Why Ravens, Seahawks Were Underachievers in 2021

Seattle Seahawks QB Russell Wilson
Seattle Seahawks QB Russell Wilson
Photo: USA Today Sports Images

NFL Offseason - Every year, we here at Football Outsiders do our best to accurately forecast the upcoming season. We run thousands of simulations taking into account major offseason personnel changes, players returning from injury, potential development from draft picks, continuity on the offensive line, and numerous other variables in an attempt to provide the best predictions out there.

Considering how difficult predicting the future can be, 2021 went quite well. There was a .74 correlation between our mean projected wins and teams' actual win totals, and a .73 correlation between our projected DVOA and teams' actual DVOA. For the most part, teams were more or less as good as we anticipated them being.

For the most part, that is, because there's always an exception or two that falls through the cracks, a team that unexpectedly gels or crumbles. You didn't exactly see too many people predicting the Cincinnati Bengals would come out of the AFC when all was said and done!

Last week, we looked at the teams that overachieved in 2021 based on our projections. Today, we'll look at those that fell short.

A quick methodological note before we begin. Our rankings are based on how many standard deviations each team beat or fell short of its projection, rather than just looking at the raw numbers. When you run thousands and thousands of simulations, it's a very rare team indeed that has an average DVOA over 20.0% or performs better than 12-5. But each NFL season only happens one time, so outlier results not only happen, but are expected—someone keeps rolling sevens all season long, it's just difficult to predict who. We're not here to explain that the best teams are better than their projections because that's how simulations work. We want to focus on the teams that got grouped wrong to begin with: the bad teams that became average or the average teams that became good.

In this installment we examine the teams that underachieved their wins projection by the most standard deviations in 2021. As we did with the overachievers, we begin in the AFC North.

1. Baltimore Ravens (1.40 Below)

Projected Wins: 10.5; +1.23 Standard Deviations
Actual Wins: 8; -0.17 Standard Deviations

Yes, the Ravens season was defined by injury, including several damaging hits that took place before the first kickoff (Marcus Peters, J.K. Dobbins, Gus Edwards). Nevertheless, Lamar Jackson managed to David Copperfield his way to victory after victory in the front part of the campaign, pulling out miraculous come-from-behind wins against the Chiefs and Colts. Jackson and Justin Tucker also somehow turned tepid ditchwater into Châteauneuf-du-Pape in Detroit, finding a way to win when Jackson put Tucker in "position" to kick the winning (and record-setting) 66-yard field goal.

Obviously, this was unsustainable, as evidenced by a pair of bludgeonings at the hands of division rivals Cincinnati. But the team's solid infrastructure and coaching seemed able to buttress even this damaged roster, and a playoff berth was assumed. Then came Week 12, when Jackson badly sprained an ankle and was lost for the season. After that, the Ravens fell into a death spiral, losing their final six games to finish 8-9 and miss the playoffs.

With the wild game of musical chairs going on in the NFL this month, it is easy to forget about the Ravens, but they should be well-placed for a rebound. The injury tsunami is unlikely to happen to such a magnitude again, and the return of dynamic playmakers such as Jackson, Dobbins, and Peters should augment studs such as tight end Mark Andrews and emerging wideout Rashod Bateman to return the offense to its accustomed spot near the top of the DVOA table (after finishing first in 2019 and 11th in 2020, Baltimore was 17th in 2021). Star safety Marcus Williams was signed to a rich deal to fill a position of need in the secondary, and a back four of Williams, Peters, Marlon Humphrey, and Chuck Clark is among the best in the league (if healthy). The Ravens will also benefit slightly from a last-place schedule.

But don't automatically assume a huge bounceback in Charm City. Jackson, despite his late-game heroics, had a mediocre season before the injury, finishing 19th among quarterbacks in DVOA. His return to form is more critical than ever given the amount of star power at the quarterback position in the AFC. Just as worrying was the trench play—or, perhaps more accurately, the edge-of-the-trenches play. Although Baltimore was first in the NFL in adjusted line yards on defense, they were 31st in adjusted sack rate as Calais Campbell aged and several developmental players failed to evolve in the manner Ravens pass-rushers generally have in the past (a fact that may have cost defensive coordinator Wink Martindale his job). Meanwhile, the offensive line was tied for 29th in adjusted sack rate allowed as All-Pro tackle Ronnie Stanley continued to struggle with injuries and retreads like Alejandro Villanueva couldn't stem the tide. (Morgan Moses was signed to play right tackle, which should be an upgrade, if a slight one.)

As mentioned above, the Ravens track record gives them the benefit of the doubt. But the team is counting on everything to simply click back into its proper 2019-2020 place. Topping eight wins might be easily attainable, and getting off this "underachievers" list doable as well. But returning to the top seed in the AFC will require a confluence of actions that will be hard to replicate.

2. Seattle Seahawks (1.38)

Projected Wins: 9.9; +0.86 Standard Deviations
Actual Wins: 7; -0.52 Standard Deviations

Well, what did Seattle expect after trading a superstar quarterback?

(… whispered aside…)

Wait, you mean the underachievement took place before the Hawks traded away the best passer in franchise history?

Just so. Seattle played under the clouds of disharmony all season, with Wilson's now-he's-happy/now-he-ain't act casting a long shadow over the team and its future. The Seahawks had won at least nine games in each of Wilson's first nine seasons under center in the Pacific Northwest, with win totals in the double-digits eight times. That extraordinary level of consistency kept their projection high, and when the team at last stumbled, it virtually assured that Seattle would end up highly placed on a list like this.

The issues were legion even beyond the palace intrigue at the team's facility in Renton, Washington. On the field, Wilson suffered a gnarly finger injury in Week 5, came back too soon, and finished with a subpar season by his lofty standards (he still was 12th in DVOA and 15th in DYAR). Meanwhile, the pass rush cratered (tied for 31st in adjusted sack rate), the overall pass defense was poor (26th in DVOA), and the long-maligned offensive line continued to struggle to protect Wilson (tied for 25th in adjusted sack rate). Unlike in previous seasons, Wilson couldn't overcome the lack of protection. Star wideout DK Metcalf fell from fifth in DYAR in 2020 to 41st in 2021, symbolic of the lack of the consistent deep passing attack the Seahawks rode to prominence under Wilson. Not much has been done as of this writing to drastically improve the problem areas.

Now Wilson is gone, traded to Denver for a huge haul of assets. The star quarterback had been Seattle's nuclear power—even as the rest of the NFC West improved by leaps and bounds, the Seahawks had to be reckoned with because No. 3 was under center. Without him, the team faces a difficult rebuilding stretch, regardless of the bounty they got in return from the Broncos. No matter the quality of the rest of the squad—and with the likes of Metcalf, Tyler Lockett, Noah Fant, Chris Carson, et al. around, they retain tons of playmakers—everything rests on the passer who succeeds Wilson. At the moment, that is Drew Lock, who inspires Absolute Zero confidence. Even if (when?) the team upgrades, will a Baker Mayfield-level player be enough to restore the team to its usual double-digit wins?

Even with the 17th game, that seems unlikely.

3. Denver Broncos (1.01)

Projected Wins: 9.3; +0.49 Standard Deviations
Actual Wins: 7; -0.52 Standard Deviations

Quick quiz: who had the better DVOA/DYAR in 2021, Teddy Bridgewater (Denver's inept passer who singlehandedly held the team back from its win projection and deserved spot atop the NFL pyramid) or Russell Wilson (Denver's new quarterback, the savior brought in to rescue the Broncos from Teddy's all-consuming mediocrity)?

You probably already guessed the answer, merely because we posed the question. Teddy was 14th/10th, Russell 15th/12th. That isn't to say that Bridgewater is better at the position than Wilson, because he demonstrably isn't. And the name of the game in the modern NFL is upgrading at quarterback. Still, the idea that Denver is a turnkey operation just waiting for a competent quarterback to return the team to its rightful place in the Super Bowl might be shining a light on the wrong dark corner.

The Broncos were chugging along at 7-6 after blowing out the Lions 38-10. They then scored 10, 13, and 13 points over the next three games, all critical losses. They managed 24 in a closing loss to Kansas City, which at least would have kept them off this list if they hadn't given up 28 and lost that one, too. A major factor was red zone production. It was a bugaboo all season for the Broncos—21st in touchdown percentage, 20th in red zone DVOA, 30th in red zone rushing DVOA (though somehow fourth in red zone passing DVOA!). They lost to the Bengals thanks to a strip-sack near the goal line late in the game, and the touchdown success rate over Denver's final three games was just 43%, bottom-third in the league.

The Broncos have a rep of a team with a great, rising defense, but they were merely 20th in DVOA, and mid-pack in adjusted sack rate and coverage of top enemy receivers. It was a unit that was greater than the sum of its parts, leading the league in fewest missed tackles and generating the ninth-best pressure rate despite trading Von Miller, but other than rookie corner Patrick Surtain there were few stars. The team swiped pass-rusher Randy Gregory out from under Dallas to help out the pressure rate, but it remains to be seen what the effect of losing Vic Fangio will have on the defense in 2022.

Denver was +13 in point differential, suggesting a 9-8 team, and that underperformance would make a bounceback likely for 2022 even without Wilson. Red zone performance isn't necessarily translatable across teams, but it is notable that Seattle with Wilson was third in touchdown percentage in 2021. Improving the run game when the offense gets down deep is critical, and with young rusher Javonte Williams leading the way that seems probable—plus, it can't get much worse.

Obviously the expectations are high for the Broncos. Merely avoiding a repeat appearance on this particular list won't be enough to satisfy fans desperate for the return to glory they seem to feel is their birthright ("We haven't been good since 2015—oh the humanity!"), but given the incredible potency of the AFC in general and the AFC West specifically, it might have to suffice.

4. New York Giants (0.70)

Projected Wins: 7.1; -0.88 Standard Deviations
Actual Wins: 4; -1.56 Standard Deviations

Hopes were not particularly high for the G-Men in 2021, and not only did they not come close to meeting them, they were so woeful it forced yet another coaching change. Joe Judge was fired after going 4-13, the third straight Giants coach to not make it to a third season at the helm. Daniel Jones was bad and injured. Saquon Barkley was the 45th-best running back in the league by DYAR. Owner John Mara was loudly booed by home fans as Eli Manning was inducted into the team's Ring of Honor. It was an epic disaster of a season, yet another in a long line for a once-proud franchise. Since 2012 the Giants have just one winning record.

2022 doesn't profile to break that skein, though staying off the underachieving list will mean at least some forward progress. That will be helped by a low initial projection, to be sure. Our way too early DVOA projections have the Giants 24th, and since those were made Denver traded for Russell Wilson, Pittsburgh added Mitch Trubisky, and the Jets and Jags considerably upgraded their roster through free agency around their second-year quarterbacks. It won't be a surprise if next season's Giants win projection is cut in half.

The lone place for optimism is the new braintrust. Mara imported Joe Schoen and Brian Daboll down from Buffalo to be his new general manager/head coach combo. Daboll is given major props league-wide for turning Josh Allen from a slab of beefcake who couldn't hit the side of the barn to an All-Pro level passer. Whether he can turn Jones into anything resembling Allen will be an interesting experiment. Wink Martindale was hired to coach the defense after a strong run in Baltimore. The defense was easily New York's strong suit in 2021, finishing 18th in DVOA as opposed to the offense's ranking of dead f-ing last. But it isn't as though the unit has oodles of talent to play with.

Big Blue's Super Bowl comes in a couple of weeks, when Schoen gets his initial crack at running a draft, one that will see the Giants have the fifth and seventh picks and six of the top 112. They could add more picks by moving players in the next few weeks, including perhaps Barkley, once the second overall pick. It has been a wildfire of disastrous personnel and coaching moves in East Rutherford over the last half-decade. For the sake of the fans of the most popular team in the largest American city, here's hoping they turn things around sooner rather than later.

5. Kansas City Chiefs (0.63)

Projected Wins: 11.5; +1.84 Standard Deviations
Actual Wins: 12; +1.21 Standard Deviations

This is a weird math thing, but stay with us. Kansas City hit the over on its projected win total, captured a division title, made the AFC Championship Game for the fourth consecutive season, and still underachieved, based on standard deviations from the mean. Trust us, the Chiefs' presence on this list doesn't factor in the 21-3 lead it blew against the Bengals in late January. But the collapse wasn't merely a case of a single bad half undoing what was otherwise a dominant season. The Chiefs had some issues in 2021, and though a good job of mitigating them righted their season and made it a success, many remain for the upcoming season as well.

Kansas City did offense and special teams just fine, finishing third overall in both disciplines. But the defense was merely 24th, and even that ranking represents a solid mid-year adjustment—after Week 8, the Chiefs defense was 31st, ahead of only the woeful Jags. Then, in a staggering turnaround, they held the next five opponents to just 9.6 points per game, turning a 4-4 record that felt kinda icky into 9-4. But that level of play wasn't sustainable, and the remainder of the year, while not as brutal as the initial eight games, felt more familiar, as the Bengals scored 34 and 27 in two encounters, the Chargers put up 28, and Buffalo went down to defeat despite a 36-point effort.

The Chiefs were also far too loose with the football, especially in the early part of the year. Overall they finished seventh from the bottom in percentage of drives ended by a turnover, and the six teams worse than them were the dregs of the league. Again, the Chiefs were so talented that they mostly got away with the sloppy play—until they didn't, losing key interceptions in the second half and overtime that allowed Cincinnati to swipe the trip to the Super Bowl out from under the Chiefs.

As of this writing, not much has changed on the roster, including the defense, one that our pre-free agency projections tabs as the 25th-best by DVOA. The Chiefs re-upped defensive end Frank Clark and brought in safety Justin Reid from Houston to presumably replace Tyrann Mathieu, who looks headed out the door, along with pass-rusher Melvin Ingram and tackle Jarran Reed. The offense made an interesting under-the-radar move in acquiring JuJu Smith-Schuster to complement Patrick Mahomes' weaponry, but its major upheaval took place last offseason, when the offensive line was rebuilt.

The Chiefs are obviously still strong, but standing largely Pat (ouch—sorry) has been complicated by the new faces Kansas City will be facing twice apiece in 2022—Russell Wilson, Davante Adams, Khalil Mack, J.C. Jackson, Chandler Jones, Randy Gregory. The AFC West has gone from simmering to supernova this offseason, while the division's standard-bearer for the last half-decade looked around and thought, "nah, we're good." Maybe the Chiefs are correct, but another season of sloppy play and shoddy defense could well mean Mahomes watches a game on Championship Sunday for the first time.

Comments

24 comments, Last at 20 Apr 2022, 2:26pm

1 When you run thousands and…

When you run thousands and thousands of simulations, it's a very rare team indeed that has an average DVOA over 20.0% or performs better than 12-5.

This implicitly assumes teams are fair die, right?

With dice, it's trivial to prove their true performance -- you just roll dice millions of times and track the actual outcomes. It converges to equally-likely outcomes (and a mean roll of 3.5) pretty quickly.

But statistics is really prone to the fallacy that increasing n decreases standard deviation. This is true in as much as increasing n decreases measurement errors -- the mean of that error should converge to zero as n approaches infinity, assuming the error is random and not biased. But baked into the concept of mean and standard deviation is that a population has a true mean and a true standard deviation, meaning after some n << infinity, the sample should converge to the true mean and standard deviation of the population.

Put it this way -- in any given population, the average adult guy is going to have a peak stature of around 69 inches tall. The population mean will converge to something close to this value. About 5% will converge to something like 74 inches. But some people are naturally 77 inches tall. If you re-ran their life 10,000 times, they would still converge to 77 inches with some small standard deviation. You can't expect a single member of the population to converge to the population mean as the same increasing n trend that the population exhibits. 

So in a roundabout way, this method seems to artificially depress the magnitude of true team performance. Some teams are actually + or -20% teams, and they should converge to something like this value no matter how many simulations are run. So how do you identify true individual performance versus true population tendency, when those are not the same things?

3 Variance

When you run thousands and thousands of simulations, it's a very rare team indeed that has an average DVOA over 20.0% or performs better than 12-5. But each NFL season only happens one time, so outlier results not only happen, but are expected—someone keeps rolling sevens all season long, it's just difficult to predict who

FO is implicitly saying that each team has a true strength, but that the small sample sizes will naturally lead to higher variation in one season than one would see if we averaged over thousands of seasons.  That's reasonable.

You can't expect a single member of the population to converge to the population mean as the same increasing n trend that the population exhibits. 

There's not saying that.  There saying that, with 32 different values being estimated, and each only being estimated once, there are bound to be some poor estimates.  That's why the real-world outliers are further from the population means than the true values would be.

If I'm estimating 32 different random variables (ignoring the dependencies), but I use a sample size of 16 (erm, 17) for such an estimate, my variance for each estimate will be much higher than if I use a sample size of 16,000. (erm, 17,000)

That's why averaging the max win values over the simulated seasons will produce a higher value than taking the max of the average win values.  Which is another way of saying that this simulation approach will consistently group the teams closer to the league mean win total (8.5) than actual seasons will.   

6 I think the natural player…

I think the natural player lifecyle. (Draft->FA + Cap) is what does it. If a team is bad, they get high draft picks and get better, good players come of rookie contracts and force a team to shed talent due to the cap.  That, and the the number of simulations is why you don't often get over 20% DVOA or 12-5, bc those are averages of a large number of runs to remove the luck aspect.  And the cap makes it hard for a team to be much more talented than the others, which is the only way to shift he average significantly from the median. Same for less talented to a degree, if your in the top 5, both pretty much any player will improve your team and the hit rate is pretty high too. 

Basically, the natural flow of the league makes it difficult for teams to have  a  large talent differential and have results that are way at the tail end either way. That is why there is only 1 perfect season and very few 0-XX ones. 

 

 

15   But statistics is really…

 

But statistics is really prone to the fallacy that increasing n decreases standard deviation.

This isn't quite the right question to ask in response to that. The reason why increasing n doesn't lead to decreases in standard deviation is because some systematic issue pops up. Imagine if you're trying to measure the rate of some (purely) random thing. You count for 100 of your time periods, then you count for 1000 of your time periods, and the standard deviation decreases by sqrt(n). Then you try counting for 1E7 of your time periods, and... you find that it doesn't decrease by nearly that amount. Because your time periods weren't perfect - your clock wanders slightly, and that wandering becomes the dominant variation over large periods.

But in this case, the "n" they're talking about are just "random throws of different variables." Increasing n just increases the precision of your final estimate. Really, you wouldn't've even needed to do a simulation, you could solve it analytically, it's just annoying and easier to Monte Carlo it. The "number of simulations" is just beating down the sampling error, and that'll always decrease.

The right question to ask though, would be about the number of different random inputs (call that "big N"), because that's what's tightening the distribution regardless of the sampling error. One of your variables (implicitly) is "can the Chiefs replace the skill players they've lost," and another one might (implicitly) be "how much does Tyreek Hill add to the Dolphins." In a sense the "simulation" treats those questions (again, implicitly) as independent... but again, in some sense, they aren't actually independent, too. The chance of the Chiefs replacing the skill players is related to how much Tyreek Hill adds to the Dolphins.

You might say "wait, but that's the same thing I'm saying," except it's not, really. It's not related to the number of simulations (which is just a way of solving a big integral which you could do by hand anyway) - it's related to the fact that you had too many degrees of freedom to begin with. (And that's a much deeper question)

But really, the main reason that the distributions tighten with "lots of simulations" is just because the central value of most of these projections is "typical" (since fundamentally the central value of a convolution of a bunch of Gaussians is just their central values added in quadrature). We don't know if Trevor Lawrence will be epic, so we 'estimate' him as a typical second-year first-overall QB. Some of those QBs are terrible, some of those QBs are awesome, and the average is 'typical'. We don't know if New Coach will be good, so we 'estimate' him as "typical first year head coach," and "typical first year head coach" is bad.

16 But statistics is really…

But statistics is really prone to the fallacy that increasing n decreases standard deviation.

I was contemplating something a little different.

Reviewers are prone to significance bias -- more papers report significance than non-significance. So reviewers are trained to expect significance. This is reflected in reviews which ask retrospectively about sample power -- the unstated begged question is that the study is actually significant, but undersampled. This reflects the assumption that an increase in n results in a decrease in standard deviation. This would occur if a study were actually significant but was underpowered, but would not result if the true result was actually non-significant. 

But because that's where the selection pressure lies, I tend to see an unspoken assumption that variance decreases toward zero as n increases. This is only true to a point; sampled variance is actually approaching its true value.

17 This is only true to a point…

This is only true to a point; sampled variance is actually approaching its true value.

This isn't what's going on here, though. What he's saying is that the distribution of team wins in the simulation is tighter than the win distribution in the NFL, and that's because they're trying thousands of different "guesses" for the way the NFL season will work - but only one of those will be the "actual' NFL season, and so it'll have higher variation than the simulation average does.

It's just reflecting lack of knowledge. You don't know how Trevor Lawrence will behave, so you give him the "typical" expectation. Yes, you run thousands of different trials on that value, but that's just math for solving an integral when you quote the average value. The "typical" expectation (which is what results in the simulation average) actually has zero variation. It's just a number.

When the "actual" Trevor Lawrence plays, he obviously doesn't play like the "typical" expectation, he plays like himself. And so the difference between him and the "typical" expectation is "unaccounted for variance," which spreads the final results.

Again, the "thousands of simulations" here is not actually an "increase in n." It's just a method for solving an integral. He could've said "when you don't know how these season-to-season decisions will work out and you just have to model their chances, it's a very rare team that ends up with an average [...] but in the NFL, the decisions either work or they don't."

It's like if a team went for 2 every time rather than kicking the extra point. You could calculate the average return after N tries, but the result after 1 try will always have more variance unless you absolutely know that the result will succeed or fail.

(edit: I should point out of course that this also ignores the fact that modeling the NFL season obviously will ignore the actual unpredictable portion of the in-season per-game results, which is about as big as the predictable part! So even if you perfectly knew how all the off-season stuff would work out, you'd still have to just randomly fuzz the results to get something that 'looks like' an NFL season).

2 I know your Baltimore…

I know your Baltimore projection needed to be something longer than "It was the injuries," but your entire analysis is just a mock-draft-like restatement of "It was the injuries."

Defensive line got great push but no sacks? See the (1-decimation) of the entire starting secondary.

QB dropped in efficiency? See the elimination of the entire backfield group and injuries throughout the receiver group.

Line was a disaster? See above.

It wasn't just that the Ravens had a ton of injuries (they did), it was also that the injuries hit entire groups at the same time, so entire position rooms were eliminated for swaths of games. The Ravens played a game with basically no members of their two-deep in either backfield. You couldn't win a game in the 1930s like that.

And the name of the game in the modern NFL is upgrading at quarterback. Still, the idea that Denver is a turnkey operation just waiting for a competent quarterback to return the team to its rightful place in the Super Bowl might be shining a light on the wrong dark corner.

Maybe.

If you borrowed this analysis, you would argue LAR wildly overpaid for Stafford, because Stafford was only around 200 DYAR better than Goff in 2019 and 2020 and Goff ran the overall better offense. But Goff was underperforming in a great car, whereas Stafford was eking out every ounce from his 1982 Datsun, which was towing a rusty trailer full of burning diapers. Once they switched teams and rosters, Stafford was 850 DYAR and 27% better in offensive DVOA than Goff.

Which is a roundabout way of saying that it's entirely reasonable that Wilson had dragged Seattle's offense as far as it could be dragged, whereas Bridgewater may have been carried by an otherwise-good team. I'm not saying that's the case, but we've done nothing to eliminate the possibility whose assumption underlined Denver's trade.

Re: Chiefs --

Is the second number the stdev of DVOA, or the stdev of wins?

5   Which is a roundabout way…

 

Which is a roundabout way of saying that it's entirely reasonable that Wilson had dragged Seattle's offense as far as it could be dragged, whereas Bridgewater may have been carried by an otherwise-good team. I'm not saying that's the case, but we've done nothing to eliminate the possibility whose assumption underlined Denver's trade.

Even the author admits that Bridgewater wasn't as good as his DYAR/DVOA would suggest, and that Wilson constitutes an upgrade at the position, which seems to supply support to the notion that Denver's offense carried Bridgewater. But instead of saying: "if Teddy Bridgewater could perform this well with that supporting cast, just imagine what a quarterback as good as Wilson might do," he goes the completely other way.

7 The exact phrase was …

The exact phrase was "waiting for a competent quarterback to return the team to its rightful place in the Super Bowl." The problem with projecting Denver as instant Super Bowl contenders now that they've upgraded from Bridgewater to Wilson is two-fold:

1) No matter his true talent level, they got a good (not great) year out of Bridgewater last year. Maybe it was a roll of 6 on the die, but that doesn't matter when it's already happened. Wilson is a smart bet to be better than Bridgewater this year, and quite possibly will improve over Bridgewater's performance last year as well. But the difference between Bridgewater2021 and Wilson2022 isn't likely to be as large as e.g. Goff2020 to Stafford2021, simply because the baseline is higher.

2) They could get better performance from the QB and still not improve that dramatically in terms of wins/losses. This is especially true given the AFC West arms race that has happened this year. The Broncos with Wilson are still probably the third best team in their own division, with fourth being closer than first.

Once again - they are a better team with Wilson than Bridgewater, and I think they were smart to make the trade. But I think it's far too glib to simply look at the Rams and say "see? Broncos will do the same thing."

8 But the difference between…

But the difference between Bridgewater2021 and Wilson2022 isn't likely to be as large as e.g. Goff2020 to Stafford2021, simply because the baseline is higher.

It's about the same baseline.

Bridgewater+Wilson(2021) was 1133 DYAR. Stafford+Goff(2020) was 1069 DYAR.

9 I'm confused what you're…

I'm confused what you're comparing, here. The baseline is the quarterback who is being replaced. 2020 Goff was 385 DYAR and -1.1% DVOA (20th and 22nd respectively), while 2021 Bridgewater was 601 / 9.7% (14/10). Let's say Wilson comes in and puts up the exact same numbers as Stafford did this year - the Rams still improved more than the Broncos, since they got less production before the trade.

19 I just don't understand the…

I just don't understand the confidence of putting Wilson at that level. He's had two top-5 DYAR passing season in Seattle. One. And he's 33, so the rushing ability is definitely less guaranteed going forward. Just looking at passing performance he's closer to (but still above!!) Andy Dalton than Aaron Rodgers.

To me this is a fascinating trade - if Wilson goes to Denver and puts up year-after-year of 1000+ DYAR performances, it's a massive indictment of Pete Carroll as a coach. It's easy to say "Carroll was holding him back" but typically when a player gets more touches, their per-touch performance goes down. Ben Roethlisberger, for instance, was absolutely being held back initially by Cowher, but when his role expanded, his per-touch performance went down. His highest DVOA ever was in year 2.

And Wilson's per-touch performance just... wasn't that high most seasons. It doesn't look the same as Roethlisberger, for instance. 

I'm not saying Wilson's not a great QB and not better than Bridgewater. I just don't see the whole "Hall of Fame" level thing (and yes, I know where Wilson is on PFR's HOFm - it's AV based, which has big issues integrating rushing numbers for QBs due to the historical importance of running, it's flat out stated in the AV methodology that it's based on average run/pass ratio from 1970 to now, and no, I do not think Lamar Jackson's 2019 was the greatest QB season ever).

18 Agreed. As a Ravens fan, it…

Agreed. As a Ravens fan, it was a very tough season. They were reduced to STARTING players at multiple positions that were not only unfamiliar to each other, but were unfamiliar with the Ravens scheme. Tyler Huntley (with RBs of Latavius Murray and Devonte Freeman!) ALMOST got them the W against Green Bay while playing with a very depleted secondary (playing the likes of Keyvon Seymour and Robert Jackson who did their best, but are not going to make people forget Marlon Humphrey, Marcus Peters). Heck, Josh Johnson started at QB! Simple reasonable return to health and some better (could it be any worse than last year) injury luck will result in a (near) totally different team this coming season. That very different team, with a very motivated #8, and playing a last place schedule is likely to earn them enough wins to find a place in the post season.  

10 fans desperate for the…

fans desperate for the return to glory they seem to feel is their birthright ("We haven't been good since 2015—oh the humanity!")

Damn Rob, I get having contempt for owners or players, but having contempt for entire fanbases is kind of off-putting. Between this and Tanier constantly joking about how little he cares or even knows about the Broncos, I'm wondering if there are any Bronco fans who will continue to read this place.

13 Drew Lock is not the presumed starter

Geno Smith is. Pete Carroll said so, though the depth chart on ESPN hasn't been upgraded to reflect this.

And why wouldn't Geno be the presumed starter? His stat line is not all that bad for his short stint. Far better than one would expect to get out of Lock.