I love that you used the term love/hate, because even as I was writing the list, I could not decide whether being a team with lots of one-off QBs ought to be construed as a compliment or an insult. Of course, given the choice, the Bucs would've just drafted Brett Favre or Steve Young and spent the whole 1990s not having to worry about it, but having a couple one-off years also means their QB room was better than some teams in the 90s, who just never had a good QB ever.
The one-offs tended to come from those teams with subpar histories at the QB position, but there are also teams that tend to be really poor (like New Orleans) who do not show up at all on this list. I'm not sure what any of this means, but for some reason the Buccaneers seem to REALLY attract this type of player. They're the best in the whole league at finding them. That much is for sure. I'm just not sure whether that's a compliment or not.
BTW, there would be no feeling better than being the man who drafted Don Majkowski to a fantasy league in 1989. I understand a similar feeling (being a Jaguar fan) with Mark Brunell in 1996. You get to walk around feeling like the biggest genius on Earth LOL.
I'm not a Washington fan, but I am still annoyed to this day by the way they handled Griffin's knee injury. Of all the names on this list, I think he could have most easily avoided one-off status, with just a little bit of common sense by the organization.
What a fun list. I think it just goes to show that, even with all the preparation and hard work that goes into it, football really is a very unpredictable game with relatively small sample sizes.
What's up buddy? I don't believe we've had the pleasure of speaking on here before. Good to meet you! Thanks for commenting.
No question about it. RG3's injury is typical Washington self-sabotage. The re-aggravation in the playoff game didn't need to happen. I don't know if it was Robert's choice to play and the organisation wasn't smart enough to overrule, or if the team wasn't smart enough to realise pressuring him to play wasn't a good idea. Either chain of actions is not acceptable at the NFL level anyway, so it doesn't matter which is the truth. Washington botched this one big time.
However, there is some personal responsibility on Robert too. People tend to forget that the Chargers did almost exactly the same thing to Philip Rivers that the Skins did to RG3, because Rivers is going to the Hall of Fame. What happened to Robert is that he went down with a horrible leg injury, and came back with the worst pocket presence in the NFL. Other QBs that went down with horrible leg injuries (Trent Green, Philip Rivers, Tom Brady, etc.) did not entirely lose their footwork in the pocket. Robert did.
I'm not sure of the reason for that, and it doesn't change the fact that his injury (and therefore, this entire scenario) was preventable, but I don't think there's a direct line of causality from the injury to the untenable footwork (if there is a line, it's one that the true greats were able to overcome, but Robert couldn't). Something else happened in the middle that made Robert unable to stand in the pocket anymore after 2012. I'm not sure what that something is. It certainly isn't public.
I still think I agree with you that Robert's one-off status (along with Elvis Grbac's, all he had to do was stay in Kansas City) is the most easily preventable on this whole list, but it's not 100% the Skins' fault. Maybe 75 or 80.
As far as football sample size goes, you are correct. Individual seasons are not long enough to figure out whether players (and teams) are good or not. QB sample sizes most often are between 550-750 touches per season, roughly the same amount of plate appearances a batter gets in baseball, and we also see one-off baseball seasons all the time. Using the midpoint, a sample size of 650 is clearly not enough to definitively determine that somebody is good at football (or baseball) or not.
In 2006,Rex Grossman's worst game was probably New Year's Eve, when he completed more passes to Packers (3) than to his own team, the Bears (2). That's been one of my favorite single-game NFL stats for almost 20 years.
It's true! It's not in the consideration for the worst game in history, because the Bears were playing for nothing, and Rex was only going to play a half whether he was great or awful, but if those weren't the circumstances that could've devolved into another of the worst games in the league courtesy of Rex Grossman. It could've been just as bad as Arizona.
I loved this piece! Seeing the Vikings' Daunte Culpepper and Case Keenum (and Gus Frerotte for that matter) makes me wonder if 2024 Sam Darnold will end up with a place on a future edition of this list
I ran the numbers just for fun. If the 2024 season ended right now (with Sam's current numbers) he would go into 10th place, just above Daniel Jones, but his continued success at the NFL level depends on whether he gets another chance at the NFL level.
I don't know how hard the Vikings intend to push JJ McCarthy, but with how well this season is going, I can't see a possibility of the team not finding a way to get Sam extended. That seems like a quick and easy way to lose the locker room, to let the QB go after a season that's gone so well. However, as we see by this list, it would not be the first time a QB is let go and never given another chance, even after being a fringe top ten QB.
That archetype (a QB that's maybe top ten, maybe not) seems to be a thing the NFL has always had a really tough time figuring out how to value throughout its history. I hope that Sam does not fall into that whirlpool, but it's entirely possible that he will.
Really honored by the mention! Love how this was approached - especially fun to see how many of these one-offs clustered around specific teams/eras. Also, as a Packers fan, that Wright->Majkowski->Favre sequence was interesting to read about. Never realized we had such a statistical anomaly of one-off seasons right before the stability of the Favre era began. Great deep dive into these forgotten seasons.
You're welcome buddy! All I did was tell the truth. Your lists about useless statistical anomalies inspired me to create my own, but I had to add my own little bit of storytelling flare to it, telling at least a bit of the story of every season involved here. I thought once I was about halfway done that I should've asked if you wanted to do a feature, maybe write one or two of the entries. It seems close enough to your niche, but oh well. I was too deep in my that point. This won't be the last time I get to ask for your help.
You're right that there did seem to be clusters of one-off players. 2017-2020 all had at least one representative on this list, but only one since. Same with there being one every year from 1993-2001, but then there being only two (both in the 2004 NFC) in the next four years. I must admit the 1990s feature a lot of one-off talent at the QB position, including several that fell just short of the cut-off for inclusion. I'm not exactly sure why that is. Maybe it requires some further study. We'll see.
As far as the Packers, my mind was blown when I saw the results. No way they bounced from one one-off straight into another. They also had Lynn Dickey before that, a man whose 1983 season comes in 36th if I would've made my list just slightly bigger. What was in the water up there in the 1980s? There are teams that are just bad at QB, and keep having to wander into one-off QBs (90s Bucs, Bears in general), but never back to back. I would've been terribly scared once Brett Favre broke out in 1994, waiting for him to get injured or something. Like you said, he definitively broke that curse, but with each new QB that comes in, you never know if it's a one-off or not.
Thank you for the compliments buddy! I'm glad you enjoyed it.
The 1994-95 cluster really grabbed me too - Erickson finding magic with the '94 Bucs, then Kramer and Mitchell both peaking in '95 with the Bears and Lions. What's wild is all three were journeymen who temporarily hit their stride just before, or during, the 1995 expansion changed the landscape. The fact that Mitchell led the Lions to the playoffs while Kramer put up the Bears' best QB season in decades, but Chicago missed the postseason, really shows how competitive that '95 NFC Central was.
For sure would be up for collaborating on something in the future if you've got ideas. These statistical oddities are endlessly fascinating, especially when there's a great story behind the numbers.
I entirely forgot about expansion. You're right to bring that up, but in my opinion, it must have something to do with the invention of free agency. I have no proof for that whatsoever, but free agency gets brought in in 1994 (what do you know?), and all of a sudden one-off QB performances shoot to the moon. I don't know what the two would have to do with each other, or why one-off performances still would've gone through spurts like they have. Free agency has never left., but I feel it must have something to do with it.
BTW, how about the first two free agent QBs in NFL history (Scott Mitchell and Erik Kramer) both being top three QBs in 1995 to get the all-new open QB market off the ground? I can almost smell a conspiracy in there somewhere LOL.
You better believe it buddy, I'm going to be talking about the 1995 NFC Central at some point. The year the three best QBs in the NFL were all in the same division. Two of them are here, the other is league MVP Brett Favre. Never been done before or since. I don't know when I'm going to talk about it, and I don't know how, but it's too good a story to stay buried.
Your point about free agency really got me thinking. Looking at this list, it seems like system/team changes might be a huge factor in these one-off seasons. Kramer and Mitchell lighting it up in '95 as early free agency successes is fascinating, but even before that, you see a pattern of QBs finding temporary magic in new situations - Hostetler with the Raiders in '93, Foles in 2013 when the Eagles brought in Chip Kelly.
The early free agency era feels especially ripe for these performances because teams were still figuring out how to evaluate QB fits. You've got this perfect storm of:
- QBs getting their first real chance to choose systems
- Teams experimenting with different offensive philosophies
- Less analytical evaluation of QB skills/fit
- Coaching staffs perhaps being more willing to adapt to new QBs
What really stands out is how few recent seasons make the list - only 5 (if I counted right) out of 30 from 2017 or later (Jackson, Tannehill, Trubisky, Keenum, Jones). Makes me wonder if modern analytics have made these lightning-in-a-bottle seasons harder to find. Teams now have so much data on QB performance in different schemes, pressure situations, throw types, etc. The kind of statistical anomaly that marked someone like Derek Anderson's 2007 seems less likely when teams can break down every aspect of QB play.
The 1995 NFC Central clustering is even more interesting through this lens - you've got this unique moment where free agency is new, teams are still figuring out how to build around QBs, and suddenly three guys hit career peaks simultaneously. It feels like a time capsule of an era when these breakout seasons were more common, before advanced metrics and sophisticated player evaluation changed how teams develop and assess QB talent.
This all makes me wonder if we'll see fewer one-off seasons going forward as teams get better at evaluating QB fit. And if factors like scheme evolution and coaching changes keep creating opportunities for these anomalies.
The relative scarcity of recent examples suggests the game might be changing in this regard.
I think you're right on both counts. There's a definite spike in the 1990s, and a lull right now. It's entirely possible that the invention of free agency created a market inefficiency, and (for some reason) the QB graveyards in Detroit and Chicago were able to capitalise on it better than anybody else could, using it to steal great seasons where they really weren't deserved, sort of like Moneyball in football.
QB free agency is still good at doing this. The 2023 Bucs stole a good season when they really weren't positioned to have one via the use of QB free agency, as did this year's Vikings.
I think it's also the case that some teams are better at the open QB market than others, as the Bucs keep doing this (Brad Johnson, Jeff Garcia, Tom Brady, Baker Mayfield are all free agent QB acquisitions that have done great things for them, with only the first round draft picks Josh Freeman and Jameis Winston breaking the chain), whereas other teams have not yet signed a QB free agent worthy of note, despite free agency having been around for 30 years now. Some (like Green Bay) because they haven't needed to, and others (like Cleveland or Houston) where it really would've helped, but they keep picking the wrong guys.
I think there are two big reasons for the lack of one-offs these days. The first is what you said. Evaluation is likely better than it used to be. There is a second though, and it's that the league is filled with more bad QBs in starting spots these days. Players aren't given chances like they used to be, unless they are first round picks. For a previous article, I compiled a top of the head list of players who would've never gotten a chance to play in the modern NFL. Here it is:
Warren Moon (the NFL would never take a CFL QB these days), Steve Young after flaming out in TB, Vinny Testaverde after flaming out in TB, Troy Aikman might’ve been run out of Dallas after starting his career with two awful seasons, Jeff Hostetler (backup for that long these days? No way), Mark Brunell, Randall Cunningham (to come back in 1998), Trent Green (same as Hostetler), Rich Gannon, Jeff Garcia, Jon Kitna, Tony Romo.
All of these men except Kitna (sixth in 2003) were top five QBs in the NFL at one point or another. None of them would've done anything if they came into the NFL today, and these are just the great players. Imagine the ones that only have one great season in them that we're discussing here.
I think the culture of tanking that exists in the NFL today prevents a lot of good QB talent from seeing the light of day, because a good QB means winning, which these days is construed as a bad thing (just ask the Giants). This also prevents one-off seasons like this from happening, because teams don't want them. With a middling rookie, you will win fewer games than with a veteran having his best career season. Like we're seeing now with the Vikings, a vet having his best career season will win you games. If you're a team like the Giants or Raiders or Patriots, who are trying to lose, why would you want a good vet on your team? That's contrary to the purpose.
Sam Darnold is a fringe top ten guy right now, and nearly got no chance to play. Imagine all the others that may be just as good, but are not given chances, in favour of unqualified rookies. I think this is also a big reason for the lack of one-offs these days. Bad rookies and second year players are always bad. Nothing one-offlike about that, and that archetype of QB takes up about a third of all starting spots in the league right now.
Fascinating breakdown by decade. That 90s spike really shows how free agency created this window of opportunity for QBs to find their footing.
Your point about modern teams preferring to lose with rookies than win with veterans really hit home. Geno Smith feels like the last of a dying breed - a Rich Gannon type who needed years to develop before finding success. The fact that he's the exception now rather than the rule shows how the tanking mindset has changed QB development.
Looking at those names of past QBs who wouldn't get chances today, it's striking how many backup/fringe starters we have now who show flashes but never get extended looks. Gardner Minshew and Taylor Heinicke put up decent numbers when called upon, but seems like teams would rather roll the dice on high picks. Sam Darnold finding some success feels almost accidental - like teams are so focused on either having a superstar QB or tanking that the middle ground of developing solid veterans has disappeared.
Makes me wonder if we'll ever see another cluster of one-off seasons like the 90s again with the current incentive structure. Even when veteran QBs play well now, it almost feels like teams view it as a problem rather than an opportunity.
As far as the final rankings, no. Because this wasn't really about the final rankings, more about the stories therein, the final rankings were exactly what my formula gave me. I didn't allow personal opinion to creep into that at all.
If I did, Derek Anderson still would've been one. Don Majkowski still would've been two. They are undeniable, the clear GOATs in terms of one-off status, but below two the order would've been slightly different if I were ordering the list personally. I have the same biases as most people. The great seasons (Neil Lomax, Rex Grossman, Scott Mitchell, Erik Kramer, etc.) would've ended up higher, and the usable years out of typically unusable players would've been lower, but then it wouldn't be measuring in terms of one-offness would it?
As far as the formula used, it begins with a rather simple regression framework to assign a score to each QB season. I combed the internet for as many QB ranking lists as I could find, and ended up with a good sample going back to 2008. I used this data to map stats like EPA/Play, CPOE, ANY/A, DVOA, DYAR, etc., plus adjustments for play count (to prevent overrating rate stats), plus adjustments for offensive era (to prevent overrating every QB that played post-2004), to a QB ranking list, which gives every season a score of where they theoretically would belong on a ranking list of QBs in their season, if every season featured exactly the same leaguewide QB quality.
Typically, (obviously) the score lands between 1 and 32. The lower the better. I did not bother to actually bind the score between one and 32, because all that would do is damp the results and add more mathematical computation, without truly changing anything. For instance, my framework believes the best season of all time is Peyton Manning's 2004, with about a -8 on a 1 to 32 scale, indicating that it was well above the calibre of a typical top QB in the league type season.
This gives me a score for every season a QB has ever had since 1980. From here, the formula for their one-off score is to take every QB who started for at least two seasons at the NFL level since 1980, and simply subtract the score of their best season from the score of their second best season. They do not need to be consecutive.
For instance, Lamar Jackson's 2019 scored about a three, indicating that in a normal season, he wouldn't have won MVP, but it is the lowest 2019 score of anybody, so he did deserve it. His second lowest score came in 2024, and was about 8.75. Therefore, his one-off score is 5.75. Same thing for Don Majkowski. His 1989 indicated he would be about 5.78 on a QB ranking list if 1989 was exactly typical. His next best score was 24.53 in 1988, hence his 18.75 one-off score.
The reason my formula could not handle Derek Anderson was that my initial formula involves a log transform in its process, and Derek Anderson outside of 2007 consistently gives negative numbers where they are not allowed, such is the depth of just how poor his play was. I had to finally tweak the formula to actually be able to put him on the scale, which gives a 2007 score of 12.39, and a next best score of 33.54 in 2008. The only person in the history of the NFL to never come within 20 spots of his best ranking on a hypothetical QB ranking list.
Funnily enough, the least one-off QB in NFL history by my formula is Tony Romo, whose 5.31 score in 2007 and 5.42 in 2014 give him a one-off score of 0.11. One tenth of one position on a ranking list. Nothing more to that. I just thought it was funny in how it demonstrates how the formula works.
I hope it makes sense in the way I've explained it!
This makes a ton of sense. Thank you. I wish you had put that information in the article, including the formula, because the stats chosen to highlight the players don’t look all that different in some cases. It also would help to ground how some guys who had really decent careers are on this list next to people like Don Majkowski.
I understand that, and if the purpose of this post was to rigorously defend my rankings I would've done just what you said. However, I don't believe the rankings were the point. I just wanted to tell the stories of these players and bring them a little bit more positive attention. I used the ranking as a framing device to do that more than I used it for anything else.
Most of the time when the stats in the best and second best seasons do not look too different, the difference is injuries. The top season had them producing at the level for a full season. The other season maybe for a half or a third, like Nick Foles. Perhaps I should've put the amount of touches each player had in with the statistics to make this a little bit clearer.
Love/hate seeing so many of my Buccaneers QBs on this list, haha, great writeup -- the Majik Man was awesome in fantasy football!
Thank you for commenting David!
I love that you used the term love/hate, because even as I was writing the list, I could not decide whether being a team with lots of one-off QBs ought to be construed as a compliment or an insult. Of course, given the choice, the Bucs would've just drafted Brett Favre or Steve Young and spent the whole 1990s not having to worry about it, but having a couple one-off years also means their QB room was better than some teams in the 90s, who just never had a good QB ever.
The one-offs tended to come from those teams with subpar histories at the QB position, but there are also teams that tend to be really poor (like New Orleans) who do not show up at all on this list. I'm not sure what any of this means, but for some reason the Buccaneers seem to REALLY attract this type of player. They're the best in the whole league at finding them. That much is for sure. I'm just not sure whether that's a compliment or not.
BTW, there would be no feeling better than being the man who drafted Don Majkowski to a fantasy league in 1989. I understand a similar feeling (being a Jaguar fan) with Mark Brunell in 1996. You get to walk around feeling like the biggest genius on Earth LOL.
I'm not a Washington fan, but I am still annoyed to this day by the way they handled Griffin's knee injury. Of all the names on this list, I think he could have most easily avoided one-off status, with just a little bit of common sense by the organization.
What a fun list. I think it just goes to show that, even with all the preparation and hard work that goes into it, football really is a very unpredictable game with relatively small sample sizes.
What's up buddy? I don't believe we've had the pleasure of speaking on here before. Good to meet you! Thanks for commenting.
No question about it. RG3's injury is typical Washington self-sabotage. The re-aggravation in the playoff game didn't need to happen. I don't know if it was Robert's choice to play and the organisation wasn't smart enough to overrule, or if the team wasn't smart enough to realise pressuring him to play wasn't a good idea. Either chain of actions is not acceptable at the NFL level anyway, so it doesn't matter which is the truth. Washington botched this one big time.
However, there is some personal responsibility on Robert too. People tend to forget that the Chargers did almost exactly the same thing to Philip Rivers that the Skins did to RG3, because Rivers is going to the Hall of Fame. What happened to Robert is that he went down with a horrible leg injury, and came back with the worst pocket presence in the NFL. Other QBs that went down with horrible leg injuries (Trent Green, Philip Rivers, Tom Brady, etc.) did not entirely lose their footwork in the pocket. Robert did.
I'm not sure of the reason for that, and it doesn't change the fact that his injury (and therefore, this entire scenario) was preventable, but I don't think there's a direct line of causality from the injury to the untenable footwork (if there is a line, it's one that the true greats were able to overcome, but Robert couldn't). Something else happened in the middle that made Robert unable to stand in the pocket anymore after 2012. I'm not sure what that something is. It certainly isn't public.
I still think I agree with you that Robert's one-off status (along with Elvis Grbac's, all he had to do was stay in Kansas City) is the most easily preventable on this whole list, but it's not 100% the Skins' fault. Maybe 75 or 80.
As far as football sample size goes, you are correct. Individual seasons are not long enough to figure out whether players (and teams) are good or not. QB sample sizes most often are between 550-750 touches per season, roughly the same amount of plate appearances a batter gets in baseball, and we also see one-off baseball seasons all the time. Using the midpoint, a sample size of 650 is clearly not enough to definitively determine that somebody is good at football (or baseball) or not.
In 2006,Rex Grossman's worst game was probably New Year's Eve, when he completed more passes to Packers (3) than to his own team, the Bears (2). That's been one of my favorite single-game NFL stats for almost 20 years.
It's true! It's not in the consideration for the worst game in history, because the Bears were playing for nothing, and Rex was only going to play a half whether he was great or awful, but if those weren't the circumstances that could've devolved into another of the worst games in the league courtesy of Rex Grossman. It could've been just as bad as Arizona.
I loved this piece! Seeing the Vikings' Daunte Culpepper and Case Keenum (and Gus Frerotte for that matter) makes me wonder if 2024 Sam Darnold will end up with a place on a future edition of this list
I ran the numbers just for fun. If the 2024 season ended right now (with Sam's current numbers) he would go into 10th place, just above Daniel Jones, but his continued success at the NFL level depends on whether he gets another chance at the NFL level.
I don't know how hard the Vikings intend to push JJ McCarthy, but with how well this season is going, I can't see a possibility of the team not finding a way to get Sam extended. That seems like a quick and easy way to lose the locker room, to let the QB go after a season that's gone so well. However, as we see by this list, it would not be the first time a QB is let go and never given another chance, even after being a fringe top ten QB.
That archetype (a QB that's maybe top ten, maybe not) seems to be a thing the NFL has always had a really tough time figuring out how to value throughout its history. I hope that Sam does not fall into that whirlpool, but it's entirely possible that he will.
Really honored by the mention! Love how this was approached - especially fun to see how many of these one-offs clustered around specific teams/eras. Also, as a Packers fan, that Wright->Majkowski->Favre sequence was interesting to read about. Never realized we had such a statistical anomaly of one-off seasons right before the stability of the Favre era began. Great deep dive into these forgotten seasons.
You're welcome buddy! All I did was tell the truth. Your lists about useless statistical anomalies inspired me to create my own, but I had to add my own little bit of storytelling flare to it, telling at least a bit of the story of every season involved here. I thought once I was about halfway done that I should've asked if you wanted to do a feature, maybe write one or two of the entries. It seems close enough to your niche, but oh well. I was too deep in my that point. This won't be the last time I get to ask for your help.
You're right that there did seem to be clusters of one-off players. 2017-2020 all had at least one representative on this list, but only one since. Same with there being one every year from 1993-2001, but then there being only two (both in the 2004 NFC) in the next four years. I must admit the 1990s feature a lot of one-off talent at the QB position, including several that fell just short of the cut-off for inclusion. I'm not exactly sure why that is. Maybe it requires some further study. We'll see.
As far as the Packers, my mind was blown when I saw the results. No way they bounced from one one-off straight into another. They also had Lynn Dickey before that, a man whose 1983 season comes in 36th if I would've made my list just slightly bigger. What was in the water up there in the 1980s? There are teams that are just bad at QB, and keep having to wander into one-off QBs (90s Bucs, Bears in general), but never back to back. I would've been terribly scared once Brett Favre broke out in 1994, waiting for him to get injured or something. Like you said, he definitively broke that curse, but with each new QB that comes in, you never know if it's a one-off or not.
Thank you for the compliments buddy! I'm glad you enjoyed it.
The 1994-95 cluster really grabbed me too - Erickson finding magic with the '94 Bucs, then Kramer and Mitchell both peaking in '95 with the Bears and Lions. What's wild is all three were journeymen who temporarily hit their stride just before, or during, the 1995 expansion changed the landscape. The fact that Mitchell led the Lions to the playoffs while Kramer put up the Bears' best QB season in decades, but Chicago missed the postseason, really shows how competitive that '95 NFC Central was.
For sure would be up for collaborating on something in the future if you've got ideas. These statistical oddities are endlessly fascinating, especially when there's a great story behind the numbers.
I entirely forgot about expansion. You're right to bring that up, but in my opinion, it must have something to do with the invention of free agency. I have no proof for that whatsoever, but free agency gets brought in in 1994 (what do you know?), and all of a sudden one-off QB performances shoot to the moon. I don't know what the two would have to do with each other, or why one-off performances still would've gone through spurts like they have. Free agency has never left., but I feel it must have something to do with it.
BTW, how about the first two free agent QBs in NFL history (Scott Mitchell and Erik Kramer) both being top three QBs in 1995 to get the all-new open QB market off the ground? I can almost smell a conspiracy in there somewhere LOL.
You better believe it buddy, I'm going to be talking about the 1995 NFC Central at some point. The year the three best QBs in the NFL were all in the same division. Two of them are here, the other is league MVP Brett Favre. Never been done before or since. I don't know when I'm going to talk about it, and I don't know how, but it's too good a story to stay buried.
Your point about free agency really got me thinking. Looking at this list, it seems like system/team changes might be a huge factor in these one-off seasons. Kramer and Mitchell lighting it up in '95 as early free agency successes is fascinating, but even before that, you see a pattern of QBs finding temporary magic in new situations - Hostetler with the Raiders in '93, Foles in 2013 when the Eagles brought in Chip Kelly.
The early free agency era feels especially ripe for these performances because teams were still figuring out how to evaluate QB fits. You've got this perfect storm of:
- QBs getting their first real chance to choose systems
- Teams experimenting with different offensive philosophies
- Less analytical evaluation of QB skills/fit
- Coaching staffs perhaps being more willing to adapt to new QBs
What really stands out is how few recent seasons make the list - only 5 (if I counted right) out of 30 from 2017 or later (Jackson, Tannehill, Trubisky, Keenum, Jones). Makes me wonder if modern analytics have made these lightning-in-a-bottle seasons harder to find. Teams now have so much data on QB performance in different schemes, pressure situations, throw types, etc. The kind of statistical anomaly that marked someone like Derek Anderson's 2007 seems less likely when teams can break down every aspect of QB play.
The 1995 NFC Central clustering is even more interesting through this lens - you've got this unique moment where free agency is new, teams are still figuring out how to build around QBs, and suddenly three guys hit career peaks simultaneously. It feels like a time capsule of an era when these breakout seasons were more common, before advanced metrics and sophisticated player evaluation changed how teams develop and assess QB talent.
This all makes me wonder if we'll see fewer one-off seasons going forward as teams get better at evaluating QB fit. And if factors like scheme evolution and coaching changes keep creating opportunities for these anomalies.
The relative scarcity of recent examples suggests the game might be changing in this regard.
If we go by ten year increments:
1980-1989: 6
1990-1999: 9
2000-2009: 6
2010-2019: 7
2020-present: Ryan Tannehill and Daniel Jones
I think you're right on both counts. There's a definite spike in the 1990s, and a lull right now. It's entirely possible that the invention of free agency created a market inefficiency, and (for some reason) the QB graveyards in Detroit and Chicago were able to capitalise on it better than anybody else could, using it to steal great seasons where they really weren't deserved, sort of like Moneyball in football.
QB free agency is still good at doing this. The 2023 Bucs stole a good season when they really weren't positioned to have one via the use of QB free agency, as did this year's Vikings.
I think it's also the case that some teams are better at the open QB market than others, as the Bucs keep doing this (Brad Johnson, Jeff Garcia, Tom Brady, Baker Mayfield are all free agent QB acquisitions that have done great things for them, with only the first round draft picks Josh Freeman and Jameis Winston breaking the chain), whereas other teams have not yet signed a QB free agent worthy of note, despite free agency having been around for 30 years now. Some (like Green Bay) because they haven't needed to, and others (like Cleveland or Houston) where it really would've helped, but they keep picking the wrong guys.
I think there are two big reasons for the lack of one-offs these days. The first is what you said. Evaluation is likely better than it used to be. There is a second though, and it's that the league is filled with more bad QBs in starting spots these days. Players aren't given chances like they used to be, unless they are first round picks. For a previous article, I compiled a top of the head list of players who would've never gotten a chance to play in the modern NFL. Here it is:
Warren Moon (the NFL would never take a CFL QB these days), Steve Young after flaming out in TB, Vinny Testaverde after flaming out in TB, Troy Aikman might’ve been run out of Dallas after starting his career with two awful seasons, Jeff Hostetler (backup for that long these days? No way), Mark Brunell, Randall Cunningham (to come back in 1998), Trent Green (same as Hostetler), Rich Gannon, Jeff Garcia, Jon Kitna, Tony Romo.
All of these men except Kitna (sixth in 2003) were top five QBs in the NFL at one point or another. None of them would've done anything if they came into the NFL today, and these are just the great players. Imagine the ones that only have one great season in them that we're discussing here.
I think the culture of tanking that exists in the NFL today prevents a lot of good QB talent from seeing the light of day, because a good QB means winning, which these days is construed as a bad thing (just ask the Giants). This also prevents one-off seasons like this from happening, because teams don't want them. With a middling rookie, you will win fewer games than with a veteran having his best career season. Like we're seeing now with the Vikings, a vet having his best career season will win you games. If you're a team like the Giants or Raiders or Patriots, who are trying to lose, why would you want a good vet on your team? That's contrary to the purpose.
Sam Darnold is a fringe top ten guy right now, and nearly got no chance to play. Imagine all the others that may be just as good, but are not given chances, in favour of unqualified rookies. I think this is also a big reason for the lack of one-offs these days. Bad rookies and second year players are always bad. Nothing one-offlike about that, and that archetype of QB takes up about a third of all starting spots in the league right now.
Fascinating breakdown by decade. That 90s spike really shows how free agency created this window of opportunity for QBs to find their footing.
Your point about modern teams preferring to lose with rookies than win with veterans really hit home. Geno Smith feels like the last of a dying breed - a Rich Gannon type who needed years to develop before finding success. The fact that he's the exception now rather than the rule shows how the tanking mindset has changed QB development.
Looking at those names of past QBs who wouldn't get chances today, it's striking how many backup/fringe starters we have now who show flashes but never get extended looks. Gardner Minshew and Taylor Heinicke put up decent numbers when called upon, but seems like teams would rather roll the dice on high picks. Sam Darnold finding some success feels almost accidental - like teams are so focused on either having a superstar QB or tanking that the middle ground of developing solid veterans has disappeared.
Makes me wonder if we'll ever see another cluster of one-off seasons like the 90s again with the current incentive structure. Even when veteran QBs play well now, it almost feels like teams view it as a problem rather than an opportunity.
I love that this list includes big names and QBs left in the dustbin of NFL history.
I'm curious about your formula! How did you math this out? Did you deviate from the formula results in your final rankings?
As far as the final rankings, no. Because this wasn't really about the final rankings, more about the stories therein, the final rankings were exactly what my formula gave me. I didn't allow personal opinion to creep into that at all.
If I did, Derek Anderson still would've been one. Don Majkowski still would've been two. They are undeniable, the clear GOATs in terms of one-off status, but below two the order would've been slightly different if I were ordering the list personally. I have the same biases as most people. The great seasons (Neil Lomax, Rex Grossman, Scott Mitchell, Erik Kramer, etc.) would've ended up higher, and the usable years out of typically unusable players would've been lower, but then it wouldn't be measuring in terms of one-offness would it?
As far as the formula used, it begins with a rather simple regression framework to assign a score to each QB season. I combed the internet for as many QB ranking lists as I could find, and ended up with a good sample going back to 2008. I used this data to map stats like EPA/Play, CPOE, ANY/A, DVOA, DYAR, etc., plus adjustments for play count (to prevent overrating rate stats), plus adjustments for offensive era (to prevent overrating every QB that played post-2004), to a QB ranking list, which gives every season a score of where they theoretically would belong on a ranking list of QBs in their season, if every season featured exactly the same leaguewide QB quality.
Typically, (obviously) the score lands between 1 and 32. The lower the better. I did not bother to actually bind the score between one and 32, because all that would do is damp the results and add more mathematical computation, without truly changing anything. For instance, my framework believes the best season of all time is Peyton Manning's 2004, with about a -8 on a 1 to 32 scale, indicating that it was well above the calibre of a typical top QB in the league type season.
This gives me a score for every season a QB has ever had since 1980. From here, the formula for their one-off score is to take every QB who started for at least two seasons at the NFL level since 1980, and simply subtract the score of their best season from the score of their second best season. They do not need to be consecutive.
For instance, Lamar Jackson's 2019 scored about a three, indicating that in a normal season, he wouldn't have won MVP, but it is the lowest 2019 score of anybody, so he did deserve it. His second lowest score came in 2024, and was about 8.75. Therefore, his one-off score is 5.75. Same thing for Don Majkowski. His 1989 indicated he would be about 5.78 on a QB ranking list if 1989 was exactly typical. His next best score was 24.53 in 1988, hence his 18.75 one-off score.
The reason my formula could not handle Derek Anderson was that my initial formula involves a log transform in its process, and Derek Anderson outside of 2007 consistently gives negative numbers where they are not allowed, such is the depth of just how poor his play was. I had to finally tweak the formula to actually be able to put him on the scale, which gives a 2007 score of 12.39, and a next best score of 33.54 in 2008. The only person in the history of the NFL to never come within 20 spots of his best ranking on a hypothetical QB ranking list.
Funnily enough, the least one-off QB in NFL history by my formula is Tony Romo, whose 5.31 score in 2007 and 5.42 in 2014 give him a one-off score of 0.11. One tenth of one position on a ranking list. Nothing more to that. I just thought it was funny in how it demonstrates how the formula works.
I hope it makes sense in the way I've explained it!
This makes a ton of sense. Thank you. I wish you had put that information in the article, including the formula, because the stats chosen to highlight the players don’t look all that different in some cases. It also would help to ground how some guys who had really decent careers are on this list next to people like Don Majkowski.
I understand that, and if the purpose of this post was to rigorously defend my rankings I would've done just what you said. However, I don't believe the rankings were the point. I just wanted to tell the stories of these players and bring them a little bit more positive attention. I used the ranking as a framing device to do that more than I used it for anything else.
Most of the time when the stats in the best and second best seasons do not look too different, the difference is injuries. The top season had them producing at the level for a full season. The other season maybe for a half or a third, like Nick Foles. Perhaps I should've put the amount of touches each player had in with the statistics to make this a little bit clearer.