The Atlanta Falcons are Shameful

A discussion of the Falcons benching Kirk Cousins, and why this decision is such a disgrace.

Okay everybody. Welcome back to my Sports Passion Project, where it’s time for an Atlanta rant.

This is not really my flavour of post, and I understand that, and offer my apologies for it. This is going to come off feeling much more like something

Smayan Srikanth

would do. Nothing against Smayan. He’s a friend, and likely would’ve written this exact post much better than I am going to write it. Nevertheless, I’m going to impersonate him as best I can on this one.

I am not telling any story today. My language will not be as elegant as it normally is, but in my mind the purpose of this Substack publication is challenging and attempting to change sports narratives that make no sense. This compulsion to correct false narratives is what led me to go in deep on Trent Green. It’s what led me to prove in excruciating detail that Tom Brady was not that great of a playoff player, and now it’s what’s forcing me to write this piece about the disaster in Atlanta.

What encouraged me to write this was the following Substack Note, written by

Matt Lombardo

:

First off, I respect

Matt Lombardo

and everything he does. None of this is a shot at him, but as somebody for whom the Atlanta Falcons have always been my NFC team of choice, I cannot disagree more with his stance that benching Kirk Cousins is the right move, and I’ll explain why.

I’m sorry you had to be treated this way Kirk.

To begin, Matt mentions Kirk’s very bad play over his last five games, which is correct. Kirk has been very bad for the last month. In fact, Kirk’s stats (-0.109 EPA/Play, -4.2 CPOE) are even worse than CJ Stroud’s (-0.062 EPA/Play, -3.5 CPOE) since the second week of November, and I’ve been calling for CJ to be benched for weeks.

So why do I defend Kirk, but not CJ?

That requires a leap in logic on my part, but I am prepared to make that leap for several reasons. Some are obvious. Some are nuanced.

I’ve talked about the economic concept of the WARP (Weak Axiom of Revealed Preference) on this publication before. It states that one chain of actions cannot reveal both action A being preferred to action B, and reveal action B being preferred to action A. If any chain of events ever does reveal this, we have judged the consumer’s (in this case, the Falcons’) preferences incorrectly.

In his Note, Matt commends the Falcons for swallowing the money and the pride to do what’s best for the team on the field, which is funny, because everybody who’s been looking at my Notes will know that I have been deriding the Falcons all season for failing to do that in a different part of their offence.

In the first round of the 2023 NFL Draft, the Atlanta Falcons selected RB Bijan Robinson out of Texas. This felt odd to all of us in the know, because the Falcons already had a top ten back in the NFL on their roster in Tyler Allgeier, who posted 134 total Rushing Yards Over Expected (RYOE) (6th), 0.65 RYOE/Attempt (7th), and had 44.7% of his rushes go for positive yards above expected (7th) in the 2022 season as a rookie.

You should be more outraged than you are about Tyler Allgeier’s usage.

This was actually one of the better rookie RB seasons in recent memory, as most RBs are like QBs. They tend to struggle their first season in the league, taking a full season to get up to speed before in year two we see what they really have, but the Falcons elected to throw this all away, and instead draft Bijan Robinson. Bijan did not come into the league as a finished product like Tyler did, and as a result the Falcon rush offence came plummeting back to Earth.

A greatly helpful 0.004 EPA/Play as a rush offence in 2022 became a brutally harmful -0.171 in 2023. There are other reasons for this beyond just the RB, but the choice to use a first round pick to predictably downgrade the RB position for no reason certainly did not help.

It should be a luxury to have Bijan be a downgrade, but nothing is easy in Atlanta…

Moving into 2024, it’s becoming more and more clear that the worst has happened for Atlanta. They bet on the wrong guy, and spent a first round pick in order to do it. It’s close as to who is better between Tyler and Bijan, but the statistics are clear. All it takes is a simple look at their Next Gen Stats pages. Bijan gets all of Atlanta’s carries, so let’s ignore the volume stats, and stick strictly to rate statistics.

In terms of RYOE/Attempt, it is 0.69 to 0.55 in favour of Tyler. In terms of proportion of rushing attempts that are successful, it is 49.1% (leading the NFL by the way) to 45.6% in favour of Tyler. There is a credible argument to be made that Bijan’s increased explosiveness helps a team more than the unstoppable (but never explosive) nature of Tyler Allgeier’s game, but the fact remains that if you want the most possible yards on the average rushing attempt, the best RB in Atlanta is Tyler Allgeier, not Bijan Robinson.

However, despite their knowledge of this truth, the Falcons in 2024 have consistently failed to swallow their pride and play the statistically superior Tyler Allgeier, instead choosing to give the inferior Bijan Robinson 291 total touches so far in 2024, to Tyler’s 124.

I am not at all attempting to say Bijan Robinson is a bad RB. His 0.55 RYOE/Attempt make him a pretty good RB. It’s not his fault that his top 10-15 level of play makes him the second best RB on his own team. He can’t choose where he got picked. What I am trying to say is that the Falcons are holding back their own rushing offence, and have been for the last two seasons, all because they don’t want to admit that they wasted the eighth overall pick on a RB worse than the one they already had.

This brings us back to the WARP. The Falcons are unwilling to swallow their pride to make the team better at the RB position, but apparently are willing to do so in order to improve the team at the QB position. They are both willing to admit they made a mistake in order to improve the offence, and not willing to do that.

This violates the WARP, meaning we must look elsewhere for their true intentions.

Their true intentions have nothing to do with Kirk’s performance in his first nine games. Regardless of what contract a player is on or what rookie is sitting on the bench, you couldn’t pay a team enough to part with a man who generated 0.160 EPA/Play on a 3.8 CPOE in the first nine games, which is what Kirk did in 2024.

If a team like Atlanta is so willing to part with that level of production, that means it must’ve come tumbling back to Earth in a collapse of epic proportions in the five games after that. This led me to take a closer look at Kirk Cousins’ stats over the last five, and what I found interested me greatly.

He wasn’t that bad. In fact, I wouldn’t use the word ‘bad’ at all to describe his performance.

I gave you the end results earlier of Kirk’s performance in these five games. -0.109 EPA/Play, on a -4.2 CPOE. That’s horrendous. I’m not trying to say it isn’t, but when you look just a little bit deeper, the analysis at such a surface level begins to fall apart.

The first game of this stretch was week ten. The Falcons were on the road in New Orleans. Kirk generates 0.03 EPA/Play as Atlanta scores just 17 points. This looks pretty bad, but it doesn’t tell you the whole story. Kirk operated at a 50% personal success rate in this game, and the Falcons scored just 17 points due to three missed field goals. Any of these FGs would’ve at least tied the game, but if Younghoe Koo could’ve made all of them, they could’ve won easily.

Instead, Atlanta loses 20-17, and Kirk begins his road to being benched.

The next game of this gauntlet is on the road in Denver, and what do you expect? This game is where Kirk’s haters can make their point. He’s awful. No ifs, ands, or buts, but the Denver Broncos are the best defence in the NFL. If we’re benching people for having bad starts against them, all of Anthony Richardson, Aaron Rodgers, Geno Smith, and Baker Mayfield are also out in the cold looking for jobs right now. This game is a total wash, and shouldn’t be taken as being indicative of Kirk’s overall performance level.

That brings us to the game three weeks ago against another one of the best defences in the NFL, at home against the Chargers, where Kirk plays pretty well, all things considered. Turnover luck gets him really badly in this one, which causes him to end the day with -0.17 EPA/Play to his name, but his 51% personal success rate is much more representative of how well he played, as he generated 0.269 EPA/Play on non-turnover plays, and got his Falcons across midfield all but one time he touched the ball, against a Charger offence that ought not to be taken lightly.

We’re starting to sense the general trend here. Kirk over the last five games has, in general, moved the ball quite well, but has been bedeviled by extremely untimely turnovers.

I repeat all the time the fact that turnovers are a function of the offence, not the defence. That is the fact of the matter. Throughout history, turnovers have between little and no correlation with performance on non-turnover defensive plays, meaning defensive turnovers can be thought of as luck and not much else. The reverse is not true. Less skilled offences do systematically turn the ball over more than better ones.

Many people get a Twitter-level understanding of this basic advanced metric idea, and are mostly able to understand that takeaways do not imply that a defence is good, but turnovers indeed do imply that an offence is bad. However, such a basic level of understanding does not take into account that while the correlation between turnovers and offensive performance exists, it can often be very light as well, and while turnover luck affects defences more than it impacts offences, it can mess with offences too.

This is why, in small samples, the key metric that ought to be studied to analyse a QB’s performance is not any top level results-based measure. EPA/Play is great. CPOE is great, but they take time to level out, and can be less useful in smaller samples. In a small sample, the metric that needs to be studied is success rate.

Why?

All plays are weighted the same. One pick six or one 75 yard touchdown pass can heavily skew somebody’s EPA/Play or passer rating in a sample that’s only a few games big, but when determining success rate there is no such thing as an outlier. There can only be two results. A play either generates positive EPA, or it doesn’t. A six yard gain on first down and ten (in the eyes of the success rate statistic) is just as good as a fourth down touchdown pass.

A player’s success rate has no respect for his turnover luck, and will look right through it. This is how I can tell you that Kirk Cousins played very badly in generating -0.17 EPA/Play against Denver, while simultaneously telling you that Kirk played very well in generating -0.17 EPA/Play against Los Angeles. From a top level results perspective, these two games are almost entirely equal, but not being able to move the ball at all against Denver and falling victim to turnover luck against the Chargers are two very different stories.

Remember the Josh Allen Theorem. Being able to move the ball will always outweigh the turnover risk in the long run. The Chargers story featured four INTs, each of them backbreaking, but I would take that every time over an offence that cannot move the ball.

This brings us to the game two weeks ago, once again on the road against one of the best defences in the NFL in Minnesota, and would you look at that? The Atlanta offence moves the ball really well against a defence against whom, if Kirk was as bad as people are claiming him to be, the Falcons should’ve had no chance against, but he generates just 0.02 EPA/Play, due to four poorly timed turnovers (two INTs, two on downs).

Once again, Kirk’s performance on non-turnover plays against another of the best defences in the NFL tells a radically different story. 0.419 EPA/Play on plays that aren’t turnovers is an extremely impressive performance, especially while the Atlanta rush offence is still needlessly handicapping itself, as has been happening in the background of all of this.

By both success rate and EPA/Play, last Monday’s game in Las Vegas is actually the worst of them all. This performance cannot be explained away by turnover luck, or by anything else, but Lamar Jackson played poorly against the Raiders too, and nobody benched him.

So ends the supposedly fireable five game stretch that seems to have ended Kirk Cousins’ days as the starter of the Atlanta Falcons, but wait a minute. How can that be?

There have been two real stinkers against Denver and Las Vegas, but the other three games are reasonable. A perfectly acceptable game against New Orleans that could’ve been an easy win with just a few made FGs, and terrific performances in terms of ball movement against tough opponents in the Chargers and the Vikings (especially the Vikings), are worthy of a benching?

I smell a rat here.

The statistics bear this out even more clearly, as in the first nine games, Kirk generated 0.160 EPA/Play (10th) on a 3.8 CPOE (9th), but did it with a 49.4% personal success rate (14th). In the next five, the results changed. Like I’ve said twice now, -0.109 EPA/Play on a -4.2 CPOE is quite bad. However, Kirk’s personal success rate over this five game stretch is 48.9%, which is not bad. It’s actually pretty good, 12th in the NFL over the last five weeks, and (key point) is not at all out of the norm for either this season or his career.

Despite what some will tell you, Kirk Cousins has not gotten radically worse over this last stretch of games. His results have gotten worse. In a game with as small a sample as football has, those two sentences have dramatically different meanings. It’s only been 184 plays of bad football.

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 about this.

The Falcons have not swallowed their pride in order to make the team better. Far from that. All the evidence points to this five game stretch of Kirk Cousins football being a cherry picked small sample of games that shows very bad results based primarily on poor turnover luck, and (upward) regression to the mean is coming as soon as turnover luck improves.

That is why Kirk Cousins isn’t the starter anymore.

Atlanta aren’t idiots. They can see all the same stats that we’re seeing. They know that Kirk hasn’t actually gotten any worse as a player, and they know that the improvement is coming if they just give it time to play out. That is why they are not going to do so.

The Falcons’ next three games are against (in order) the 28th, 21st, and 31st ranked defences in the NFL. I believe that, if given the chance to play against this competition level, things will turn around, and Kirk will get back to being at his regular results level again. The Falcons do not want that.

They want these last five games to be the lasting memory of Kirk Cousins as a Falcon. They want to set the narrative that this five game sample was so untenable that their hand was forced into playing Michael Penix, because when just looking at the football reference stats (one TD to nine INT over the last five games), that narrative seems plausible, but as we’ve just gone over, the top level statistics are a lie, based on poor turnover luck in a small sample of games against three of the top seven defences in the NFL, and each of the best two.

Let me be clear. The Falcons’ hands are not tied. They absolutely do have a choice. Kirk Cousins was a top 8-15 QB before this poor stretch. He was the 12th best QB by success rate during this poor stretch, and he will continue to be a top 8-15 QB from here onwards. There is nothing mandating that they have to trade this cow for the handful of magic beans that is rookie QB Michael Penix, but the Falcons have elected to do it anyway, because Penix was a first round pick, and Kirk Cousins wasn’t.

Do not misinterpret this action as anything other than what it is. In the modern NFL economy, first round pick QB trumps big free agent signing QB. Therefore, the vanity of the Atlanta brass is hurt less by admitting defeat on the big free agent signing than it would be to admit defeat on the first round draft choice. In my opinion, their actions indicate that their vanity was leading them to be scanning the horizon for an excuse to use Michael Penix, and an unrepresentative set of poor results caused by poor turnover luck was exactly the excuse they needed.

Looking at the situation this way, the Falcons are now compliant with the Weak Axiom of Revealed Preference. The Falcons are willing to harm both their rush and their pass offence for the sake of the egos in the front office not having to admit that their first round pick was a mistake, and they have now succeeded in doing it.

I hate this about the NFL. A house divided against itself cannot stand, and in the modern NFL, it feels like half the teams are wholly committed to hurting themselves in order to stroke the egos of the humans at the top, just so that they don’t have to admit to having made a mistake. I very nearly wrote this exact same article about the Anthony Richardson-Joe Flacco situation earlier this year, and it was shameful to see the Colts make abundantly clear just how badly they wanted to lose games, but Joe Flacco is no Kirk Cousins, and therefore the Colts are not as shameful as the Falcons.

I’ve been a closeted Atlanta supporter for a long time. I’ve always dug the subtle underdog energy they’ve always had, but that ends today, because they (like a lot of teams) have just made it abundantly clear that what happens off the field is more important than what happens on it. Quality of the team be damned. Penix is going to start over Cousins. That is shameful, and I cannot support it.

It’s not fair to me. It’s not fair to you. It’s not fair to Kirk Cousins, and it’s not fair to anybody in the Atlanta fanbase to treat the on-field play as such an afterthought like this, and I don’t feel the need to support a business that so clearly values the vanity of those at the top over the quality of the product. It begins to feel a lot less like football, and a lot more like politics. I have no interest in organisational politics, but the Atlanta Falcons do. It’s quite clearly the only thing they are interested in.

Stay ready Kirk. You may get the chance to go back to a real team once all this is over.

Thanks so much for reading.