TANNEHILL VERSUS THE WORLD: THE IN-DEPTH COMPARISON TANNEHILL’S CRITICS DON’T WANT YOU TO READ

"cbrad wrote:

Well it's not difficult to test. Here are some GB forums:

http://www.footballsfuture.com/phpBB2/viewforum.php?f=17
http://packerrats.com/forumdisplay.php?9-Green-Bay-Packers
https://www.packerforum.com/forums/packer-fan-forum.20/
http://forums.prosportsdaily.com/forumdisplay.php?89-Green-Bay-Packers

Basically the topics have little to do with the QB or it's appreciation of the QB (even entire threads on that), and there's some debate as to how much others benefit from Rodgers. But no QB controversy or anything close to it I can find in ~10 minutes browsing some threads.

So jdang is right here it seems (as in not a lot of critical debate).

Being up here in GB country I can tell you that they LOVE Rodgers. However, at times I do hear and read criticism about him from GB fans. I got into a debate with a Packers fan I work with at the beginning of last season. Just to be clear, he's a very informed and smart guy (he's a cardiologist) and has been a Packers fan his entire life. He was saying that he thought the Packers should trade Rodgers "soon" because he's fading.

Anyhow, and to save space because that debate was pretty long, I think fans get emotional over their teams and a lot of times don't use critical thinking. It's mostly knee-jerk comments and opinions based upon the "right now". Rodgers has a string of bad games and some fans want him gone. I've seen it with Pats fans and even with Dolphins fans when Marino was playing. Currently we see it with RT. Fans get tunnel vision because they believe the "QB's win Super Bowls" mantra that the media puts out there and think they need a QB who is Rodgers or Marino in order to win a Super Bowl. In reality you need a good team to win a Super Bowl.

"dirtylandry wrote:

how that guy is qb still there, if you ask fans, is surprising

It depends on which fans you ask. A lot felt like Dalton would have won the SB with Houstons team last year.

I can't really get too enthused about rankings based articles. For me the difference from the average is more important than the rank. The z-score in particular tells you far more than ranking.

"dirtylandry wrote:

I'm actually tired of reading and debating Tannehill. I just feel the past (based on who we should have drafted or traded for after Tannehill) is a waste of time. I also feel the past performances of him, whether it was himself, the line, the coaching, the run game, no TE, or WRs, is history. Miami has made strides and not sat on their hands. Their FO has changed, their coaching has changed, their personnel has changed and their plans have changed. With this transformation, it is all on Tannehill now. There are no more excuses, whether it is him or not. Jamil Douglas, Billy Turner and Dallas Thomas are gone. We have a team to support him and it is not crazy to expect him to carry us for a game or two. We should expect elite play, because in year 2, Gase has quietly built an offensive juggernaut

Do you realize you actually just debated it, even though you're tired of it? :lol: As someone else said, you will debate it, and you will like it! :lol:

Sorry dirty... I had to. :lol: :up:

"Pauly wrote:

I can't really get too enthused about rankings based articles. For me the difference from the average is more important than the rank. The z-score in particular tells you far more than ranking.

I just decided not to comment on the article itself (up till now obviously!) because the "methodology" is useless to put it nicely and it's not like QBR from ESPN where I think it's really important to push back on a serious attempt to pass off something that isn't even a statistic as a stat.

But since you bring it up lol..

Let me say that while what you say is true because z-scores actually measure things and ranks don't (for those who don't know.. z-scores are standard deviation units and actually measure the distance between points while rank isn't a measure because the "distance" between rank 3 and 4 need not mean the same thing as the "distance" between rank 10 and 11), you could use z-scores everywhere and the "methodology" would still be.. how should I put this nicely? Useless lol.

The biggest single issue is that multiple stats that shouldn't be treated equally in importance are weighted equally when combined into a single stat. For example, completions per game has essentially a ZERO correlation with win% yet the rank on completions is treated equal to stuff like Y/A which has a relatively high correlation??

The absolute minimum would be to weight the different components by correlation to win%, or absent that just choose only those stats that have similar correlations to win% (at least the components of passer rating have similar correlations to win% or the first thing you'd point out with passer rating is what I just said).

But with so many components I'd argue just weighting things by correlation to win% might be problematic too because some of those components are highly correlated with each other, meaning you're measuring similar things in two or more components effectively recreating the problem, etc..

Anyway, there are ways to correct for all that but no need to even go that far. Other issues are the use of subjective cutoffs and pff-style stuff (e.g. Fahey) mixed in with a pure stats-based approach. It's probably best to keep those worlds separate because mixing in pff-style stuff with stats just ruins the stats based approach.

OK.. rant (mostly) over. Just going to point out something you might be interested in: there are ways of turning multiple rankings into a single measure (bypassing the need to use stuff like z-scores). There's something called non-metric multidimensional scaling that will take a whole set of different rankings and reduce its dimensionality. So if you have N rankings you can estimate where the points lie in any dimension less than N. In our case you'd estimate where they lie in a 1-D space, and when you do that you get a true measure, just like z-scores.

"dirtylandry wrote:

The line was only intact for 4-5 games

LOL, I said we had the pieces. It's not my fault those pieces were broken...we still had em! =)

"dirtylandry wrote:

how that guy is qb still there, if you ask fans, is surprising

To be fair, he was pretty mediocre the first few years here. But year 5, he blew up. Year 6, came back down. If he has another strong year, you can accept a QB that has a good year, a mediocre year, etc.

Dalton was the best argument for being patient with Tanny. If it's just a one hit wonder though, I'm looking at possible projects in the draft.

"cbrad wrote:

The biggest single issue is that multiple stats that shouldn't be treated equally in importance are weighted equally when combined into a single stat. For example, completions per game has essentially a ZERO correlation with win% yet the rank on completions is treated equal to stuff like Y/A which has a relatively high correlation??

That's exactly why I wasn't going to comment either. That article had a ton of work/research behind it and I wasn't going to try to discredit the stats as not being a true measure. Completion percentage is definitely not equal to TD's or pass attempts, for instance. The rushing stats aren't exactly relevant since we know Tannehill was told to stay in the pocket the first few years and never run.

And of course, the whole "unavoidable sacks" stat is BS since it's calculation formula is BS. The "interceptable passes" is also a convenient stat for a team that throws half their passes within 1 yard of the LOS. I put very little faith in Fahey in general, so I'm bothered by any analysis based on Fahey's analysis. For instance, when we twist a stat into overall accuracy....Tannehill's dump off passes to a RB in 2013 should be more accurate than Cam or Carr throwing 6-10 yards. That's not a metric to measure with, it's common sense since our line stunk and we had more 3 and outs than anyone back then. Yet here's a metric saying that Tannehill was thriving when he clearly wasn't.

I do appreciate the effort because a ton of work went into that article, but there was a lot taken for granted as well.

Why is it a negative that a raw QB with limited college starts wasn't top 5 in the league his first couple of years? Especially considering the lackluster offensive weapons during those years, the bad oline, and the coaching carousel?

"resnor wrote:

Why is it a negative that a raw QB with limited college starts wasn't top 5 in the league his first couple of years? Especially considering the lackluster offensive weapons during those years, the bad oline, and the coaching carousel?

Well, when you put it like that......makes sense. I was really expecting more from Bill Lazor's offense. Forgetting talent, the playcalling was questionable

Sent from my iPhone using Tapatalk

"resnor wrote:

Why is it a negative that a raw QB with limited college starts wasn't top 5 in the league his first couple of years? Especially considering the lackluster offensive weapons during those years, the bad oline, and the coaching carousel?

Because we had to wait for him to develop with no guarantee he would develop. Let's hope he keeps developing but don't be so critical of posters who didn't want their NFL team to groom a developmental QB.

I don't even expect top 5 (long shot) but his peak has been top 12'ish. Top 8 is what I want my QB to be, if possible.

"jdang307 wrote:

Because we had to wait for him to develop with no guarantee he would develop. Let's hope he keeps developing but don't be so critical of posters who didn't want their NFL team to groom a developmental QB.

I don't even expect top 5 (long shot) but his peak has been top 12'ish. Top 8 is what I want my QB to be, if possible.

Think about it- the Fins led the league in 3 and outs and had one of the league's lowest scoring offenses for several of the years mentioned here. And that proves Tannehill's growth is better than Marino's? Oh hell no- I think it shows that the team really sucked in all facets of the offense. Just because we're doing better now doesn't mean there was phenomenal growth in our QB.

You can't have it both ways. If the line and the receivers are the reason Tannehill had some rough years, then they're also a reason why he had a good year. You can't blame all the bad on others, credit all the good on Tannehill, and then cherry pick stats to call something objective. The truth of the matter is that Tannehill is a solid young QB with tons of potential. Either you believe that or you don't, but stats aren't ever going to prove anything about one component in a team sport.

The only true comparison we have is what Moore has done on the field in the exact same role with the exact same surroundings; that's the only head-to-head, unbiased data available here. And remember, Moore hadn't had serious reps in about five years- Tannehill got well over 90% of those in practice. So it's still an unfair comparison really...Moore should get bonus points somehow. Yet nobody wants to hear those head-to-head stats-

Highest scoring game? Matt Moore
Best TD/Int ratio? Matt Moore
Best third down efficiency? Matt Moore
Best overall QB rating? Matt Moore

Can we see how stats work? Matt Moore is clearly the better QB if we only look at the numbers. Football is not about numbers though, it's about execution and winning football games. Stats only tell a part of that story...it's never the story itself.

So I'll say it again- Tannehill is an insanely talented young QB with amazing potential.

Let's not try voting him into the HOF just yet though...maybe we should wait until we consistently beat a few playoff teams and actually win in the playoffs. Or hell, back to back winning records would work. But you're not going to sell that the most sacked QB over the last 5 seasons with one of the worst conversion percentages on 3rd down is a top 5 QB. That's about as objective as I can possibly be.

"jdang307 wrote:

Because we had to wait for him to develop with no guarantee he would develop. Let's hope he keeps developing but don't be so critical of posters who didn't want their NFL team to groom a developmental QB.

I don't even expect top 5 (long shot) but his peak has been top 12'ish. Top 8 is what I want my QB to be, if possible.

We "had to wait for him to develop"?? Seriously? You have to wait for basically EVERY QB to develop. Even Andrew Luck hasn't been an elite QB, and he had a much better pedigree and resume. That attitude, actually, I think is what has fostered so much of the disagreements and arguments on the site.

"dirtylandry wrote:

Well, when you put it like that......makes sense. I was really expecting more from Bill Lazor's offense. Forgetting talent, the playcalling was questionable

Sent from my iPhone using Tapatalk

I did some number crunching on some other threads, and Lazor's playcalling was really bad whenever the Phins were playing from behind. The TL;DR version is that Tannehill suffered a 15 point drop off in passer rating when behind with Lazor's playcalling but when he was under Sherman or Gase there is no drop off when behind. Also in year 2 of Lazor (when every man and his dog knew the restrictions Lazor had put on audibles/changing plays at the LOS) Tannehill had a really bad rating when blitzed, but under Gase Tannehill had great numbers.

"resnor wrote:

We "had to wait for him to develop"?? Seriously? You have to wait for basically EVERY QB to develop. Even Andrew Luck hasn't been an elite QB, and he had a much better pedigree and resume. That attitude, actually, I think is what has fostered so much of the disagreements and arguments on the site.

You just put forth an argument about why he's taking longer to develop (limited college starts).

Andrew Luck in year 3, threw for 40 TDs. Something only 8 QBs have done, EVER. 4700 yards, 7.7 ypa. 7th highest passer rating. 40 TDs is something Tanny may never pull off. You hope he gets to 30-34 consistently like Rivers.

Andrew Luck on a torn shoulder had a damn good year individually in 2016 (and let's not talk about their surrounding circumstances .. nobody on here is trading Indy for Miami). Tanny, had a promising year.

Indy fans hope Luck gets even better. But he's developed.

Kirk Cousins, year 4, first year starting. Developed. Marcus Mariota, looks pretty good In a new offense, 2nd year starter. He had college starts but not in a pro-style offense, he came from a run and shoot. Wilson. Carr. There are plenty of QBs with different situations that developed pretty quick in the NFL.

So when you take someone raw at #8 overall, and you start him right away, you're literally giving him a college education in the pros. It is what it is, and if some fans who spend a lot of money disagree with that, I think it's okay. I don't care, and that's why I don't use it as an excuse.

He was drafted and started in the NFL. These are real games. If he started 1 game in college, or 44, doesn't matter to me. He was inserted into an NFL game and I'll judge his performance as is. You get handed the keys to the franchise, you're getting judged day 1 like everyone else.

He's here, he's solidified the next two years starting at least. What's in the past, is that. But if someone wants to criticize that it took 5 years just to get here, I think that's fair.

"cbrad wrote:

I just decided not to comment on the article itself (up till now obviously!) because the "methodology" is useless to put it nicely and it's not like QBR from ESPN where I think it's really important to push back on a serious attempt to pass off something that isn't even a statistic as a stat.

But since you bring it up lol..

Let me say that while what you say is true because z-scores actually measure things and ranks don't (for those who don't know.. z-scores are standard deviation units and actually measure the distance between points while rank isn't a measure because the "distance" between rank 3 and 4 need not mean the same thing as the "distance" between rank 10 and 11), you could use z-scores everywhere and the "methodology" would still be.. how should I put this nicely? Useless lol.

The biggest single issue is that multiple stats that shouldn't be treated equally in importance are weighted equally when combined into a single stat. For example, completions per game has essentially a ZERO correlation with win% yet the rank on completions is treated equal to stuff like Y/A which has a relatively high correlation??

The absolute minimum would be to weight the different components by correlation to win%, or absent that just choose only those stats that have similar correlations to win% (at least the components of passer rating have similar correlations to win% or the first thing you'd point out with passer rating is what I just said).

But with so many components I'd argue just weighting things by correlation to win% might be problematic too because some of those components are highly correlated with each other, meaning you're measuring similar things in two or more components effectively recreating the problem, etc..

Anyway, there are ways to correct for all that but no need to even go that far. Other issues are the use of subjective cutoffs and pff-style stuff (e.g. Fahey) mixed in with a pure stats-based approach. It's probably best to keep those worlds separate because mixing in pff-style stuff with stats just ruins the stats based approach.

OK.. rant (mostly) over. Just going to point out something you might be interested in: there are ways of turning multiple rankings into a single measure (bypassing the need to use stuff like z-scores). There's something called non-metric multidimensional scaling that will take a whole set of different rankings and reduce its dimensionality. So if you have N rankings you can estimate where the points lie in any dimension less than N. In our case you'd estimate where they lie in a 1-D space, and when you do that you get a true measure, just like z-scores.

Great rant. Very entertaining and informative as usual.

I like to read Fahey and think that he makes some very good points in his analysis. However I do think that his own biases affect his attempts to turn his opinions into data. I question his accuracy in regards to Tannehill because he seems to have taken a position as a Tannehill advocate. I agree with the direction of his opinion but not the degree.

"KeyFin wrote:

It looks like a lot of that article was written at the start of 2016...although it alludes to being the full season.

Is this a different article now? The original article when first posted is NOT the article that is linked to now.

So the author eliminated Kirk Cousins. Weird.

"Pauly wrote:

I did some number crunching on some other threads, and Lazor's playcalling was really bad whenever the Phins were playing from behind. The TL;DR version is that Tannehill suffered a 15 point drop off in passer rating when behind with Lazor's playcalling but when he was under Sherman or Gase there is no drop off when behind. Also in year 2 of Lazor (when every man and his dog knew the restrictions Lazor had put on audibles/changing plays at the LOS) Tannehill had a really bad rating when blitzed, but under Gase Tannehill had great numbers.

You had me at "number crunching".....

"jdang307 wrote:

Is this a different article now? The original article when first posted is NOT the article that is linked to now.

So the author eliminated Kirk Cousins. Weird.

It's the same article. Both articles had Cousins eliminated.

"danmarino wrote:

It's the same article. Both articles had Cousins eliminated.

Naw, earlier in the first paragraph it said something to the effect of Wilson not being good from the pocket, and it looked old. Hence my first post. Now when I click it, that paragraph is completely gone. Weird.

I thought you had me on ignore? Lol.

"jdang307 wrote:

Naw, earlier in the first paragraph it said something to the effect of Wilson not being good from the pocket, and it looked old. Hence my first post. Now when I click it, that paragraph is completely gone. Weird.

It's possible. Inspect the page source and it says:

<meta property="article: published_time" content="2017-06-15T01:19:58+00:00" />
<meta property="article:modified_time" content="2017-06-15T12:10:06+00:00" />

Your post #2 in this thread was supposedly posted at 11:45am ET yesterday, which is right before the article was last modified according to the page source (assuming there's no time zone issue.. in any case it was modified).

"jdang307 wrote:

You just put forth an argument about why he's taking longer to develop (limited college starts).

Andrew Luck in year 3, threw for 40 TDs. Something only 8 QBs have done, EVER. 4700 yards, 7.7 ypa. 7th highest passer rating. 40 TDs is something Tanny may never pull off. You hope he gets to 30-34 consistently like Rivers.

Andrew Luck on a torn shoulder had a damn good year individually in 2016 (and let's not talk about their surrounding circumstances .. nobody on here is trading Indy for Miami). Tanny, had a promising year.

Indy fans hope Luck gets even better. But he's developed.

Kirk Cousins, year 4, first year starting. Developed. Marcus Mariota, looks pretty good In a new offense, 2nd year starter. He had college starts but not in a pro-style offense, he came from a run and shoot. Wilson. Carr. There are plenty of QBs with different situations that developed pretty quick in the NFL.

So when you take someone raw at #8 overall, and you start him right away, you're literally giving him a college education in the pros. It is what it is, and if some fans who spend a lot of money disagree with that, I think it's okay. I don't care, and that's why I don't use it as an excuse.

He was drafted and started in the NFL. These are real games. If he started 1 game in college, or 44, doesn't matter to me. He was inserted into an NFL game and I'll judge his performance as is. You get handed the keys to the franchise, you're getting judged day 1 like everyone else.

He's here, he's solidified the next two years starting at least. What's in the past, is that. But if someone wants to criticize that it took 5 years just to get here, I think that's fair.

Every QB takes time to develop. Every. Damn. One. So, if people want to have knee jerk reactions, then fine, they can do that.

But they should be honest, and just admit that was what they were doing, and they should expect more reasonable people to dispute their knee jerk reactions.

"resnor wrote:

Every QB takes time to develop. Every. Damn. One. So, if people want to have knee jerk reactions, then fine, they can do that.

But they should be honest, and just admit that was what they were doing, and they should expect more reasonable people to dispute their knee jerk reactions.

How many games did it take Marino to develop?

Or Russell Wilson.

100.0 passer rating in his rookie year 2012 (9 regular season games with 100+ rating that year). And it wasn't a fluke either. 101.2 in 2013, 95.0 in 2014, 110.1 in 2015 and 92.6 in 2016. Played great right out of the gate:
http://www.pro-football-reference.com/players/W/WilsRu00.htm

Not every QB needs development time.

Ryan Tannehill Is the NFL’s Most Underrated Quarterback

https://brickwallblitz.wordpress.com/2017/06/17/ryan-tannehill-is-the-nfls-most-underrated-quarterback/