in expected wins might be a few games.

  6. This is an interesting way to consider value. Thanks for compiling the stats.

    I had the same feeling as Jim above. I would phrase his argument in a slightly different way, the mapping of these average stats and value is probably not as "linear" as we are making it out to be. Marginal increases in compl % and ypa above a threshold probably have a higher correlation with additional offensive production and wins. And that is probably where the true value of a Manning or a Brady is realized.