I always like UNADJUSTED yards per pass attempt. To me its the best stat. 7.6 or better is good. 8 or higher is great and Ryan Tannehill 2019 was GREAT leading league with a 9.1 or so. YAC is included as it means QB is hitting targets going running speed, he's executing most plays exactly as the route combinations were designed.
It punishes checkdown dink and dunkers who dont move chains. Backup QBs stay in the league by having high completion rates, awful YPA in the 6's. Also guys who sling deep incompletions get behind the chains and fail to keep the ball.
In general though it shows an exciting offense in one game. Two guys with 8 YPA and a close finish is a fun game to watch/rewatch.
When I go back to the old days, I often like to use NY/A myself, if we're trying to measure how good a QB actually was, independent of any weight at all on turning the ball over. It's basically what Y/A is, but with a punishment for sacks taken. 2019 Ryan Tannehill is a good example in my opinion of what NY/A can do.
Ryan was by far the best QB in the NFL in terms of simple Y/A in 2019. Fair enough. He's actually the best in terms of NY/A too, but with his sk%+ of 75 that year, he falls basically all the way down to Patrick Mahomes in a razor tight race for the top spot. This is why he's actually worse than Patrick from an EPA/Play perspective that year, in addition to Patrick's electric receiver group.
My only opinion on stats scaled in yards is that the punishment for losing yards is not big enough. This is not a Y/A criticism or a NY/A criticism or any specific stat criticism, it's that scaling in terms of yards understates just how awful losing yards is in the pursuit of scoring points. There's nothing worse than a negative yardage play. In something like NY/A, you can make up for a six yard sack with a six yard pass, but in actual football, it takes a lot more than a six yard pass to make up for a six yard sack.
This is why basic rushing yards for RBs is such a bad stat as well. Negative yardage runs are not punished enough. The very most important thing in football are your three opportunities to get to the next checkpoint. If you waste one, you are significantly hindering your team's chances of scoring, and stats scaled in yards have a really hard time evaluating the meaning of this wasted opportunity. They basically can't do it.
So while we can argue about Y/A and NY/A and all this, the lead we're burying is that it's not necessarily the yards that are important, it's series conversion that's important. Always getting to the next checkpoint. Yards have a high correlation with this, but not a perfect one. If we stay in the 2019 season again, Lamar Jackson is tenth that year in NY/A. 12th in basic Y/A, but he was by far the best QB around, and everybody knew it.
A lot of it is to do with rushing value, but more of it has to do with the fact that his Ravens team was never wasting chances. Never going backwards. Y/A and NY/A and any other stat scaled in yards did not reward this enough, but the ones that could understand the value of a down wasted (mainly EPA/Play and the eye test) valued it immensely.
In short, I'm not the biggest ANY/A guy either. I simply used it here because it's a stat a lot of people know and like. I would prefer simple Y/A or NY/A over it, but this whole conversation buries the lead that stats that are scaled in yards have the fundamental flaw that they believe going six yards forward can compensate for going six yards backwards, when it can't.
I think getting those 0 yards on INC punishes YPA. You keep wasting downs. The problem isn't yards its INTs and strip sacks being filtered out. Which, are certainly huge but only Jameis gives you a high YPA and loses games with his INTs. I haven't seen any mediocre QB get high YPA. Ive seen a lot of Brady getting low YPA especially early career. Flacco with a cannon but too many INC. Eli Manning being below avg most of his career. That Guy threw INTs on screen passes. He would just lapse into scattershot and his YPA told the story of a guy who missed way too often.
It's true. You're correct about this Y/C stuff. It's a difference between a statistic explaining 60% of how good a player is and 70%. The new fangled fancy stuff does not buy you as much as many people think it does. I use it because they do still buy you something, but it's not a ton.
Nothing too in depth, but, in the spirit of this publication, went through looking at some of the best ANY/A results seasons from some of the worst career xANY/A quarterbacks. Biggest differential I saw poking around, I was pleasantly surprised to see, already has a His Year article: 2007 Derek Anderson.
There are some other seasons with larger absolute differences to their career xANY/A, but most involve bad quarterbacks having around league average seasons, results-wise. The next best I found was 1987 Marc Wilson, who actually managed a top-ten ANY/A season. Even in that season, the skills model doesn't love him, but he got some good results in losing efforts anyway.
In terms of the reverse (worst ANY/A season from some of the best career xANY/A quarterbacks), there's the obvious "young QB struggles" (TB Steve Young) and "aging veterans that probably should have retired" (2015 Peyton Manning). But sandwiched right in the middle of Fran Tarkenton's career is the 1971 season that saw him manage just 3.90 ANY/A. Even accounting for era, this is a bad result, despite fairly good skills model numbers (4th in the league that season). His Int% is a bit higher than his career norms, but still below league average on a rate basis, and he only goes from 16th to 14th when you change to NY/A (taking out the effect of interceptions).
When I did my one-off QBs article, I did have one of the 'really bad QBs having a league average season' archetypes in third place with 1986 Randy Wright, but 2007 Derek Anderson is absolutely ridiculous. It's ridiculous in that it's not ridiculous, which is out of place in such a ridiculous career.
What I said in my one-offs article I meant. This is the worst QB in NFL history, who also happened to be about the 12th best QB in the league (in both results and skills) one time. Derek Anderson's 2007 is a 0.086 xEPA/Play, 6.19 xANY/A season that results in 0.085 EPA/Play, and 6.29 ANY/A. It's one of the most normal seasons on the record, which would've never let me or anybody predict that from the end of 2007 and onwards, this man was going to have a career xEPA/Play of -0.206, and xANY/A of 3.65. Both figures are worse than JaMurcus Russell's career skill metrics. They're worse than any QB with multiple qualified seasons.
I do not know what happened to Derek Anderson, but I'm perfectly okay with speaking in extremes about it. This is the worst QB ever to be given the privilege of qualifying for the leaderboards multiple times, and I wholeheartedly mean that. Akili Smith is better than this. Jack Concannon is better, but somehow, some way, this guy managed to be pretty good for a year there. It's one of the most inexplicable things ever, and brings his career stats from worse than JaMarcus Russell into the range of extremely bad.
You've actually stolen my idea David. I also planned to look into really bad players having really good seasons. I still plan to do that, but you've pre-empted me a little bit LOL.
1971 in specific for Fran is a really interesting time for his results to drop off despite the skill not dropping off. He was embroiled in a fairly nasty contract dispute when he asked for a $300000 loan on top of his salary to fund some business interests of his. The Giants said no, a weirdly poor results season ensues, and Fran is then traded to Minnesota, and never plays for the Giants again. I'm not saying anybody played badly on purpose or anything, but that is very interesting timing if nothing else.
I'm sure there's more to find in there! I just manually went through the worst 20 or so in the career xANY/A table (under the thinking that the absolute differences will be larger), but there's a lot more names on the list to cover. You probably already have the data to be able to do it more systematically.
I did come away feeling for you as a suffering Jags fan. I knew Gabbert and Bortles back-to-back was not a pretty combination, but I hadn't realized the depths of how historically bad they were. Gabbert, for instance had the second largest absolute difference between career xANY/A and individual season ANY/A when he produced 5.71 ANY/A in San Francisco in 2015--but his xANY/A is so bad that it was still below league average for the season (using whatever pro-football-reference uses as the qualifying cutoff in the ANY/A leaderboard).
The QB problem in Jacksonville is beyond words, but we brought it on ourselves. The Jags had never had a negative xEPA/Play QB ever, up until the point when Shahid Khan bought the team, and the incumbent regime decided to try to fix a David Garrard problem that really really was not broken. David had a 4.5 CPOE and 0.127 xEPA/Play in 2010. QB was not an issue, but the geniuses that this team has become under the Khan ownership decided to draft Blaine Gabbert anyways, and the fallout of that decision has been nothing short of disastrous.
After 16 seasons of never having a single negative xEPA/Play season from our QBs, we booted a good guy out to the curb, and since then it's been like the curse of David Garrard. Over the fifteen years since that horrible decision, we have had a positive xEPA/Play QB just five times. One of those times was Gardner Minshew in 2020, after which we made the Garrard mistake a second time, kicking him to the curb in order to replace him with Trevor Lawrence.
At least Trevor is mostly in the positives, but even in his best season, it's only 0.147 xEPA/Play. That's not better than David Garrard could do, meaning we went through this entire 15 year QB odyssey basically to get back to the point where we could have David Garrard playing for us again. I love David Garrard. He was a good player, but not the kind of guy you wait 15 years for.
This is organisational incompetence of the highest order, and it never ceases to amaze that you can see this abrupt change right at the point where the franchise is sold. From 1995-2010, the Jaguars were not always at the top of the league, but were a team to take seriously, and occasionally did get to the top of the league. From 2011-present, we have been a doormat. We've been a team you have to at least take seriously three times (2017, 2022, 2023), but we've only been a team to truly fear once (2017).
There are some teams even more incompetent than this, but it's a bit heartbreaking to see what this once great team has reduced itself to over the last 15 years. This is true for the QB position, and most other positions too.
If we take being underrated by 0.7 as the norm for the 1978-2003 era, both of them basically exactly met their expected results.
This does come with the caveat there's an implicit assumption that elite QBs have elite receivers, because elite QBs tend to create elite receivers. It's not explicitly codified in the model anywhere, but implicitly it happens in the calibration process. Guys with good individual stats are expected to have good receivers, because those good individual QB stats go a lot of the way to creating those good receivers, which pulls their expected results up just a little bit. Take a look at guys like Drew Brees or Peyton Manning for example. Each of these guys basically never went without at least one elite receiver, and both actually end up on the underrated side. Not significantly underrated, but their elite receivers gave no boost either, because it's simply expected for an elite QB to have elite receivers.
Boosts more often go the other way, when there is no reason for a QB with such a low level of skill to have such a great offensive environment around him, but for some reason he does anyway, like CJ Stroud the last couple years. Or sometimes, in the case of the Shanahan/McVay guys, that there's no level of individual QB skill that can create an offensive environment THAT good. Underrating often happens due to the exact opposite story, where a QB should have been good enough to create elite receivers for himself, but for some reason never could, like Brian Sipe, Chad Pennington, Bobby Hebert, etc..
I don't even think the offensive environments that Joe and Steve played in were all that abnormal. Joe had the number one and two WRs in the NFL in 1989 with Jerry Rice and Jim Taylor. Fair, but Peyton Manning had both the best and second best WR in the NFL on multiple different occasions according to my ranking system. Does this mean these guys were lucky? A little bit, but in larger part, it means that they were so elite that they created this ridiculous WR room for themselves.
The same goes for all the top xEPA/Play guys. It's not abnormal for any of them to have elite WR rooms, so should we really perceive any of them as having a better WR room than any of the others? This is also why guys like Aaron Rodgers and Tom Brady end up on the overrated side. My model doesn't see either of them as quite good enough to create the 'elite WRs for an entire career' treatment for themselves, but both of them pretty much got it anyway.
It all comes down to whether a QB is merely in his offensive environment, or whether he plays a part in creating it, and I think it's the latter in most of these cases.
I always like UNADJUSTED yards per pass attempt. To me its the best stat. 7.6 or better is good. 8 or higher is great and Ryan Tannehill 2019 was GREAT leading league with a 9.1 or so. YAC is included as it means QB is hitting targets going running speed, he's executing most plays exactly as the route combinations were designed.
It punishes checkdown dink and dunkers who dont move chains. Backup QBs stay in the league by having high completion rates, awful YPA in the 6's. Also guys who sling deep incompletions get behind the chains and fail to keep the ball.
In general though it shows an exciting offense in one game. Two guys with 8 YPA and a close finish is a fun game to watch/rewatch.
When I go back to the old days, I often like to use NY/A myself, if we're trying to measure how good a QB actually was, independent of any weight at all on turning the ball over. It's basically what Y/A is, but with a punishment for sacks taken. 2019 Ryan Tannehill is a good example in my opinion of what NY/A can do.
Ryan was by far the best QB in the NFL in terms of simple Y/A in 2019. Fair enough. He's actually the best in terms of NY/A too, but with his sk%+ of 75 that year, he falls basically all the way down to Patrick Mahomes in a razor tight race for the top spot. This is why he's actually worse than Patrick from an EPA/Play perspective that year, in addition to Patrick's electric receiver group.
My only opinion on stats scaled in yards is that the punishment for losing yards is not big enough. This is not a Y/A criticism or a NY/A criticism or any specific stat criticism, it's that scaling in terms of yards understates just how awful losing yards is in the pursuit of scoring points. There's nothing worse than a negative yardage play. In something like NY/A, you can make up for a six yard sack with a six yard pass, but in actual football, it takes a lot more than a six yard pass to make up for a six yard sack.
This is why basic rushing yards for RBs is such a bad stat as well. Negative yardage runs are not punished enough. The very most important thing in football are your three opportunities to get to the next checkpoint. If you waste one, you are significantly hindering your team's chances of scoring, and stats scaled in yards have a really hard time evaluating the meaning of this wasted opportunity. They basically can't do it.
So while we can argue about Y/A and NY/A and all this, the lead we're burying is that it's not necessarily the yards that are important, it's series conversion that's important. Always getting to the next checkpoint. Yards have a high correlation with this, but not a perfect one. If we stay in the 2019 season again, Lamar Jackson is tenth that year in NY/A. 12th in basic Y/A, but he was by far the best QB around, and everybody knew it.
A lot of it is to do with rushing value, but more of it has to do with the fact that his Ravens team was never wasting chances. Never going backwards. Y/A and NY/A and any other stat scaled in yards did not reward this enough, but the ones that could understand the value of a down wasted (mainly EPA/Play and the eye test) valued it immensely.
In short, I'm not the biggest ANY/A guy either. I simply used it here because it's a stat a lot of people know and like. I would prefer simple Y/A or NY/A over it, but this whole conversation buries the lead that stats that are scaled in yards have the fundamental flaw that they believe going six yards forward can compensate for going six yards backwards, when it can't.
I think getting those 0 yards on INC punishes YPA. You keep wasting downs. The problem isn't yards its INTs and strip sacks being filtered out. Which, are certainly huge but only Jameis gives you a high YPA and loses games with his INTs. I haven't seen any mediocre QB get high YPA. Ive seen a lot of Brady getting low YPA especially early career. Flacco with a cannon but too many INC. Eli Manning being below avg most of his career. That Guy threw INTs on screen passes. He would just lapse into scattershot and his YPA told the story of a guy who missed way too often.
It's true. You're correct about this Y/C stuff. It's a difference between a statistic explaining 60% of how good a player is and 70%. The new fangled fancy stuff does not buy you as much as many people think it does. I use it because they do still buy you something, but it's not a ton.
Nothing too in depth, but, in the spirit of this publication, went through looking at some of the best ANY/A results seasons from some of the worst career xANY/A quarterbacks. Biggest differential I saw poking around, I was pleasantly surprised to see, already has a His Year article: 2007 Derek Anderson.
There are some other seasons with larger absolute differences to their career xANY/A, but most involve bad quarterbacks having around league average seasons, results-wise. The next best I found was 1987 Marc Wilson, who actually managed a top-ten ANY/A season. Even in that season, the skills model doesn't love him, but he got some good results in losing efforts anyway.
In terms of the reverse (worst ANY/A season from some of the best career xANY/A quarterbacks), there's the obvious "young QB struggles" (TB Steve Young) and "aging veterans that probably should have retired" (2015 Peyton Manning). But sandwiched right in the middle of Fran Tarkenton's career is the 1971 season that saw him manage just 3.90 ANY/A. Even accounting for era, this is a bad result, despite fairly good skills model numbers (4th in the league that season). His Int% is a bit higher than his career norms, but still below league average on a rate basis, and he only goes from 16th to 14th when you change to NY/A (taking out the effect of interceptions).
When I did my one-off QBs article, I did have one of the 'really bad QBs having a league average season' archetypes in third place with 1986 Randy Wright, but 2007 Derek Anderson is absolutely ridiculous. It's ridiculous in that it's not ridiculous, which is out of place in such a ridiculous career.
What I said in my one-offs article I meant. This is the worst QB in NFL history, who also happened to be about the 12th best QB in the league (in both results and skills) one time. Derek Anderson's 2007 is a 0.086 xEPA/Play, 6.19 xANY/A season that results in 0.085 EPA/Play, and 6.29 ANY/A. It's one of the most normal seasons on the record, which would've never let me or anybody predict that from the end of 2007 and onwards, this man was going to have a career xEPA/Play of -0.206, and xANY/A of 3.65. Both figures are worse than JaMurcus Russell's career skill metrics. They're worse than any QB with multiple qualified seasons.
I do not know what happened to Derek Anderson, but I'm perfectly okay with speaking in extremes about it. This is the worst QB ever to be given the privilege of qualifying for the leaderboards multiple times, and I wholeheartedly mean that. Akili Smith is better than this. Jack Concannon is better, but somehow, some way, this guy managed to be pretty good for a year there. It's one of the most inexplicable things ever, and brings his career stats from worse than JaMarcus Russell into the range of extremely bad.
You've actually stolen my idea David. I also planned to look into really bad players having really good seasons. I still plan to do that, but you've pre-empted me a little bit LOL.
1971 in specific for Fran is a really interesting time for his results to drop off despite the skill not dropping off. He was embroiled in a fairly nasty contract dispute when he asked for a $300000 loan on top of his salary to fund some business interests of his. The Giants said no, a weirdly poor results season ensues, and Fran is then traded to Minnesota, and never plays for the Giants again. I'm not saying anybody played badly on purpose or anything, but that is very interesting timing if nothing else.
I'm sure there's more to find in there! I just manually went through the worst 20 or so in the career xANY/A table (under the thinking that the absolute differences will be larger), but there's a lot more names on the list to cover. You probably already have the data to be able to do it more systematically.
I did come away feeling for you as a suffering Jags fan. I knew Gabbert and Bortles back-to-back was not a pretty combination, but I hadn't realized the depths of how historically bad they were. Gabbert, for instance had the second largest absolute difference between career xANY/A and individual season ANY/A when he produced 5.71 ANY/A in San Francisco in 2015--but his xANY/A is so bad that it was still below league average for the season (using whatever pro-football-reference uses as the qualifying cutoff in the ANY/A leaderboard).
The QB problem in Jacksonville is beyond words, but we brought it on ourselves. The Jags had never had a negative xEPA/Play QB ever, up until the point when Shahid Khan bought the team, and the incumbent regime decided to try to fix a David Garrard problem that really really was not broken. David had a 4.5 CPOE and 0.127 xEPA/Play in 2010. QB was not an issue, but the geniuses that this team has become under the Khan ownership decided to draft Blaine Gabbert anyways, and the fallout of that decision has been nothing short of disastrous.
After 16 seasons of never having a single negative xEPA/Play season from our QBs, we booted a good guy out to the curb, and since then it's been like the curse of David Garrard. Over the fifteen years since that horrible decision, we have had a positive xEPA/Play QB just five times. One of those times was Gardner Minshew in 2020, after which we made the Garrard mistake a second time, kicking him to the curb in order to replace him with Trevor Lawrence.
At least Trevor is mostly in the positives, but even in his best season, it's only 0.147 xEPA/Play. That's not better than David Garrard could do, meaning we went through this entire 15 year QB odyssey basically to get back to the point where we could have David Garrard playing for us again. I love David Garrard. He was a good player, but not the kind of guy you wait 15 years for.
This is organisational incompetence of the highest order, and it never ceases to amaze that you can see this abrupt change right at the point where the franchise is sold. From 1995-2010, the Jaguars were not always at the top of the league, but were a team to take seriously, and occasionally did get to the top of the league. From 2011-present, we have been a doormat. We've been a team you have to at least take seriously three times (2017, 2022, 2023), but we've only been a team to truly fear once (2017).
There are some teams even more incompetent than this, but it's a bit heartbreaking to see what this once great team has reduced itself to over the last 15 years. This is true for the QB position, and most other positions too.
Interesting how dispute the strong reputation of their offensive environments, Montana and Young get fairly inconsequential boost.
If we take being underrated by 0.7 as the norm for the 1978-2003 era, both of them basically exactly met their expected results.
This does come with the caveat there's an implicit assumption that elite QBs have elite receivers, because elite QBs tend to create elite receivers. It's not explicitly codified in the model anywhere, but implicitly it happens in the calibration process. Guys with good individual stats are expected to have good receivers, because those good individual QB stats go a lot of the way to creating those good receivers, which pulls their expected results up just a little bit. Take a look at guys like Drew Brees or Peyton Manning for example. Each of these guys basically never went without at least one elite receiver, and both actually end up on the underrated side. Not significantly underrated, but their elite receivers gave no boost either, because it's simply expected for an elite QB to have elite receivers.
Boosts more often go the other way, when there is no reason for a QB with such a low level of skill to have such a great offensive environment around him, but for some reason he does anyway, like CJ Stroud the last couple years. Or sometimes, in the case of the Shanahan/McVay guys, that there's no level of individual QB skill that can create an offensive environment THAT good. Underrating often happens due to the exact opposite story, where a QB should have been good enough to create elite receivers for himself, but for some reason never could, like Brian Sipe, Chad Pennington, Bobby Hebert, etc..
I don't even think the offensive environments that Joe and Steve played in were all that abnormal. Joe had the number one and two WRs in the NFL in 1989 with Jerry Rice and Jim Taylor. Fair, but Peyton Manning had both the best and second best WR in the NFL on multiple different occasions according to my ranking system. Does this mean these guys were lucky? A little bit, but in larger part, it means that they were so elite that they created this ridiculous WR room for themselves.
The same goes for all the top xEPA/Play guys. It's not abnormal for any of them to have elite WR rooms, so should we really perceive any of them as having a better WR room than any of the others? This is also why guys like Aaron Rodgers and Tom Brady end up on the overrated side. My model doesn't see either of them as quite good enough to create the 'elite WRs for an entire career' treatment for themselves, but both of them pretty much got it anyway.
It all comes down to whether a QB is merely in his offensive environment, or whether he plays a part in creating it, and I think it's the latter in most of these cases.