The commonality in choosing Robinson over Allgeier and Penix over Cousins are the new guys are better athletes, maybe even far better athletes. I don't think there is anything nefarious in the choices, and that the Falcons know they are making the wrong decision and aren't swallowing their pride, as they should. My gut tells me your opinions about both players are spectacularly wrong, meaning the Falcons will do much better in the near future following their program than they would following yours. I think Robinson's athleticism will eventually win out, and deployed the right way, he could have some Saquon Barkley 2024, OJ Simpson 1973 in him. But beyond that, instead of browbeating the Falcons and saying they act in bad faith, you should at least acknowledge that your opinion is fully the contrarian one.
I will also say that my read of historical QB statistics is that INTs might be a leading indicator or non-INT data that will then also fall off. Pretty hard to find quarterbacks over a long period of time with INT rates above average and non-INT production above average. It's not sustainable.
Comments like this make me glad that I inserted the warning that the quality would not be of my normal level at the top, because I missed the entire purpose of my argument. With an offseason looming so imminently, Falcons' decisions need not be beneficial in the near future, as in next season. They must be beneficial in the immediate future, as in three days from now. Decisions pertaining to the near future can happen in the offseason. For now, we need to worry about now, and I don't believe benching Kirk Cousins helps this team in their playoff race, which they are very much still alive in.
Which brings me to your arguments. Are Bijan Robinson and Michael Penix better athletes than Tyler Allgeier and Kirk Cousins? Almost certainly, but if a team was looking for a bastion of athleticism at their QB position, they would not have drafted Michael Penix in the first place. Bo Nix would've been much better for that. Even with this, Kirk Cousins is sneakily mobile when he needs to be. We'll see if this is more or less mobile than Penix. Quite frankly, I'm not sure. Penix's player comparison on draft night was Ryan Fitzpatrick after all. It's not like they were comparing him to Michael Vick or anything.
I assume, given this stance, that you're going with the Denny Crum 'don't get me good basketball players, get me guys that can jump. I'll teach them how to play' stance, which is a perfectly acceptable stance to have, although I would tread lightly on it given the Falcons' recent success (or lack thereof) of developing good players. In addition, this athleticism does not make them the better football players this year, which when the offseason is coming in three games' time, this year is all that matters.
There is no correlation at all between rookie QBs playing their rookie season and their future level of play, so the Falcons do not need to 'learn what they have' in Michael Penix. They do not need to get him any experience. All of that is just excuses for losing, and will not be tolerated here.
Framed the way you've framed it, the Falcons do pass the WARP, harming both their rush and pass offence for the sake of this nebulous concept of 'athleticism,' so this is a theoretically acceptable set of preferences, but it would be an odd set to have for a team that's one game out of the playoffs right now, unless perhaps the goal is to miss the playoffs, which I'm absolutely taking into account as a prime suspect for the motivation behind this move.
You may be correct that these decisions improve the Falcons for 2025. That's a guess. I don't know the answer. Nobody knows the answer, but my fundamental stance is that decisions for 2025 should be handled in the 2025 offseason. There is no need whatsoever to do that now, because this team has a real playoff chance still. Once again, NO correlation whatsoever between a rookie QB playing in his rookie year and his future level of play. That's a proven fact. If there was any correlation at all, even an extremely weak one, that would sway my opinion on this, but since there isn't, the Falcons are going from a top 15 QB in the NFL to a rookie who is almost certain to be worse. Why would you do this?
I understand the final sentence of your first paragraph David. I really do, but I simply refuse to acknowledge that 'trying your best to win and make the playoffs in 2024' is a contrarian opinion. Tanking has become such an accepted thing in North American sports that an article like this criticizing a team for (in my opinion) knowingly lessening their own playoff chances is absolutely a contrarian opinion, but damn it. I just can't bring myself to accept that reality, and since I am my own media company here, I don't have to.
As far as your INT argument, it isn't true. First of all, Kirk Cousins has not had an INT rate above average for a long time. He's had it for five games, and even if it did last for a a whole season, Brett Favre spent 2005 leading the league in turnovers, and went right back to being a top ten guy afterwards. Eli Manning in 2013, exact same phenomenon. Josh Allen got bad turnover luck two seasons in a row, in 2022 and 2023. In 2024 (in my opinion), Josh has slightly regressed as a player, and yet has become one of the best in the league again at avoiding INTs. If you for some reason think Kirk Cousins is fundamentally different to all of these players, let's look at the QB who (by my similarity measures) is the most similar to Kirk Cousins in NFL history, Tommy Kramer.
Compare Tommy's 1985 to his 1986, and you'll see that, even in older ages, turnover luck is just that. Luck, and it will pass if you give it long enough. That's why I claim that the Falcons are undercutting Kirk on purpose, deliberately not giving his bad luck time to pass, in an effort to not generate any QB controversy around starting Penix next season, which even by then may or may not be the best move for them.
Also David, I'm not sure if you read the series, but there's a QB named Trent Green who was never that great at avoiding turnovers (exempting 2003. That was his worst prime season anyway), but was nevertheless one of the elite QBs in the league for a long time. I'm not saying Kirk Cousins' playstyle can sustain this high INT, high value on non-INT statistical profile, but there are some that can.
Even though it is not my post, I appreciate you sharing your thoughts, David.
Robinson and Penix may have better physical tools but that doesn't mean they will necessarily be more effective players. And even if they do become better players in the long run, that doesn't mean right now Cousins and Allgeier aren't their best options this season.
To your INT point, I think you have the causality arrow backwards. INT are strong explanatory stats but they have very little predictive power, which is why NY/A is just as predictive as ANY/A attempt despite leaving out TDs and INTs and why turnover margin for teams often regresses to the mean in the following season.
If Cousins were struggling to make good reads and throw accurate passes, then we would expected higher interception rates. However, his NY/A is still a solid 6.5 in these past 5 games which means he is still completing passes that are gaining solid yardage and avoiding sacks well. Additionally, Cousins has been good at avoiding INTs his whole career.
Look at Dak in 2022 compared to 2023. His passer rating and ANY/A were lower than usually do to high INT rates, but his NY/A was still strong. The following year Dak was an MVP candidate but his NY/A was only 0.17 yards better than in 2022. Dak was more or less the same guy but his interceptions went way down, which was the primary driver of his improved results.
To add on: Dak's personal success rate in 2022: 52.5%
In 2023: 52.6%.
One thing I've noticed in analysing this phenomenon is that inexplicably bad seasons like this seem to coincide with inexplicable lapses in accuracy. For instance, despite the identical success rates, Dak's CPOE in 2022 was 0.1. In 2023, it was 3.9. Pass accuracy and success rate have a much smaller correlation than results and success rate do, and high success rates combined with uncharacteristically bad pass accuracy numbers always seem to be present for these difficult to explain low points.
Even in THE turnover luck season, Jameis Winston's career success rate coming into 2019 was 49.6%, and his success rate in 2019 was 49.1%. However, the throw accuracy was just 0.7, compared to a career total of 2.2. I'm not sure if there's anything to take from this, other than the accuracy tends to bounce back, and the turnovers tend to go away again, AS LONG AS the personal success rate did not dip with it.
For instance, Aaron Rodgers' 2021 rate of 52.1% contrasts starkly with 2022's 46.6%, indicating that this was not a temporary dip in form, more of a structural break, and his Jets performance possibly could've been predicted. Also, Dak's precipitous decline to a 45.3% personal success rate in 2024 indicates that this falloff is perhaps more meaningful than all the other falloffs he's had, although his history of falling off and coming back does cause me to pause some on that hypothesis.
“Imagine being a baseball team, and deciding a player isn’t a starting calibre player anymore based on a sample of 184 plate appearances, and at that, a 184 PA sample where his exit velocities and barrel rates are not out of line with his career norms. That is what the Atlanta Falcons have just done, and that is what’s so shameful”
This seems to me to be a category error. Insofar as I can tell, all of the tools for evaluating QBs are outcome dependent formulas; exit velocity is an outcome independent measurement, requiring no human calculation or algorithms to establish.
If Atlanta Braves third baseman Austin Riley goes 184 at bats with low batting average and slugging percentage, EV is a simple measurements to supplement management’s observation about whether there has been a decline in his bat speed, about whether he is on time and in balance. It is physically impossible to have a high EV with poor contact and a slow bat.
Watching Cousins against the Raiders, he looked like he was guiding his passes instead of stepping into them. He looked tentative, which the announcers attributed to lack of confidence and others to possible injury. (He got banged up against the Saints and spent one day on the injury list due to a sore shoulder.) I wish there was something as straightforward as EV that helped confirm or refute the observation, but I’m not seeing anything that provides any illumination.
If Cousins has a physical or psychological problem that keeps him from committing to zipping passes that need to be zipped — even if management just has a good faith basis for making that conclusion — I guess I’d be reluctant to call the decision to replace him shameful.
Perhaps my allegory was flawed, because you're correct that even personal success rate is dependent on the outcome of the play. The reason I choose to use it is because it's less dependent on performance from non-QB players than any other stat except sack rate (which QBs mostly own on their own), which in Kirk's case did go up the tiniest bit in this stretch but remained better than league average. However, even if the exact analogy was incorrect, the simple assertion that Atlanta replaced a QB based upon poor results in a small sample during which his underlying peripherals (such as they exist in this game) did not change remains true.
Success rate can predict EPA/Play with an R-Square of about 80% at the season level, so while it cannot be treated as an underlying peripheral in all cases, because it is indeed determined by EPA, in the case that somebody's results dramatically plummet, but their success rate stays the same, it cannot continue. It is obvious that the poor results are caused by an extreme negative skew as a result of some really bad negative outliers. If these outliers were sacks, it would be extremely concerning, and I wouldn't really have a problem with this decision, but since they're turnovers (which are luck-based most of the time), I have very little concern for Kirk's future.
One of these things has to give. In general throughout the NFL's history, the success rate tends to hold, and the actual on field results tend to have to bend the knee, and regress (in this case regress upwards) to the mean. We actually saw this a bit on Sunday, where Michael Penix played eerily similar to Kirk Cousins. Fine, but fine is about it. Quite inaccurate but at least very good at sack avoidance, and nevertheless got very good results because he was playing the Giants.
At least against the Giants, Atlanta replaced Kirk Cousins with Kirk Cousins, and it's a crafty move the Falcons have pulled here, because given the very weak competition they're going against in the final three weeks, nobody is going to notice. Blinded by the very good on-field results.
As far as the Raiders game, we're just guessing. It was the worst game of this five game stretch by far, about that there is no guessing, but the cause is just guesswork. The commentators guessed that he was tentative because of a lack of confidence. I would say no wonder, given the fact that the Falcons were intent on putting him in bad positions all night (36 plays for a rush offence working at a 31% success rate is a bad gameplan), and that his aDoT and TTT numbers were not very far out of his ordinary, so perhaps we can just chalk this up to players have bad games sometimes and not think about it any further.
Something was clearly off though, because the Falcons called one solitary first down pass all night, compared to 23 first down rushes, despite the pass game, even while having a bad night, still being successful on 41 percent of all plays, compared to 31 percent for the rush offence. I don't know if they were trying on purpose to ensure Kirk had a bad day by putting him in must-pass spots constantly, if they were trying to send a message, or if we've just watched a real contender for the worst offensive gameplan in NFL history. Once again, we're guessing.
However, one first down pass to 23 first down rushes on your 24 first down plays is not even a play call distribution we would've seen in the 1970s. Regardless of how bad some people think Kirk had played in the weeks before, it was not bad enough to justify this. I smell a rat on Monday night too. However, the alternative of the Falcons just having extraordinarily bad offensive play callers was too viable to include this supposition in the actual article itself.
So yes, regardless of what guess we make about Monday night (mine is a split between either sabotage, or being constantly put in poor positions for less sinister reasons. The masses' is injury or a lack of self confidence), something went wrong there. That doesn't change the fact that Kirk had just played very well against perhaps the best offence in the NFL the week before, and even if he hadn't done that, it doesn't change the fact that the Falcons had made a commitment to get a guy who has always been about the 12th best QB in the league, he played exactly like the 12th best QB in the league, and the Falcons are electing to throw the whole thing away with a horrifying dead cap number to get out of exactly the situation that they wanted to begin with.
This is a massive organisational failure, and terribly unfair to both Kirk Cousins (who got bait and switched) and Michael Penix (who is basically going to be on his big boy contract already due to the Falcons paying both his salary and Kirk's next season, with the depleted roster to match, and is going to have to play like it right out of the gate).
We've seen this season with the Bengals that winning while moving to a big boy QB contract is not easy, and Michael Penix might be as good as Joe Burrow. We don't know, but even if he is, will that change anything about the Falcons? Joe Burrow is not that much better than Kirk Cousins. It's putting Michael in the horrible sink-or-swim position of needing to play like a great QB immediately.
All of this could've been avoided by just letting Kirk take the lumps this season and possibly next one, then cutting him in 2026, keeping Michael away from the public eye for a while, but the Falcons wouldn't do it. My opinion for the reason for this is that Terry Fontenot has been trying (and failing) to make the Falcons winners for a long time. It's very possible that his job is on the line, and so he instigated a decision being made that screws up the timeline for everybody, and stands a real chance of ruining Michael Penix's career by putting too much pressure on him too quick.
Not your usual style, Robbie, but I really liked this! I'll definitely be keeping the concept of WARP in mind, I hadn't heard of that before, but I really like it. One pop-psych/economics idea I'm familiar with is the idea of "econs" or actors that behave in completely rational ways, with the idea being that real people aren't econs. I think there's some ambiguity about whether the Falcons treatment of Allgeier/Robinson and Cousins/Penix is part of a common cause/effect like you suggest (ego), or just the humans in charge behaving irrationally and in contradictory ways (as we all do).
To me, the Cousins situation reads as the Falcons giving up on this season. They're more likely to miss the playoffs than make it and they're not contenders anyway, so why not give their young QB a few reps against some weak defenses to learn more about where he is right now. I guess they signed Cousins as a stopgap until Penix is ready and they're moving up the timeline a bit in a season they're writing off. This happens sometimes with NHL teams, but not normally with teams as close to the playoffs as the Falcons are. Probably still disappointing if you're a Falcons fan, but maybe there's a bit of hope that the future is brighter with Penix than Cousins.
I can cite book after book about revealed preference theory, but generally, it's pretty strong. It can fail in all kinds of different ways in specific circumstances, but in general, it will only fail for three general reasons:
1) Unknown alternatives. An agent does not directly know all alternatives to the choice they are making, and once shown a better alternative, they reveal a different preference. That is not the case here. NFL QBs and RBs are not like grocery store bundles. There are only a finite number of them, and I'm willing to assume that an NFL GM knows all of them.
2) Time inconsistency. Preferences will have changed by the time the two ends of the WARP comparison are being made. This is obviously not true, as the Falcons are making the Bijan Robinson decision and the Michael Penix decision at the same time.
3) Irrationality. Somebody likes A better than B, and B better than C, but for some reason will pick C over A. You are right that this is the only condition on which my use of the WARP in this specific context can fail. Perhaps I've just built a habit of ignoring this condition, as it's somewhat demeaning to simply accuse somebody of irrationality every time they make a choice that doesn't make sense. It's a much better use of time to try to devise a set of preferences that fit their actions, because most of the time humans are not irrational.
This is what I did, and I found a set of preferences that work given the data I have, that the Falcons were not willing to admit they made a mistake on a first round pick, and were willing to make their team worse to avoid doing so. My friend David Harris devised a different set, that the Falcons are willing to get worse in both the pass and rush offence in order to get younger and more athletic. I don't know why any team would do this midseason, which is why it wasn't my hypothesis, but it's technically valid, as it does clear the WARP. There are other ideas too.
What cannot be argued is that the Falcons definitively ARE NOT swallowing their pride to make the team better. If they were willing to do that they would've benched Bijan, so to me that isn't an acceptable explanation for what happened to Kirk, although people that aren't economists will give it anyway LOL.
You're exactly right Aaron. They are giving up on the season. That's why this is so shameful. All this team needs is one Washington loss to be right back in the hunt. Washington plays Philadelphia this week. To me, it's absurd to give up on the season right now. If the Falcons were truly acting in good faith, they would've waited until after that Philadelphia game. Instead, they chose to pull the trigger now, which casts doubt to me as to whether they wanted to make the playoffs in the first place.
It really is trading a cow for magic beans. If Michael Penix is as good as Kirk Cousins is, this move will have been a home run. Why not just stick with what you had and avoid having to hit a HR on the move? There is no correlation between a rookie QB playing in his first season and his future level of play, no matter how many coaches and media try to tell you there is one, so to me there is no benefit to playing a rookie QB. If you're making decisions for the good of 2025, there is no rush. You can do that in 2025. There's no need to tank 2024 like this.
That's the crux of this whole thing. My anger is not about 2025 and beyond. Kirk Cousins is 35+ after all. It's about right now, where Atlanta is trading a top 15 QB in the NFL for magic beans.
Ha, it's interesting you say you like to ignore reason #3 WARP can fail, because I feel like sometimes I'm too quick to handwave things I don't understand away as people acting irrationally. I agree, regardless of the reason, benching Cousins is not about swallowing pride to make the team better.
Two things I didn't realize when I wrote my comment were that Cousins was signed before Penix was drafted and drafting Penix at #8 OA seems to have been a bit of a reach. Knowing that, I think your theory makes a lot more sense. It's definitely true in the NHL that teams will boost their first round picks longer than they should and over better players.
Two things that still don't make sense to me are why they signed Cousins in the first place if they liked Penix so much they were planning to reach for him and why they seem to have given up on the season already. I suppose time inconsistency could play a role in both of those: their evaluation of Penix could have changed after they signed Cousins and they decided future seasons were more important than the current season so worthwhile to get Penix to get some reps in a real game (this is a very hockey way of thinking about things, I'm not sure if it applies to the NFL), but I'm just playing devil's advocate at this point. Whatever the reason, they're trashing this season and blaming Cousins for it.
We can indeed blame time inconsistency for ending up with this situation at all. The Falcons didn't know that Penix was going to fall to them in the Draft and (because Atlanta) somehow couldn't figure out that reaching for a QB when they had guaranteed money tied up in one for years into the future was a bad idea. Unless Penis turns out to be a much better player than Kirk Cousins, this has been a mistake from the start, because merely replacing Kirk Cousins is not good enough when you had to spend a first round pick to do it. The NFL is not like most other leagues. Years of age are simply not that valuable. Teams get older to get better all the time.
As to why they're giving up on the season, it would be hard to bench a QB who just took your team to the playoffs, wouldn't it? There's a reason the Vikings were trying so hard NOT to start Sam Darnold this year.
There's a long and storied history in the NFL of teams throwing in their rookie QBs, strictly to prevent a playoff calibre team from making the playoffs. Just in recent memory, Ryan Fitzpatrick was benched in the middle of 2020 despite generating 0.245 EPA/Play in 2020, just because the owner didn't think the tank was over yet. They were fined draft picks for that one. That was extremely egregious. The Baltimore Ravens tried to do the same thing in 2018, swapping out Joe Flacco's best season in years for a Lamar Jackson that was not near ready, but failed and made the playoffs anyway, getting killed in their playoff game.
The 2017 Browns' coaches had incentives in their contracts for losing, so played Deshone Kizer over Cody Kessler and went 0-16. I think there were fines over that too.
As you can see, deliberately losing in this way for this reason is not common, but is also not unprecedented in the NFL either.
I love it, and I completely agree.
The commonality in choosing Robinson over Allgeier and Penix over Cousins are the new guys are better athletes, maybe even far better athletes. I don't think there is anything nefarious in the choices, and that the Falcons know they are making the wrong decision and aren't swallowing their pride, as they should. My gut tells me your opinions about both players are spectacularly wrong, meaning the Falcons will do much better in the near future following their program than they would following yours. I think Robinson's athleticism will eventually win out, and deployed the right way, he could have some Saquon Barkley 2024, OJ Simpson 1973 in him. But beyond that, instead of browbeating the Falcons and saying they act in bad faith, you should at least acknowledge that your opinion is fully the contrarian one.
I will also say that my read of historical QB statistics is that INTs might be a leading indicator or non-INT data that will then also fall off. Pretty hard to find quarterbacks over a long period of time with INT rates above average and non-INT production above average. It's not sustainable.
Comments like this make me glad that I inserted the warning that the quality would not be of my normal level at the top, because I missed the entire purpose of my argument. With an offseason looming so imminently, Falcons' decisions need not be beneficial in the near future, as in next season. They must be beneficial in the immediate future, as in three days from now. Decisions pertaining to the near future can happen in the offseason. For now, we need to worry about now, and I don't believe benching Kirk Cousins helps this team in their playoff race, which they are very much still alive in.
Which brings me to your arguments. Are Bijan Robinson and Michael Penix better athletes than Tyler Allgeier and Kirk Cousins? Almost certainly, but if a team was looking for a bastion of athleticism at their QB position, they would not have drafted Michael Penix in the first place. Bo Nix would've been much better for that. Even with this, Kirk Cousins is sneakily mobile when he needs to be. We'll see if this is more or less mobile than Penix. Quite frankly, I'm not sure. Penix's player comparison on draft night was Ryan Fitzpatrick after all. It's not like they were comparing him to Michael Vick or anything.
I assume, given this stance, that you're going with the Denny Crum 'don't get me good basketball players, get me guys that can jump. I'll teach them how to play' stance, which is a perfectly acceptable stance to have, although I would tread lightly on it given the Falcons' recent success (or lack thereof) of developing good players. In addition, this athleticism does not make them the better football players this year, which when the offseason is coming in three games' time, this year is all that matters.
There is no correlation at all between rookie QBs playing their rookie season and their future level of play, so the Falcons do not need to 'learn what they have' in Michael Penix. They do not need to get him any experience. All of that is just excuses for losing, and will not be tolerated here.
Framed the way you've framed it, the Falcons do pass the WARP, harming both their rush and pass offence for the sake of this nebulous concept of 'athleticism,' so this is a theoretically acceptable set of preferences, but it would be an odd set to have for a team that's one game out of the playoffs right now, unless perhaps the goal is to miss the playoffs, which I'm absolutely taking into account as a prime suspect for the motivation behind this move.
You may be correct that these decisions improve the Falcons for 2025. That's a guess. I don't know the answer. Nobody knows the answer, but my fundamental stance is that decisions for 2025 should be handled in the 2025 offseason. There is no need whatsoever to do that now, because this team has a real playoff chance still. Once again, NO correlation whatsoever between a rookie QB playing in his rookie year and his future level of play. That's a proven fact. If there was any correlation at all, even an extremely weak one, that would sway my opinion on this, but since there isn't, the Falcons are going from a top 15 QB in the NFL to a rookie who is almost certain to be worse. Why would you do this?
I understand the final sentence of your first paragraph David. I really do, but I simply refuse to acknowledge that 'trying your best to win and make the playoffs in 2024' is a contrarian opinion. Tanking has become such an accepted thing in North American sports that an article like this criticizing a team for (in my opinion) knowingly lessening their own playoff chances is absolutely a contrarian opinion, but damn it. I just can't bring myself to accept that reality, and since I am my own media company here, I don't have to.
As far as your INT argument, it isn't true. First of all, Kirk Cousins has not had an INT rate above average for a long time. He's had it for five games, and even if it did last for a a whole season, Brett Favre spent 2005 leading the league in turnovers, and went right back to being a top ten guy afterwards. Eli Manning in 2013, exact same phenomenon. Josh Allen got bad turnover luck two seasons in a row, in 2022 and 2023. In 2024 (in my opinion), Josh has slightly regressed as a player, and yet has become one of the best in the league again at avoiding INTs. If you for some reason think Kirk Cousins is fundamentally different to all of these players, let's look at the QB who (by my similarity measures) is the most similar to Kirk Cousins in NFL history, Tommy Kramer.
Compare Tommy's 1985 to his 1986, and you'll see that, even in older ages, turnover luck is just that. Luck, and it will pass if you give it long enough. That's why I claim that the Falcons are undercutting Kirk on purpose, deliberately not giving his bad luck time to pass, in an effort to not generate any QB controversy around starting Penix next season, which even by then may or may not be the best move for them.
Also David, I'm not sure if you read the series, but there's a QB named Trent Green who was never that great at avoiding turnovers (exempting 2003. That was his worst prime season anyway), but was nevertheless one of the elite QBs in the league for a long time. I'm not saying Kirk Cousins' playstyle can sustain this high INT, high value on non-INT statistical profile, but there are some that can.
Even though it is not my post, I appreciate you sharing your thoughts, David.
Robinson and Penix may have better physical tools but that doesn't mean they will necessarily be more effective players. And even if they do become better players in the long run, that doesn't mean right now Cousins and Allgeier aren't their best options this season.
To your INT point, I think you have the causality arrow backwards. INT are strong explanatory stats but they have very little predictive power, which is why NY/A is just as predictive as ANY/A attempt despite leaving out TDs and INTs and why turnover margin for teams often regresses to the mean in the following season.
If Cousins were struggling to make good reads and throw accurate passes, then we would expected higher interception rates. However, his NY/A is still a solid 6.5 in these past 5 games which means he is still completing passes that are gaining solid yardage and avoiding sacks well. Additionally, Cousins has been good at avoiding INTs his whole career.
Look at Dak in 2022 compared to 2023. His passer rating and ANY/A were lower than usually do to high INT rates, but his NY/A was still strong. The following year Dak was an MVP candidate but his NY/A was only 0.17 yards better than in 2022. Dak was more or less the same guy but his interceptions went way down, which was the primary driver of his improved results.
To add on: Dak's personal success rate in 2022: 52.5%
In 2023: 52.6%.
One thing I've noticed in analysing this phenomenon is that inexplicably bad seasons like this seem to coincide with inexplicable lapses in accuracy. For instance, despite the identical success rates, Dak's CPOE in 2022 was 0.1. In 2023, it was 3.9. Pass accuracy and success rate have a much smaller correlation than results and success rate do, and high success rates combined with uncharacteristically bad pass accuracy numbers always seem to be present for these difficult to explain low points.
Even in THE turnover luck season, Jameis Winston's career success rate coming into 2019 was 49.6%, and his success rate in 2019 was 49.1%. However, the throw accuracy was just 0.7, compared to a career total of 2.2. I'm not sure if there's anything to take from this, other than the accuracy tends to bounce back, and the turnovers tend to go away again, AS LONG AS the personal success rate did not dip with it.
For instance, Aaron Rodgers' 2021 rate of 52.1% contrasts starkly with 2022's 46.6%, indicating that this was not a temporary dip in form, more of a structural break, and his Jets performance possibly could've been predicted. Also, Dak's precipitous decline to a 45.3% personal success rate in 2024 indicates that this falloff is perhaps more meaningful than all the other falloffs he's had, although his history of falling off and coming back does cause me to pause some on that hypothesis.
“Imagine being a baseball team, and deciding a player isn’t a starting calibre player anymore based on a sample of 184 plate appearances, and at that, a 184 PA sample where his exit velocities and barrel rates are not out of line with his career norms. That is what the Atlanta Falcons have just done, and that is what’s so shameful”
This seems to me to be a category error. Insofar as I can tell, all of the tools for evaluating QBs are outcome dependent formulas; exit velocity is an outcome independent measurement, requiring no human calculation or algorithms to establish.
If Atlanta Braves third baseman Austin Riley goes 184 at bats with low batting average and slugging percentage, EV is a simple measurements to supplement management’s observation about whether there has been a decline in his bat speed, about whether he is on time and in balance. It is physically impossible to have a high EV with poor contact and a slow bat.
Watching Cousins against the Raiders, he looked like he was guiding his passes instead of stepping into them. He looked tentative, which the announcers attributed to lack of confidence and others to possible injury. (He got banged up against the Saints and spent one day on the injury list due to a sore shoulder.) I wish there was something as straightforward as EV that helped confirm or refute the observation, but I’m not seeing anything that provides any illumination.
If Cousins has a physical or psychological problem that keeps him from committing to zipping passes that need to be zipped — even if management just has a good faith basis for making that conclusion — I guess I’d be reluctant to call the decision to replace him shameful.
Perhaps my allegory was flawed, because you're correct that even personal success rate is dependent on the outcome of the play. The reason I choose to use it is because it's less dependent on performance from non-QB players than any other stat except sack rate (which QBs mostly own on their own), which in Kirk's case did go up the tiniest bit in this stretch but remained better than league average. However, even if the exact analogy was incorrect, the simple assertion that Atlanta replaced a QB based upon poor results in a small sample during which his underlying peripherals (such as they exist in this game) did not change remains true.
Success rate can predict EPA/Play with an R-Square of about 80% at the season level, so while it cannot be treated as an underlying peripheral in all cases, because it is indeed determined by EPA, in the case that somebody's results dramatically plummet, but their success rate stays the same, it cannot continue. It is obvious that the poor results are caused by an extreme negative skew as a result of some really bad negative outliers. If these outliers were sacks, it would be extremely concerning, and I wouldn't really have a problem with this decision, but since they're turnovers (which are luck-based most of the time), I have very little concern for Kirk's future.
One of these things has to give. In general throughout the NFL's history, the success rate tends to hold, and the actual on field results tend to have to bend the knee, and regress (in this case regress upwards) to the mean. We actually saw this a bit on Sunday, where Michael Penix played eerily similar to Kirk Cousins. Fine, but fine is about it. Quite inaccurate but at least very good at sack avoidance, and nevertheless got very good results because he was playing the Giants.
At least against the Giants, Atlanta replaced Kirk Cousins with Kirk Cousins, and it's a crafty move the Falcons have pulled here, because given the very weak competition they're going against in the final three weeks, nobody is going to notice. Blinded by the very good on-field results.
As far as the Raiders game, we're just guessing. It was the worst game of this five game stretch by far, about that there is no guessing, but the cause is just guesswork. The commentators guessed that he was tentative because of a lack of confidence. I would say no wonder, given the fact that the Falcons were intent on putting him in bad positions all night (36 plays for a rush offence working at a 31% success rate is a bad gameplan), and that his aDoT and TTT numbers were not very far out of his ordinary, so perhaps we can just chalk this up to players have bad games sometimes and not think about it any further.
Something was clearly off though, because the Falcons called one solitary first down pass all night, compared to 23 first down rushes, despite the pass game, even while having a bad night, still being successful on 41 percent of all plays, compared to 31 percent for the rush offence. I don't know if they were trying on purpose to ensure Kirk had a bad day by putting him in must-pass spots constantly, if they were trying to send a message, or if we've just watched a real contender for the worst offensive gameplan in NFL history. Once again, we're guessing.
However, one first down pass to 23 first down rushes on your 24 first down plays is not even a play call distribution we would've seen in the 1970s. Regardless of how bad some people think Kirk had played in the weeks before, it was not bad enough to justify this. I smell a rat on Monday night too. However, the alternative of the Falcons just having extraordinarily bad offensive play callers was too viable to include this supposition in the actual article itself.
So yes, regardless of what guess we make about Monday night (mine is a split between either sabotage, or being constantly put in poor positions for less sinister reasons. The masses' is injury or a lack of self confidence), something went wrong there. That doesn't change the fact that Kirk had just played very well against perhaps the best offence in the NFL the week before, and even if he hadn't done that, it doesn't change the fact that the Falcons had made a commitment to get a guy who has always been about the 12th best QB in the league, he played exactly like the 12th best QB in the league, and the Falcons are electing to throw the whole thing away with a horrifying dead cap number to get out of exactly the situation that they wanted to begin with.
This is a massive organisational failure, and terribly unfair to both Kirk Cousins (who got bait and switched) and Michael Penix (who is basically going to be on his big boy contract already due to the Falcons paying both his salary and Kirk's next season, with the depleted roster to match, and is going to have to play like it right out of the gate).
We've seen this season with the Bengals that winning while moving to a big boy QB contract is not easy, and Michael Penix might be as good as Joe Burrow. We don't know, but even if he is, will that change anything about the Falcons? Joe Burrow is not that much better than Kirk Cousins. It's putting Michael in the horrible sink-or-swim position of needing to play like a great QB immediately.
All of this could've been avoided by just letting Kirk take the lumps this season and possibly next one, then cutting him in 2026, keeping Michael away from the public eye for a while, but the Falcons wouldn't do it. My opinion for the reason for this is that Terry Fontenot has been trying (and failing) to make the Falcons winners for a long time. It's very possible that his job is on the line, and so he instigated a decision being made that screws up the timeline for everybody, and stands a real chance of ruining Michael Penix's career by putting too much pressure on him too quick.
That's shameful.
Not your usual style, Robbie, but I really liked this! I'll definitely be keeping the concept of WARP in mind, I hadn't heard of that before, but I really like it. One pop-psych/economics idea I'm familiar with is the idea of "econs" or actors that behave in completely rational ways, with the idea being that real people aren't econs. I think there's some ambiguity about whether the Falcons treatment of Allgeier/Robinson and Cousins/Penix is part of a common cause/effect like you suggest (ego), or just the humans in charge behaving irrationally and in contradictory ways (as we all do).
To me, the Cousins situation reads as the Falcons giving up on this season. They're more likely to miss the playoffs than make it and they're not contenders anyway, so why not give their young QB a few reps against some weak defenses to learn more about where he is right now. I guess they signed Cousins as a stopgap until Penix is ready and they're moving up the timeline a bit in a season they're writing off. This happens sometimes with NHL teams, but not normally with teams as close to the playoffs as the Falcons are. Probably still disappointing if you're a Falcons fan, but maybe there's a bit of hope that the future is brighter with Penix than Cousins.
I can cite book after book about revealed preference theory, but generally, it's pretty strong. It can fail in all kinds of different ways in specific circumstances, but in general, it will only fail for three general reasons:
1) Unknown alternatives. An agent does not directly know all alternatives to the choice they are making, and once shown a better alternative, they reveal a different preference. That is not the case here. NFL QBs and RBs are not like grocery store bundles. There are only a finite number of them, and I'm willing to assume that an NFL GM knows all of them.
2) Time inconsistency. Preferences will have changed by the time the two ends of the WARP comparison are being made. This is obviously not true, as the Falcons are making the Bijan Robinson decision and the Michael Penix decision at the same time.
3) Irrationality. Somebody likes A better than B, and B better than C, but for some reason will pick C over A. You are right that this is the only condition on which my use of the WARP in this specific context can fail. Perhaps I've just built a habit of ignoring this condition, as it's somewhat demeaning to simply accuse somebody of irrationality every time they make a choice that doesn't make sense. It's a much better use of time to try to devise a set of preferences that fit their actions, because most of the time humans are not irrational.
This is what I did, and I found a set of preferences that work given the data I have, that the Falcons were not willing to admit they made a mistake on a first round pick, and were willing to make their team worse to avoid doing so. My friend David Harris devised a different set, that the Falcons are willing to get worse in both the pass and rush offence in order to get younger and more athletic. I don't know why any team would do this midseason, which is why it wasn't my hypothesis, but it's technically valid, as it does clear the WARP. There are other ideas too.
What cannot be argued is that the Falcons definitively ARE NOT swallowing their pride to make the team better. If they were willing to do that they would've benched Bijan, so to me that isn't an acceptable explanation for what happened to Kirk, although people that aren't economists will give it anyway LOL.
You're exactly right Aaron. They are giving up on the season. That's why this is so shameful. All this team needs is one Washington loss to be right back in the hunt. Washington plays Philadelphia this week. To me, it's absurd to give up on the season right now. If the Falcons were truly acting in good faith, they would've waited until after that Philadelphia game. Instead, they chose to pull the trigger now, which casts doubt to me as to whether they wanted to make the playoffs in the first place.
It really is trading a cow for magic beans. If Michael Penix is as good as Kirk Cousins is, this move will have been a home run. Why not just stick with what you had and avoid having to hit a HR on the move? There is no correlation between a rookie QB playing in his first season and his future level of play, no matter how many coaches and media try to tell you there is one, so to me there is no benefit to playing a rookie QB. If you're making decisions for the good of 2025, there is no rush. You can do that in 2025. There's no need to tank 2024 like this.
That's the crux of this whole thing. My anger is not about 2025 and beyond. Kirk Cousins is 35+ after all. It's about right now, where Atlanta is trading a top 15 QB in the NFL for magic beans.
Ha, it's interesting you say you like to ignore reason #3 WARP can fail, because I feel like sometimes I'm too quick to handwave things I don't understand away as people acting irrationally. I agree, regardless of the reason, benching Cousins is not about swallowing pride to make the team better.
Two things I didn't realize when I wrote my comment were that Cousins was signed before Penix was drafted and drafting Penix at #8 OA seems to have been a bit of a reach. Knowing that, I think your theory makes a lot more sense. It's definitely true in the NHL that teams will boost their first round picks longer than they should and over better players.
Two things that still don't make sense to me are why they signed Cousins in the first place if they liked Penix so much they were planning to reach for him and why they seem to have given up on the season already. I suppose time inconsistency could play a role in both of those: their evaluation of Penix could have changed after they signed Cousins and they decided future seasons were more important than the current season so worthwhile to get Penix to get some reps in a real game (this is a very hockey way of thinking about things, I'm not sure if it applies to the NFL), but I'm just playing devil's advocate at this point. Whatever the reason, they're trashing this season and blaming Cousins for it.
We can indeed blame time inconsistency for ending up with this situation at all. The Falcons didn't know that Penix was going to fall to them in the Draft and (because Atlanta) somehow couldn't figure out that reaching for a QB when they had guaranteed money tied up in one for years into the future was a bad idea. Unless Penis turns out to be a much better player than Kirk Cousins, this has been a mistake from the start, because merely replacing Kirk Cousins is not good enough when you had to spend a first round pick to do it. The NFL is not like most other leagues. Years of age are simply not that valuable. Teams get older to get better all the time.
As to why they're giving up on the season, it would be hard to bench a QB who just took your team to the playoffs, wouldn't it? There's a reason the Vikings were trying so hard NOT to start Sam Darnold this year.
There's a long and storied history in the NFL of teams throwing in their rookie QBs, strictly to prevent a playoff calibre team from making the playoffs. Just in recent memory, Ryan Fitzpatrick was benched in the middle of 2020 despite generating 0.245 EPA/Play in 2020, just because the owner didn't think the tank was over yet. They were fined draft picks for that one. That was extremely egregious. The Baltimore Ravens tried to do the same thing in 2018, swapping out Joe Flacco's best season in years for a Lamar Jackson that was not near ready, but failed and made the playoffs anyway, getting killed in their playoff game.
The 2017 Browns' coaches had incentives in their contracts for losing, so played Deshone Kizer over Cody Kessler and went 0-16. I think there were fines over that too.
As you can see, deliberately losing in this way for this reason is not common, but is also not unprecedented in the NFL either.