DallasInDC
Well-Known Member
- Messages
- 5,136
- Reaction score
- 5,019
I know, I know, the last thing we need in the zone is another QB ranking thread….but bear with me for a moment…the analysis I am going to share will tell us what most of us already know but some of it may (or may not) surprise you. Plus I would like your feedback on how convincing of an argument I am making to a coworker.
Last week at work I was talking to a coworker, who is a Panthers fan, who is always ribbing me about Tony Romo. He regurgitates the same old Media crap we have all grown accustom to hearing, i.e. isn’t a top tier QB, can’t win big games, throws more INTs than other top QBs, yada, yada, yada. In our discussion last week, he tells me that Tony Romo was only good last year because of the running game and without Murray he will go back to being outside the top 10 QB in where he has always been. Now I know statistically Romo has been a top 10 QB since he became a starter and typically is in the top 3-6 in most statistical categories. Since these arguments seem to go know where….I make my point, he makes his point, know one budges….I figured I need to provide some statistical analysis that will demonstrate my point and force him to respond to the data. Know I know there will always be the intangible part of the argument that I will never be able to overcome, but at least he will have to admit the data says what it says. So onto my analysis:
I decided to look at only the QBS that are typically included in any top 10 discussion. They include the following 13 QBs:
A. Rodgers (24) P. Manning (1) D. Brees (32) T. Brady (199)
B. Roethlisberger (11) A. Luck (1) R. Wilson (75) P. Rivers (4)
M. Ryan (3) E. Manning (1) J. Flacco (72) E. Manning (1)
Tony Romo (UDFA)
Note: I included their draft position in parenthesis. Eight of the thirteen QBs are first round picks and six of those were top four picks. Only two were outside of the top 100 picks and there was only one undrafted QB as you all know.
I looked at their body of work over the past 5 years (I chose five years to award consistency of QB play). I used a weighted average giving more weight to the most recent seasons and less weight to older seasons. I did this to give more consideration to current performance and to account for potential declining play over the 5 year period. I also only compared their stats to each other only, i.e. leaving out the other 19+ QBs when doing the rankings. I looked at three categories of stats:
Traditional QB stats that include QB Rating, Comp %, YDS/ATT (TD %), INT/ATT (INT %), TD/ATT, TD/INT, and TD% to INT % differential (I thought this was going to be identical to TD/INT but the results were different enough to leave it). I know that most of these stats are components of the QBR and may seem a little redundant, but thought it was important to see the ranking impacts of each of these. I also looked at my total analysis with only each of the components eluding the QB rating and did not see much difference in results)
Total QBR Rating which provides a different view of QB success that tries to assess value to wins based on each plays contribution to the win/loss.
Regular season win success, which includes each QBs overall win/loss percentage over the 5 year period. I also included a ranking of defensive yards allowed and rushing yards gained with the assumption that wins are predicated not only on QB play but the impact of how much pressure defenses put on (or alleviate from) the offense, as well as the pressure that a good rushing attack takes off (puts on) a QB by forcing defenses to be less (more) predictable, e.g. if a QB’s defense is giving up more yards than the peer set, chances are the other team is scoring more points at most, and are putting the offense in worse field position at best. This would be a variable that the offense (mostly QB) would have to overcome to be successful. The same concept if the rushing yards are lower, then the QB will have to overcome the lower offensive production most likely against a defense that knows they will have to rely on the pass. I did not include post season appearances in my analysis because of the various non QB related factors that influence playoff appearances, e.g. a team with a 8-8 records reaches the post season due to winning the division vs. a team with a 10-6 record not getting in. Also, I did not look at post season success because there are other influencing factors that have a much bigger impact on wins/losses that cannot be accounted for in which each team will have at most 4 games each season). This includes weather, good (or bad) luck, home field advantage, referee decisions, matchups, momentum, etc. I included the regular season because each team (and in most cases QBs) has 16 games and the external influences even themselves out with the larger number of games. I also wanted to account to some degree for the QBs get the credit for wins crowd knowing good well that playoff success and SB wins do not fall squarely on QB play or leadership. Note for QBS who did not play at all in the season, I eliminated the defense and rushing ranking for that year as well, e.g. Luck and Wilson didn’t play in 2010/2011 so I excluded their defense/rush ranking for that year from the averages.
Here are the results:
http://i77.***BLOCKED***/albums/j61/dallasallen2/QB%20Stats%201_zps7piipv5z.jpg
Interesting note, Rodgers and P. Manning are the only QBs who are above the top 10 average which clearly demonstrates their elite status. Romo, Brees, and Brady are above the top 10 average in 6 of the 11 categories. The rest of the QBs, however, typically perform outside of the top 10 average.
I then took the data above and ranked each stat from 1-13 and color coded the results as follows 1-3 green; 4-5 blue (to round out the top 5); 6-10 red (the back end of the top 10); 11-13 grey (outside of top 10 rankings).
http://i77.***BLOCKED***/albums/j61/dallasallen2/QB%20Stats%202_zpsyiqcucbi.jpg
Looking at the color coded rankings, you will start to see a pattern. Stafford, E. Manning, and Flacco should really not be in the top 10 QB discussion (even with consideration of their SB rings). Across the board statistically they are outside of the top 10 or at best the back end of the top 10. A couple of interesting observations, Eli Manning has had the least amount of team support (Def and Rush Impacts) which bares itself out in the Win %.,,,mostly driven by the weighted averages of the past two years. Wilson overall is much better than I expected with a top 5 ranking in 6 categories, however, you can see that he was aided the most by a top defense and running game. What is clear, is Rogers and P. Manning are the elite QBs Brady is in the bottom half of the elite status and directly below them is Romo and Brees.
Finally, I added all of the ranking to come up with a master ranking of the 13 QBs. I ranked them in three categories just to see what the impact was of each category:
While the results are not overly surprising, I expected Romo to be in the 4-5 category and not necessarily # 3 overall. Goes to show you how lucky we are to have an UDFA that outperforms 6 other highly touted first round picks (including 5 top four picks). This fact alone should have made him a media and fan favorite. Even looking at the Win% category alone he finishes tied for 6 which indicates he has done as much to help his team win as the other top tier QBs. I did expect both Brady and Brees to be higher than Romo. I also expected Wilson to be lower than he is, which tells me that Wilson performs much better than I thought, but is definitely aided by having the best defense and rushing attack of the 13 QBs. The data tells me that Luck is overrated in general in QB rankings and is given overall credit for team wins and draft status. Finally, it tells me that Manning, Flacco and Stafford really shouldn’t even be in the conversation and that Roethlisberger and Rivers are overrated as well.
As far as my original premises, based on the data the results are what they are; they are based on metrics applied consistently across the board. You can argue if Brady or Brees should be higher (although it appears that they are starting to decline). But at the end of the day there is no way you can argue that Tony Romo is not a top 5 QB in this league. I am sure there are many different ways to look at this and there is a lot to criticize about this approach so I welcome your feedback before I unleash it on my co-worker.
Last week at work I was talking to a coworker, who is a Panthers fan, who is always ribbing me about Tony Romo. He regurgitates the same old Media crap we have all grown accustom to hearing, i.e. isn’t a top tier QB, can’t win big games, throws more INTs than other top QBs, yada, yada, yada. In our discussion last week, he tells me that Tony Romo was only good last year because of the running game and without Murray he will go back to being outside the top 10 QB in where he has always been. Now I know statistically Romo has been a top 10 QB since he became a starter and typically is in the top 3-6 in most statistical categories. Since these arguments seem to go know where….I make my point, he makes his point, know one budges….I figured I need to provide some statistical analysis that will demonstrate my point and force him to respond to the data. Know I know there will always be the intangible part of the argument that I will never be able to overcome, but at least he will have to admit the data says what it says. So onto my analysis:
I decided to look at only the QBS that are typically included in any top 10 discussion. They include the following 13 QBs:
A. Rodgers (24) P. Manning (1) D. Brees (32) T. Brady (199)
B. Roethlisberger (11) A. Luck (1) R. Wilson (75) P. Rivers (4)
M. Ryan (3) E. Manning (1) J. Flacco (72) E. Manning (1)
Tony Romo (UDFA)
Note: I included their draft position in parenthesis. Eight of the thirteen QBs are first round picks and six of those were top four picks. Only two were outside of the top 100 picks and there was only one undrafted QB as you all know.
I looked at their body of work over the past 5 years (I chose five years to award consistency of QB play). I used a weighted average giving more weight to the most recent seasons and less weight to older seasons. I did this to give more consideration to current performance and to account for potential declining play over the 5 year period. I also only compared their stats to each other only, i.e. leaving out the other 19+ QBs when doing the rankings. I looked at three categories of stats:
Traditional QB stats that include QB Rating, Comp %, YDS/ATT (TD %), INT/ATT (INT %), TD/ATT, TD/INT, and TD% to INT % differential (I thought this was going to be identical to TD/INT but the results were different enough to leave it). I know that most of these stats are components of the QBR and may seem a little redundant, but thought it was important to see the ranking impacts of each of these. I also looked at my total analysis with only each of the components eluding the QB rating and did not see much difference in results)
Total QBR Rating which provides a different view of QB success that tries to assess value to wins based on each plays contribution to the win/loss.
Regular season win success, which includes each QBs overall win/loss percentage over the 5 year period. I also included a ranking of defensive yards allowed and rushing yards gained with the assumption that wins are predicated not only on QB play but the impact of how much pressure defenses put on (or alleviate from) the offense, as well as the pressure that a good rushing attack takes off (puts on) a QB by forcing defenses to be less (more) predictable, e.g. if a QB’s defense is giving up more yards than the peer set, chances are the other team is scoring more points at most, and are putting the offense in worse field position at best. This would be a variable that the offense (mostly QB) would have to overcome to be successful. The same concept if the rushing yards are lower, then the QB will have to overcome the lower offensive production most likely against a defense that knows they will have to rely on the pass. I did not include post season appearances in my analysis because of the various non QB related factors that influence playoff appearances, e.g. a team with a 8-8 records reaches the post season due to winning the division vs. a team with a 10-6 record not getting in. Also, I did not look at post season success because there are other influencing factors that have a much bigger impact on wins/losses that cannot be accounted for in which each team will have at most 4 games each season). This includes weather, good (or bad) luck, home field advantage, referee decisions, matchups, momentum, etc. I included the regular season because each team (and in most cases QBs) has 16 games and the external influences even themselves out with the larger number of games. I also wanted to account to some degree for the QBs get the credit for wins crowd knowing good well that playoff success and SB wins do not fall squarely on QB play or leadership. Note for QBS who did not play at all in the season, I eliminated the defense and rushing ranking for that year as well, e.g. Luck and Wilson didn’t play in 2010/2011 so I excluded their defense/rush ranking for that year from the averages.
Here are the results:
http://i77.***BLOCKED***/albums/j61/dallasallen2/QB%20Stats%201_zps7piipv5z.jpg
Interesting note, Rodgers and P. Manning are the only QBs who are above the top 10 average which clearly demonstrates their elite status. Romo, Brees, and Brady are above the top 10 average in 6 of the 11 categories. The rest of the QBs, however, typically perform outside of the top 10 average.
I then took the data above and ranked each stat from 1-13 and color coded the results as follows 1-3 green; 4-5 blue (to round out the top 5); 6-10 red (the back end of the top 10); 11-13 grey (outside of top 10 rankings).
http://i77.***BLOCKED***/albums/j61/dallasallen2/QB%20Stats%202_zpsyiqcucbi.jpg
Looking at the color coded rankings, you will start to see a pattern. Stafford, E. Manning, and Flacco should really not be in the top 10 QB discussion (even with consideration of their SB rings). Across the board statistically they are outside of the top 10 or at best the back end of the top 10. A couple of interesting observations, Eli Manning has had the least amount of team support (Def and Rush Impacts) which bares itself out in the Win %.,,,mostly driven by the weighted averages of the past two years. Wilson overall is much better than I expected with a top 5 ranking in 6 categories, however, you can see that he was aided the most by a top defense and running game. What is clear, is Rogers and P. Manning are the elite QBs Brady is in the bottom half of the elite status and directly below them is Romo and Brees.
Finally, I added all of the ranking to come up with a master ranking of the 13 QBs. I ranked them in three categories just to see what the impact was of each category:
- Ranked only the traditional QB stats excluding QBR and Win % stats (including def and rush stats)
- Ranked the traditional QB stats and the QBR stats and excluded Win % stats (including def and rush stats)
- Win % category stats only (includes total win%, defensive ranking – higher yards mean higher ranking, Rushing ranking – lower yards means higher ranking)
- Ranked all of the stats including the Win % stats
While the results are not overly surprising, I expected Romo to be in the 4-5 category and not necessarily # 3 overall. Goes to show you how lucky we are to have an UDFA that outperforms 6 other highly touted first round picks (including 5 top four picks). This fact alone should have made him a media and fan favorite. Even looking at the Win% category alone he finishes tied for 6 which indicates he has done as much to help his team win as the other top tier QBs. I did expect both Brady and Brees to be higher than Romo. I also expected Wilson to be lower than he is, which tells me that Wilson performs much better than I thought, but is definitely aided by having the best defense and rushing attack of the 13 QBs. The data tells me that Luck is overrated in general in QB rankings and is given overall credit for team wins and draft status. Finally, it tells me that Manning, Flacco and Stafford really shouldn’t even be in the conversation and that Roethlisberger and Rivers are overrated as well.
As far as my original premises, based on the data the results are what they are; they are based on metrics applied consistently across the board. You can argue if Brady or Brees should be higher (although it appears that they are starting to decline). But at the end of the day there is no way you can argue that Tony Romo is not a top 5 QB in this league. I am sure there are many different ways to look at this and there is a lot to criticize about this approach so I welcome your feedback before I unleash it on my co-worker.
