Ryan Tannehill

11318 replies

CBcbradForum Veteran
Nov 28, 2019, 12:11 PM

"Fin D wrote:

Right because you’re deciding what Top 10 means.

You can make Top 10 mean numerous different things from Top 10 all time, Top 10 in a certain era, Top 10 versus every other QBs best same game stretch, Top 10 versus every other QBs best same game stretch that year (your particular take) or Top 10 during the stretch, which is my and most other people’s take when we say it.

That’s the issue with being entirely and singularly stat focused.

Top 10 in passer rating in a given year or during some stretch of play, which was precisely the stat people were referencing when saying he was playing at a "top 10 level". Tannehill never did that in Miami when the proper comparisons are made. In Tennessee, we'll see. Right now his stats are still consistent with a short streak of great play, but like I said if he continues this to year's end it will be statistically significant. I mean.. his current z-score is +1.7106 so he has a tremendous head start.

Fin DFin DForum Veteran
Nov 28, 2019, 12:41 PM

"cbrad wrote:

Top 10 in passer rating in a given year or during some stretch of play, which was precisely the stat people were referencing when saying he was playing at a "top 10 level". Tannehill never did that in Miami when the proper comparisons are made. In Tennessee, we'll see. Right now his stats are still consistent with a short streak of great play, but like I said if he continues this to year's end it will be statistically significant. I mean.. his current z-score is +1.7106 so he has a tremendous head start.

You're finding the the range/variables/etc. to use that best fits your argument. There is no need to normalize the results versus QBs in different sections of a given season, unless you have a narrative to push. All that does, is effectively strip out the counter argument by erasing its importance too the stats. You are and have been ignoring known handicaps that Thill had to deal with beyond his control and outside of his abilities, that other QBs did NOT have to deal with, then wanting to normalize his numbers versus that. And as always, your approach is predicated on the notion that everything comes out in the wash and everyone has the same issues to deal with.

When i say plays at Top 10 level, I am referring to that given stretch or period when had the things I said he needed versus the QBs in that same time frame.

I ask you again, what are the ways/tools a QB can use deal with a poor oline?

CBcbradForum Veteran
Nov 28, 2019, 01:03 PM

"Fin D wrote:

You're finding the the range/variables/etc. to use that best fits your argument. There is no need to normalize the results versus QBs in different sections of a given season, unless you have a narrative to push. All that does, is effectively strip out the counter argument by erasing its importance too the stats. You are and have been ignoring known handicaps that Thill had to deal with beyond his control and outside of his abilities, that other QBs did NOT have to deal with, then wanting to normalize his numbers versus that. And as always, your approach is predicated on the notion that everything comes out in the wash and everyone has the same issues to deal with.

When i say plays at Top 10 level, I am referring to that given stretch or period when had the things I said he needed versus the QBs in that same time frame.

No Fin-D it's the other way around. It's ONLY if you have an agenda to push that you would compare a "best consecutive 8 game stretch" to a full year's passer ratings, which is precisely what everyone else was doing. THAT is deliberately trying to mislead people with stats.

The comparisons I'm making are the comparisons you'd want to make if you start with the data and infer a hypothesis from the data instead of the other way around. That is, you want to compare "like vs. like" by equating sample size, or if you can't equate them use a statistical test that takes into account the different sample sizes (which lead to different levels of uncertainty).

Also, the reasons it's fine to compare "best N game stretches" during different portions of a season are: 1) random variation means that the "best N game stretch" could occur at any time, and 2) average passer rating by game number is nearly constant throughout the season (very minor differences). So the comparisons I'm making are what you want to make IF you want the data to speak for itself.

Finally, the claim Tannehill had to deal with issues so much worse than other QB's had to deal with is pure speculation. You're using that kind of assumption (and that's all it is, an assumption) to try to reject proper statistical analysis.

Fin DFin DForum Veteran
Nov 28, 2019, 01:46 PM

"cbrad wrote:

No Fin-D it's the other way around. It's ONLY if you have an agenda to push that you would compare a "best consecutive 8 game stretch" to a full year's passer ratings, which is precisely what everyone else was doing. THAT is deliberately trying to mislead people with stats.

The comparisons I'm making are the comparisons you'd want to make if you start with the data and infer a hypothesis from the data instead of the other way around. That is, you want to compare "like vs. like" by equating sample size, or if you can't equate them use a statistical test that takes into account the different sample sizes (which lead to different levels of uncertainty).

Also, the reasons it's fine to compare "best N game stretches" during different portions of a season are: 1) random variation means that the "best N game stretch" could occur at any time, and 2) average passer rating by game number is nearly constant throughout the season (very minor differences). So the comparisons I'm making are what you want to make IF you want the data to speak for itself.

Finally, the claim Tannehill had to deal with issues so much worse than other QB's had to deal with is pure speculation. You're using that kind of assumption (and that's all it is, an assumption) to try to reject proper statistical analysis.

Yeah no.

Saying he's playing at a top 10 level based on the other QBs that week is perfectly acceptable and common. ONLY YOU CBRAD, decides to construct some BS overly complicated measure so you never have to admit you've been wrong.

It is not speculation. We know for a fact, that he wasn't;t allowed to break the pocket or call audibles for huge portions of his Miami career. We know for a fact, Lazor called long developing pass plays with a crap oline. We know for a fact Sherman had Thill yell go or go go depending on pass or run. We know for a fact, most of his career in Miami the OCs called very little running plays relative tot he rest of the league. We also know the QBs you want to cherry pick time frames of DID NOT have to deal with those issues. This all happened and for you to keep denying these things and then pushing your own speculation that his situation was no different than anyone else's is absurd...no matter how many times you find a way to try (and fail) and justify it.

And yes, i know you'll make another post full of reaches then tell me you're not discussing it with me anymore all while completely ignoring that you once again, engaged me.

CBcbradForum Veteran
Nov 28, 2019, 02:04 PM

"Fin D wrote:

Yeah no.

Saying he's playing at a top 10 level based on the other QBs that week is perfectly acceptable and common. ONLY YOU CBRAD, decides to construct some BS overly complicated measure so you never have to admit you've been wrong.

It's not "perfectly acceptable" even if it's common. You won't get anything published in a scientific journal if you present a statistic without a measure of its uncertainty, and that's precisely what's missing when people deliberately ignore sample size. And even if you think what I'm doing is BS, it's not. What I'm doing is called proper statistical methodology, which you neither understand nor respect.

"Fin D wrote:

It is not speculation. We know for a fact, that he wasn't;t allowed to break the pocket or call audibles for huge portions of his Miami career. We know for a fact, Lazor called long developing pass plays with a crap oline. We know for a fact Sherman had Thill yell go or go go depending on pass or run. We know for a fact, most of his career in Miami the OCs called very little running plays relative tot he rest of the league. We also know the QBs you want to cherry pick time frames of DID NOT have to deal with those issues. This all happened and for you to keep denying these things and then pushing your own speculation that his situation was no different than anyone else's is absurd...no matter how many times you find a way to try (and fail) and justify it.

None of that implies Tannehill had it worse than any other QB. I'm about to go and have Thanksgiving dinner with a bunch of Redskins, Vikings and Ravens fans. I can guarantee you the degree to which those Redskins fans think their OL is the worst, their team has way too many holes for any QB to succeed, exceed by FAR anything you hear from Dolphins fans.

And Vikings fans are just expecting the worst come playoff time. It's like they expect a miracle play to occur that causes them to lose.

So personally, I think all these opinions you and others have about how bad Tannehill had it relative to the rest of the league are because you don't pay anywhere near as much attention to the rest of the league. So no the facts you list (most of which I'm not denying btw) aren't at all evidence Tannehill had it worse than anyone else.

"Fin D wrote:

And yes, i know you'll make another post full of reaches then tell me you're not discussing it with me anymore all while completely ignoring that you once again, engaged me.

You're wrong about the first part: I haven't made any post with "reaches" about Tannehill. What I'm posting is data driven analysis, not hypothesis driven analysis. But you're right about the second part. You're back to acting like proper statistical methodology is crap even when you don't have any clue about it. So yes it's time to end this discussion until the next time, and I'll let you end it. And no I don't mind engaging you until you start with that crap about proper statistical methodology being crap.

Fin-OFin-OForum Veteran
Nov 28, 2019, 02:30 PM

LOL.

Sooooo many memories.

Why not just put your pitchforks down and compromise?

When Ryan has had protection and a running game? He is good. Like good slightly beyond game manager.

When Ryan's passing is the focal point of an offense? He has been bad. Not horrendous, just not good enough.

Really is that simple, happy Turkey Day team!!

"The reason so many people misunderstand so many issues is not that these issues are so complex, but that the people do not want a factual or analytical explanation that leaves them emotionally unsatisfied"

Fin-OFin-OForum Veteran
Nov 28, 2019, 02:33 PM

"cbrad wrote:

It's not "perfectly acceptable" even if it's common. You won't get anything published in a scientific journal if you present a statistic without a measure of its uncertainty, and that's precisely what's missing when people deliberately ignore sample size. And even if you think what I'm doing is BS, it's not. What I'm doing is called proper statistical methodology, which you neither understand nor respect.

None of that implies Tannehill had it worse than any other QB. I'm about to go and have Thanksgiving dinner with a bunch of Redskins, Vikings and Ravens fans. I can guarantee you the degree to which those Redskins fans think their OL is the worst, their team has way too many holes for any QB to succeed, exceed by FAR anything you hear from Dolphins fans.

And Vikings fans are just expecting the worst come playoff time. It's like they expect a miracle play to occur that causes them to lose.

So personally, I think all these opinions you and others have about how bad Tannehill had it relative to the rest of the league are because you don't pay anywhere near as much attention to the rest of the league. So no the facts you list (most of which I'm not denying btw) aren't at all evidence Tannehill had it worse than anyone else.

You're wrong about the first part: I haven't made any post with "reaches" about Tannehill. What I'm posting is data driven analysis, not hypothesis driven analysis. But you're right about the second part. You're back to acting like proper statistical methodology is crap even when you don't have any clue about it. So yes it's time to end this discussion until the next time, and I'll let you end it. And no I don't mind engaging you until you start with that crap about proper statistical methodology being crap.

:jt0323::jt0323:

"The reason so many people misunderstand so many issues is not that these issues are so complex, but that the people do not want a factual or analytical explanation that leaves them emotionally unsatisfied"

CBcbradForum Veteran
Nov 28, 2019, 03:44 PM

"Fin-O wrote:

:jt0323::jt0323:

Just an update: the Redskins fans here are saying there’s a good chance they will NOT take a QB with their first pick.. owner apparently likes the QB and they knew going in that he was raw.

Good news for those of us wanting a QB.

It’s actually hilarious here. One Redskins fan is incessantly trying to convince a 1 year old that he HAS to become a Redskins fan lol.

PAPaulyForum Veteran
Nov 28, 2019, 03:57 PM

"Fin D wrote:

Yeah no.

Saying he's playing at a top 10 level based on the other QBs that week is perfectly acceptable and common. ONLY YOU CBRAD, decides to construct some BS overly complicated measure so you never have to admit you've been wrong.

It is not speculation. We know for a fact, that he wasn't;t allowed to break the pocket or call audibles for huge portions of his Miami career. We know for a fact, Lazor called long developing pass plays with a crap oline. We know for a fact Sherman had Thill yell go or go go depending on pass or run. We know for a fact, most of his career in Miami the OCs called very little running plays relative tot he rest of the league. We also know the QBs you want to cherry pick time frames of DID NOT have to deal with those issues. This all happened and for you to keep denying these things and then pushing your own speculation that his situation was no different than anyone else's is absurd...no matter how many times you find a way to try (and fail) and justify it.

And yes, i know you'll make another post full of reaches then tell me you're not discussing it with me anymore all while completely ignoring that you once again, engaged me.

I have to agree with cbrad. Whatever test you want to apply:- season rating, career rating, best N games, games against “top” defenses, rating when under pressure - it has to applied equally to all QBs. Then it is only meaningful if a sufficient number of QBs generate enough data for it to be significant.

It is fair to say that Tannehill has the potential to be a top 10 QB and has played for large stretches at top 10 level. It is not fair to say he has been a top 10 QB.
It is fair to say he has been held back by poor coaching/FO, and that with optimum coaching/FO he would do better. It is also fair to say that at least 26 QBs can make the same argument (Tom Brady, Lamar Jackson, Pat Mahomes and Drew Brees are the 4 I would definitely say you can’t make that argument for). For example Seahawks fans have complained for years about Russel Wilson being held back by the FO neglecting the OL and WR positions. What is not fair to say is that if all QBs got optimum coaching/FO that Tannehill would definitely be top 10.

Fin DFin DForum Veteran
Nov 28, 2019, 06:47 PM

"Pauly wrote:

I have to agree with cbrad. Whatever test you want to apply:- season rating, career rating, best N games, games against “top” defenses, rating when under pressure - it has to applied equally to all QBs. Then it is only meaningful if a sufficient number of QBs generate enough data for it to be significant.

It is fair to say that Tannehill has the potential to be a top 10 QB and has played for large stretches at top 10 level. It is not fair to say he has been a top 10 QB.
It is fair to say he has been held back by poor coaching/FO, and that with optimum coaching/FO he would do better. It is also fair to say that at least 26 QBs can make the same argument (Tom Brady, Lamar Jackson, Pat Mahomes and Drew Brees are the 4 I would definitely say you can’t make that argument for). For example Seahawks fans have complained for years about Russel Wilson being held back by the FO neglecting the OL and WR positions. What is not fair to say is that if all QBs got optimum coaching/FO that Tannehill would definitely be top 10.

If Thill puts up Top 10 numbers in given week, then he is being compared to all other QBs that week.

If during one game, Thill put up 120ish rating with only two other QBs putting ratings that high, according to you and Cbrad I cannot say Thill played like a Top 3 QB that week, because first you'd have to take the BEST week of all the other QBs throughout the year first. That would be absurd and unnecessary.

And again, if you agree with what you wrote that I bolded, then you DO NOT AGREE with Cbrad, but agree with me. Here's what I said that he started in on:

"Fin D wrote:

Every time he gets those things from teammates and coaches, (even the few times here) he's played at a Top 10 level.

There is nothing wrong with what i said and nothing that required him to redefine what Top 10 means.

The GuyThe GuyForum Veteran
Nov 28, 2019, 08:59 PM

The problem with the discussion above is that we aren’t reliably measuring the degree to which quarterbacks across the league are helped or harmed by their surroundings. Consequently there can be wild speculation about any quarterback in that regard.

resnorresnorForum Veteran
Nov 28, 2019, 09:09 PM

But your solution is to ignore it and act like every QB has it the same.

Fin DFin DForum Veteran
Nov 28, 2019, 09:19 PM

"The Guy wrote:

The problem with the discussion above is that we aren’t reliably measuring the degree to which quarterbacks across the league are helped or harmed by their surroundings. Consequently there can be wild speculation about any quarterback in that regard.

Name a Top 10 QB that has a shitty oline, not allowed to audible, team is in the bottom 5 in run plays called and was told not to break the pocket.

Fin-OFin-OForum Veteran
Nov 28, 2019, 10:23 PM

"Fin D wrote:

Name a Top 10 QB that has a ****ty oline, not allowed to audible, team is in the bottom 5 in run plays called and was told not to break the pocket.

That honestly was only for the less than 2 seasons we had Bill Lazor, he wasn't allowed to change the formation but WAS allowed to change the play from what we've heard.

The told not to break the pocket is likely hyperbolic as well, no coach would say "no matter what, don't leave the pocket"....even a goof like Philbin.

The other two points however are hard to argue against.

We should all be happy for Ryan, he is a fantastic situation and I've enjoyed watching him thrive.

"The reason so many people misunderstand so many issues is not that these issues are so complex, but that the people do not want a factual or analytical explanation that leaves them emotionally unsatisfied"

The GuyThe GuyForum Veteran
Nov 28, 2019, 10:38 PM

"resnor wrote:

But your solution is to ignore it and act like every QB has it the same.

No, my response to that is to point out that the better quarterbacks in the league vary in their performance from year to year at a level significantly higher than the worse ones, which suggests that variation in quarterbacks' surroundings is a less powerful variable than quarterbacks' individual ability.

The GuyThe GuyForum Veteran
Nov 28, 2019, 10:40 PM

"Fin D wrote:

Name a Top 10 QB that has a ****ty oline, not allowed to audible, team is in the bottom 5 in run plays called and was told not to break the pocket.

The task should be to determine the degree to which the average QB was helped or harmed by his surroundings and to determine the degree to which Tannehill deviated from that average, not to determine whether there was a QB who experienced surroundings identical to Tannehill's.

Fin DFin DForum Veteran
Nov 28, 2019, 10:47 PM

"The Guy wrote:

The task should be to determine the degree to which the average QB was helped or harmed by his surroundings and to determine the degree to which Tannehill deviated from that average, not to determine whether there was a QB who experienced surroundings identical to Tannehill's.

I keep asking people and no one ever actually answers.......

What are the ways/tools a QB can use to counter the pass rush?

But to your point....its not 0. If it's not zero, then it is some percentage. If it was as low as 5%, what would Thill's stats look like with players and coaches better than Brian Hartline, Legeduu Nanee & Mike Sherman?

CBcbradForum Veteran
Nov 29, 2019, 01:03 AM

"The Guy wrote:

No, my response to that is to point out that the better quarterbacks in the league vary in their performance from year to year at a level significantly higher than the worse ones, which suggests that variation in quarterbacks' surroundings is a less powerful variable than quarterbacks' individual ability.

That's only true if all surrounding variables truly "average out" over the time period you're looking at.

Remember when I looked at the change in win% over 5 year periods (for each team) and found that the standard deviation in win% was about half of what you'd expect if team strength was random from year to year? That affects calculations of statistical significance.

For example, a t-test shows that the probability Wilson's and Tannehill's game-by-game adjusted ratings "come from the same QB" is a ridiculously tiny 0.095%, clearly less than the 5% generally used to claim statistical significance. However, if we assume that half of the difference in their adjusted ratings is due to their surroundings, we'd have to remove ~6.3 passer rating points from Wilson's ratings and the t-test now gives you a probability of 9.5% that the two sets of adjusted ratings come from the same QB, which is no longer statistically significant (and btw it's total coincidence that the stat went from 0.095% to 9.5% and was 100x larger: it's actually 0.0951 and 9.53).

Not saying that represents reality. Just pointing out that once you assume "partial" averaging out we're back to where we started: we can't use the stats available to determine what percentage of QB performance is due to the surroundings. Or at least I don't yet see how to do that.

The GuyThe GuyForum Veteran
Nov 29, 2019, 09:14 AM

"cbrad wrote:

That's only true if all surrounding variables truly "average out" over the time period you're looking at.

Remember when I looked at the change in win% over 5 year periods (for each team) and found that the standard deviation in win% was about half of what you'd expect if team strength was random from year to year? That affects calculations of statistical significance.

For example, a t-test shows that the probability Wilson's and Tannehill's game-by-game adjusted ratings "come from the same QB" is a ridiculously tiny 0.095%, clearly less than the 5% generally used to claim statistical significance. However, if we assume that half of the difference in their adjusted ratings is due to their surroundings, we'd have to remove ~6.3 passer rating points from Wilson's ratings and the t-test now gives you a probability of 9.5% that the two sets of adjusted ratings come from the same QB, which is no longer statistically significant (and btw it's total coincidence that the stat went from 0.095% to 9.5% and was 100x larger: it's actually 0.0951 and 9.53).

Not saying that represents reality. Just pointing out that once you assume "partial" averaging out we're back to where we started: we can't use the stats available to determine what percentage of QB performance is due to the surroundings. Or at least I don't yet see how to do that.

You don't think the problem you outlined above is addressed by comparing multiple QBs over many years?

For example, compare the passer ratings of the group comprised of Peyton Manning, Tom Brady, Drew Brees, Aaron Rodgers, and Russell Wilson, to the group comprised of Ryan Tannehill, Andy Dalton, Cam Newton, Joe Flacco, and Eli Manning, over the seven seasons from 2012 to 2018.

If the passer ratings of the former group vary at a significantly higher level than those of the latter group, despite the year-to-year variation in passer ratings for all of them, what should the conclusion be?

Surely we can't reasonably conclude that the former group's passer ratings are significantly higher because they enjoyed significantly better surroundings than the latter group over the entirety of that seven-year period.

AGAGuyNamedAlexForum Veteran
Nov 29, 2019, 09:27 AM

"The Guy wrote:

You don't think the problem you outlined above is addressed by comparing multiple QBs over many years?

For example, compare the passer ratings of the group comprised of Peyton Manning, Tom Brady, Drew Brees, Aaron Rodgers, and Russell Wilson, to the group comprised of Ryan Tannehill, Andy Dalton, Cam Newton, Joe Flacco, and Eli Manning, over the seven seasons from 2012 to 2018.

If the passer ratings of the former group vary at a significantly higher level than those of the latter group, despite the year-to-year variation in passer ratings for all of them, what should the conclusion be?

Surely we can't reasonably conclude that the former group's passer ratings are significantly higher because they enjoyed significantly better surroundings than the latter group over the entirety of that seven-year period.

I dont think anyone is arguing that a guy can be elite just due to surroundings.

Just that QB's of a similar talent level with perform better/worse depending how their teams are tailored to thei talent.

Vertical LimitVertical LimitForum Veteran
Nov 29, 2019, 09:47 AM

Another guy we dont talk about much that is suddenly having success is Jordan Phillips.. 7.5 sacks at his position puts him at 2nd in the league.. he turned his career around..

The GuyThe GuyForum Veteran
Nov 29, 2019, 10:15 AM

"AGuyNamedAlex wrote:

I dont think anyone is arguing that a guy can be elite just due to surroundings.

Just that QB's of a similar talent level with perform better/worse depending how their teams are tailored to their talent.

Sure, and that's true for all QBs, and then the questions become: 1) how likely are the surroundings to be optimized in that way for any QB, and 2) to what degree will that optimization enhance a QB's performance?

The GuyThe GuyForum Veteran
Nov 29, 2019, 10:17 AM

"Vertical Limit wrote:

Another guy we dont talk about much that is suddenly having success is Jordan Phillips.. 7.5 sacks at his position puts him at 2nd in the league.. he turned his career around..

Buffalo is doing it largely with smoke and mirrors, having for example the weakest strength of schedule in the league this year, but they're also surrendering an opposing passer rating of 78.5, which is fourth in the league and to which Phillips is likely contributing to significantly.

CBcbradForum Veteran
Nov 29, 2019, 11:01 AM

"The Guy wrote:

You don't think the problem you outlined above is addressed by comparing multiple QBs over many years?

For example, compare the passer ratings of the group comprised of Peyton Manning, Tom Brady, Drew Brees, Aaron Rodgers, and Russell Wilson, to the group comprised of Ryan Tannehill, Andy Dalton, Cam Newton, Joe Flacco, and Eli Manning, over the seven seasons from 2012 to 2018.

If the passer ratings of the former group vary at a significantly higher level than those of the latter group, despite the year-to-year variation in passer ratings for all of them, what should the conclusion be?

Surely we can't reasonably conclude that the former group's passer ratings are significantly higher because they enjoyed significantly better surroundings than the latter group over the entirety of that seven-year period.

First thing is to remember what the question is. The question is whether we can estimate the relative influence of the QB vs. the QB's surroundings on QB performance, in particular to determine which one is larger even if we can't quantify either.

So for purposes of this discussion we need to look at variance in game-by-game passer ratings, not year-by-year ratings because year-by-year ratings already average out a lot of the effect of the surroundings – an advantage if your goal is to estimate QB ability, but a disadvantage if your goal is to estimate the effect of the surroundings. Variance in game-by-game ratings is determined by variance in QB ability, variance in the QB's teammates' abilities, and variance in opponent players' abilities (which is larger than for the QB's teammates because the opponent changes).

Once adjusted to a common year (2019), Tannehill's average game-by-game rating (literally the average across games, which is different than a passing attempt weighted career average rating) is 91.62 with a standard deviation of 26.76, while Wilson's average is 104.24 with a standard deviation of 28 (league average is 91). Note how large those standard deviations are. The difference in means is 12.6 yet the standard deviations are more than twice that!

To make the argument you're making you'd have to show that the influence of the QB multiplied by variation in QB ability is larger than the total influence of the QB's surrounding cast and opponent multiplied by their variation in abilities. It's hard to say which is larger because the influence of the QB on passer rating is huge – the QB is the only constant in that formula – but the QB is also only one person! And variances add (mathematically). So the variations in ability for each of the other players add.

What you are trying to do is compare year-by-year ratings which already wash out tons of variation in the surroundings. Naturally you do that and the main reason for the difference, with large sample size whether it's one QB or many, is that the QB was the main difference. But that's not the right stat to answer the question of which has more influence.

The GuyThe GuyForum Veteran
Nov 29, 2019, 11:45 AM

"cbrad wrote:

First thing is to remember what the question is. The question is whether we can estimate the relative influence of the QB vs. the QB's surroundings on QB performance, in particular to determine which one is larger even if we can't quantify either.

So for purposes of this discussion we need to look at variance in game-by-game passer ratings, not year-by-year ratings because year-by-year ratings already average out a lot of the effect of the surroundings – an advantage if your goal is to estimate QB ability, but a disadvantage if your goal is to estimate the effect of the surroundings. Variance in game-by-game ratings is determined by variance in QB ability, variance in the QB's teammates' abilities, and variance in the opponent and well as the opponent's abilities.

Once adjusted to a common year (2019), Tannehill's average game-by-game rating (literally the average across games, which is different than a passing attempt weighted career average rating) is 91.62 with a standard deviation of 26.76, while Wilson's average is 104.24 with a standard deviation of 28 (league average is 91). Note how large those standard deviations are. The difference in means is 12.6 yet the standard deviations are more than twice that!

To make the argument you're making you'd have to show that the influence of the QB multiplied by variation in QB ability is larger than the total influence of the QB's surrounding cast and opponent multiplied by their variation in abilities. It's hard to say which is larger because the influence of the QB on passer rating is huge – the QB is the only constant in that formula – but the QB is also only one person! And variances add (mathematically). So the variations in ability for each of the other players add.

What you are trying to do is compare year-by-year ratings which already wash out tons of variation in the surroundings. Naturally you do that and the main reason for the difference, with large sample size whether it's one QB or many, is that the QB was the main difference. But that's not the right stat to answer the question of which has more influence.

But the quarterback's surroundings don't typically change significantly on a game-by-game basis, whereas they can on a year-by-year basis (through player acquisitions and losses, coaching changes, etc.).

Note for example the huge increase in Tom Brady's season passer rating in 2007, presumably as a function of the receiving trio of Randy Moss, Wes Welker, and Donte Stallworth. His passer rating probably didn't vary from game to game any more that season than it has from game to game in any other season of his career, but I suspect it varied from game to game at a level significantly higher that year than it has in any other year of his career.

So if it's true that a QB's performance is determined more by his surroundings than by his own ability (or vice-versa), shouldn't a year-by-year comparison, across multiple QBs and teams, provide the best test of that hypothesis?

There's nothing inherently stopping us from finding elevated standard deviations analogous to the ones you pointed out above, but on a within-group basis in the comparison I outlined in the post you quoted. Whether those would outweigh the between-group variation would be the question.