Ryan Tannehill

"The Guy wrote:

How does that comport with the fact that Tannehill played the best football of his career under Gase?

Tannehill’s best statistical year was his first year under Lazor.

Titans have the 26th ranked run blocking o line and 31st ranked pass pro one this year according to pro football outsiders. Whether or not you buy into that ranking in full or not doesn’t much matter. It hasn’t been a carry tbe qb group lets put it that way.

Is it possible that Ryan’s just now getting back to pre injury in 2016 form when he was playing a plus qb and having missed what 2.5 years of game action 2018 can be chalked up to rust and concern about the stability of his knee? I think the rust thing in terms of timing played out on tape in 18. He was a tick late with the ball out of hand relative to route concept a lot in 18. He’s not this year. Maybe the knee thing is a reach maybe it’s not. I just know it doesn’t seem to be a concern this year. No knee brace and moving free and easy pre injury like.

Also he doesn’t have a utter eye sore in his lead leg in Dallas thomas anymore either. Another plus. Just about every time he dropped back thru 2016 week 5 after which thomas was cut by gase that dude was in his lead leg or throwing arm even.

And don’t get me started on tbe cutting billy turner after being in a dire left tackle situation for one week and throwing him in there witb little to no reps even. That was just overreaction by gase. He’s playing a decent athletic right guard in Green Bay these days. Which is where he should have been in Miami all along but I digress.

there’s been a lot of dumpster fire o line play in Miami in tannehills tenure. O lines that just about to a man graded out bottom level at their position. Be it run or pass pro.

It shouldn't matter, we were never going to keep him here. He had six years to show us something. Last season....maybe call that a rehab season after the knee injury, but the point being...He had his opportunities. Sure, playing under Philbin and Adam F Gase stunted him, but his best season coincidentally was the 2016 one under Adam Gase...Right before the knee injury. I think the knee injury changed a lot more than we all think. It was a major setback for a player that was finally peaking at the right time of his career.
I've noticed a trend with players that have knee injuries or basically any injuries where they miss a ton of time. That trend being the first year back healthy, they spend most of that year shaking off the rust and gaining confidence back (unless you're Adrian Peterson). We saw it last year with Dalvin Cook and Deshaun Watson. Both came off of knee injuries that required a lot of time to rehab and both looked a step off. This year Cook leads the league in rushing and Watson is in the running for MVP.
It's no surprise to me that a change of scenery, with better coaching, better running game, better O line, and a year of rehabilitation under his belt that Tannehill is flourishing in Tennessee. I'm happy for the guy.

"cbrad wrote:

Yeah, any OL that surrenders more "jailbreak" pressure will also surrender more pressure in general. There's no plausible mechanism for that not to be the case. Furthermore, you could probably model the percent of "jailbreak" pressure fairly well by treating the probability that any player on the OL makes a mistake as independent of the probability that any other player will make a mistake, since mistakes are made by individuals. And that would mean the probability of "jailbreak" pressure is just a calculation of joint probability, which means you could probably infer the distribution of pressure due to 1, 2 or 3 players from the overall pressure statistic.

Can't test that directly yet, but I bet it's a good approximation.

Which brings us to another important point: people claiming Tannehill had more "jailbreak" pressure than most other teams aren't actually providing any evidence for it. It's all assumption based on watching primarily the Dolphins OL, which I agree was bad. But this season has been interesting for me regarding OL play because I am watching more games by other teams, and man.. there are a LOT of teams that allow tons of this "jailbreak" pressure. Dolphins don't look at all special with regards to the percent of "jailbreak" pressure when you watch the rest of the league. Lots of QB's have to keep scrambling around to avoid it.

Actually there is a plausible mechanism for what's being described. If you are a coach and you know your team is going to give up pressure fairly quickly you can game plan around it ie bubble screens, RB and TE left in for extra blockers, run heavy game plan, move the QB from position. You do all of these accepting that it will make your offense more and more anemic, but keeps the pressure relatively low.

What you can't do is game plan around it when you have no choice but to go for the win in the game or make a serious effort to pick up 3rd and longs. At that time the defense will have the ability to bring extra pressure giving more jailbreak pressure opportunity. thus you can have more jail break pressures without a noticeable uptick in number of pressures across the average of the game.

I'm actually not saying that's what happened in Miami cause I didn't research that, I just wanted to bring up that its not out of the realm of possibility to see that discrepancy.

"cbrad wrote:

Note that the section "How is OSM calculated" gives you absolutely no idea how it's calculated.

I noticed that too and immediately discarded OSM as a thing. With that said though, Tannehill has looked good this season and he is playing well. I really don't need an advanced metric or a fancy algorithm to show me that.

"Pauly wrote:

Tannehill’s best statistical year was his first year under Lazor.

By the absolute slimmest of margins however. Let me post the z-scores again because we're going to look at this when this season is over:

Tannehill z-scores:
2012: -0.8259
2013: -0.3630
2014: 0.3940
2015: -0.1698
2016: 0.3739
2018: -0.0167

So yes 2014 was technically the best but as you can see 2014 and 2016 are essentially identical. They were also the only years he was above average. What is Tannehill's z-score right now? It's a very impressive 1.0474. The average z-score for the entire offense for SB winners is 0.9389 so if Tannehill's current level is his average level this year then that's more than sufficient to build a SB team around.

But again.. sample size is small and nothing Tannehill has done so far statistically or otherwise is abnormal for 136 passing attempts.

And so it's clear what the statistical test will be at the end of the year, we'll do a 2-sample t-test for unequal samples sizes on his game-by-game ratings after adjusting ratings to a common year and compare his entire career in Miami vs. Tennessee to test whether those two sets of ratings "come from the same QB". We can also do an ANOVA to test whether the adjusted ratings from all years "come from the same QB" (this test separates the data by year).

Anyway.. that will show whether you can explain whatever Tannehill does in Tennessee given what he did in Miami by random variation alone, or whether there is something statistically significant with his performance in Tennessee.

"Phins_to_Win wrote:

Actually there is a plausible mechanism for what's being described. If you are a coach and you know your team is going to give up pressure fairly quickly you can game plan around it ie bubble screens, RB and TE left in for extra blockers, run heavy game plan, move the QB from position. You do all of these accepting that it will make your offense more and more anemic, but keeps the pressure relatively low.

What you can't do is game plan around it when you have no choice but to go for the win in the game or make a serious effort to pick up 3rd and longs. At that time the defense will have the ability to bring extra pressure giving more jailbreak pressure opportunity. thus you can have more jail break pressures without a noticeable uptick in number of pressures across the average of the game.

I'm actually not saying that's what happened in Miami cause I didn't research that, I just wanted to bring up that its not out of the realm of possibility to see that discrepancy.

That alone won't do it. You have to explain why it would be plausible for all coaches other than Miami's coach to NOT game plan when you know your OL is bad because if "game planning around issues with the OL" is not unique to Miami it won't affect the overall stats. And the only way around that critique would be to start off by assuming that only Miami's OL is so bad that you'd need to game plan around it. The problem with that is we know many coaches game plan like you're describing.

Besides.. given how bad Gase was, if anything it would be the other way around right? We did NOT game plan as well as we should have which would make it less likely for your scenario to occur specifically in Miami.

Ehh the first 5 weeks of 2016 the team was a dumpster fire around the qb.

that’s the only reason 2014 and 2016 were close in terms of his “best season”
And of course he got hurt before he could finish the season.

"cbrad wrote:

A case study in what happens when you don't adjust stats for era. I mean.. Tannehill's career passer rating is 87.8 while Marino's is 86.4, so if you don't adjust for era it's pretty clear: Tannehill > Marino using stats right?

Yeah.. adjust for era and Tannehill is average while Marino is one of the best ever, which IMO is a far more accurate representation.

You misunderstand my post. Was I making a comparison that Tannehill is better than Marino? SACRILEGE!!!!!! Marino is the GREATEST quarterback to ever play the game!!! I don't care who says otherwise.

The point I was making was simple. We asked Tannehill to produce. He produced. The remainder of the team, most specifically the coaching staff did not and yet, who is still frowned upon and looked down upon? Tannehill!

I bet Titans fan are elated they have Tannehill.

"jdallen1222 wrote:

12 pages on a Tannehill thread lol.

LOL. Once again Ryan Tannehill is proving to be the most polarizing player on this forum!

All I know is the guy is 3-1 since he took the starting job from Mariotta. In the end that's all that matters and it remains to be seen how far he can take them. It'll be interesting to watch.

Good for him and the Titans.

"cbrad wrote:

That alone won't do it. You have to explain why it would be plausible for all coaches other than Miami's coach to NOT game plan when you know your OL is bad because if "game planning around issues with the OL" is not unique to Miami it won't affect the overall stats. And the only way around that critique would be to start off by assuming that only Miami's OL is so bad that you'd need to game plan around it. The problem with that is we know many coaches game plan like you're describing.

Besides.. given how bad Gase was, if anything it would be the other way around right? We did NOT game plan as well as we should have which would make it less likely for your scenario to occur specifically in Miami.

Actually the fact that we ran an inordinate amount of screen and short passes under Gase can in a way explain why our pressure % would be mediocre but our line terrible.

"The_Dark_Knight wrote:

The point I was making was simple. We asked Tannehill to produce. He produced. The remainder of the team, most specifically the coaching staff did not and yet, who is still frowned upon and looked down upon? Tannehill!

Very well.. but let's be clear. Statistically speaking Tannehill's production was that of an average QB (a tad bit below actually). Those of us who "frown" on Tannehill are only doing so because we want consistently above average production.

"The_Dark_Knight wrote:

I bet Titans fan are elated they have Tannehill.

I'm sure they are. The question is whether that will last or not. It's only been 136 passing attempts. Either way, what's happening right now is really good because it provides a good test of the effects of different surroundings on Tannehill.

Let's just all remember this isn't the first such "test". For many years people predicted next year was the year Tannehill would be "top 10" or so. Even last year, in the 2018 preseason, there were lots of posters claiming this was finally the year Tannehill would produce because he had a coach that believed in him (can't fault Gase for not believing in Tannehill) and because Gase had "his team".

So we'll see!

"AGuyNamedAlex wrote:

Actually the fact that we ran an inordinate amount of screen and short passes under Gase can in a way explain why our pressure % would be mediocre but our line terrible.

The entire NFL has moved in the direction of more screen passes. It's not like the Dolphins are an outlier. And the bad OL argument has been a constant for us, well before Gase. Also, we had a good OL with Gase during certain stretches like in 2016.

So it's going to have to be something truly unique to the Dolphins over an extended period of time.

"Hoops wrote:

Ehh the first 5 weeks of 2016 the team was a dumpster fire around the qb.

that’s the only reason 2014 and 2016 were close in terms of his “best season”
And of course he got hurt before he could finish the season.

Of course if you remove all the bad games you end up with better stats. Here's another interesting result using ranks for the moment: Tannehill was ranked #12 by passer rating among all starting QB's at the end of 2016. He was also ranked #12 when you ONLY consider that "best consecutive 8 game stretch" from games 6-13 if you compare to all other QB's "best consecutive 8 game stretches".

Just pointing out what the data look like from a statistician's perspective. It's consistent with random variation. So sure one can try to explain everything by looking for an explanation, but it's also consistent with the type of random variation you generally see for most QB's across the league. Anyway.. we'll see what happens this year.

"cbrad wrote:

That alone won't do it. You have to explain why it would be plausible for all coaches other than Miami's coach to NOT game plan when you know your OL is bad because if "game planning around issues with the OL" is not unique to Miami it won't affect the overall stats. And the only way around that critique would be to start off by assuming that only Miami's OL is so bad that you'd need to game plan around it. The problem with that is we know many coaches game plan like you're describing.

Besides.. given how bad Gase was, if anything it would be the other way around right? We did NOT game plan as well as we should have which would make it less likely for your scenario to occur specifically in Miami.

Well we aren't talking about just bad teams, we are talking about averages.

So we will have teams that have a great OL that will push their game plan and not attempt to cover for bad OL plays. They will still see pressure, but they will most likely see pressure based on the overall performance of an OL that is much higher and consistent then most other teams in the NFL.

Then we will have teams that only have a good OL, that may flirt with opening up the offense but might implement some of the OL covers as needed.

Then you will have the bad OL that the coach's entire game plan revolves around covering for the bad OL play until they absolutely have to try to win the game.

It is very possible that all 3 of these scenarios give up the exact same amount of pressure in the game(I would wager there are more tiers then just the 3 I'm showing here), but as you move down the list your offense suffers 1st from the very process of trying to cover for the bad O-line and 2nd from the dangerous point where you have to open up the offense and pray the Offensive line doesn't get your QB killed.

I'm not pointing out that Miami was unique in its bad OL play, I'm saying they fell into the 3rd example, and probably suffered equally with jailbreak pressures as other really bad O-line teams did, once the team had to stop covering for the OL weakness. But these teams would all be taking more jailbreak situations then the other instances.

What I am pointing out is that it is very possible to have 1 team on one end of the spectrum give up the exact same number of pressures as a team on the other end of the spectrum, but you should be still expecting considerably different results from the 2 teams that has nothing to do with QB play. And it is possible the jail break scenario happens as a larger percentage of the pressures in the 3rd example.

I'm not saying Miami is unique, I'm saying there are probably one of only 6 or 7 teams that experience the same situation as Miami did.

"cbrad wrote:

The entire NFL has moved in the direction of more screen passes. It's not like the Dolphins are an outlier. And the bad OL argument has been a constant for us, well before Gase. Also, we had a good OL with Gase during certain stretches like in 2016.

So it's going to have to be something truly unique to the Dolphins over an extended period of time.

And the league has a problem with oline play.

"Phins_to_Win wrote:

Well we aren't talking about just bad teams, we are talking about averages.

So we will have teams that have a great OL that will push their game plan and not attempt to cover for bad OL plays. They will still see pressure, but they will most likely see pressure based on the overall performance of an OL that is much higher and consistent then most other teams in the NFL.

Then we will have teams that only have a good OL, that may flirt with opening up the offense but might implement some of the OL covers as needed.

Then you will have the bad OL that the coach's entire game plan revolves around covering for the bad OL play until they absolutely have to try to win the game.

It is very possible that all 3 of these scenarios give up the exact same amount of pressure in the game(I would wager there are more tiers then just the 3 I'm showing here), but as you move down the list your offense suffers 1st from the very process of trying to cover for the bad O-line and 2nd from the dangerous point where you have to open up the offense and pray the Offensive line doesn't get your QB killed.

I'm not pointing out that Miami was unique in its bad OL play, I'm saying they fell into the 3rd example, and probably suffered equally with jailbreak pressures as other really bad O-line teams did, once the team had to stop covering for the OL weakness. But these teams would all be taking more jailbreak situations then the other instances.

What I am pointing out is that it is very possible to have 1 team on one end of the spectrum give up the exact same number of pressures as a team on the other end of the spectrum, but you should be still expecting considerably different results from the 2 teams that has nothing to do with QB play. And it is possible the jail break scenario happens as a larger percentage of the pressures in the 3rd example.

I'm not saying Miami is unique, I'm saying there are probably one of only 6 or 7 teams that experience the same situation as Miami did.

OK.. so let's say we take your third scenario where the coach attempts to cover up for OL weaknesses, and as long as the game situation permits this leads to equal amounts of overall pressure (by your assumption or just hypothetically). So you're saying that you'd see more "jailbreak" pressure when the coach can no longer cover up for OL weaknesses. I agree.

But how is it possible that in that case the OL doesn't give up more pressure in general? Again, if you game plan so that you reduce or increase the probability any player on the OL gives up pressure, then the conditional probability that 2 players give up pressure, or that 3 or more give up pressure is mathematically determined. So you can change the total amount of pressure but not the relative frequency of different types of pressure.. unless of course you get rid of "random variation" in how players play, which is untenable IMO.

"cbrad wrote:

OK.. so let's say we take your third scenario where the coach attempts to cover up for OL weaknesses, and as long as the game situation permits this leads to equal amounts of overall pressure (by your assumption or just hypothetically). So you're saying that you'd see more "jailbreak" pressure when the coach can no longer cover up for OL weaknesses. I agree.

But how is it possible that in that case the OL doesn't give up more pressure in general? Again, if you game plan so that you reduce or increase the probability any player on the OL gives up pressure, then the conditional probability that 2 players give up pressure, or that 3 or more give up pressure is mathematically determined. So you can change the total amount of pressure but not the relative frequency of different types of pressure.. unless of course you get rid of "random variation" in how players play, which is untenable IMO.

Dude, you watched the games right? Teams were routinely rushing only 4, and getting 2 or 3 guys through to the QB. So teams were getting significant pressure WITHOUT blitzing. Which then allows there to be more people out in coverage, which then causes they're to be no open receivers, especially when they aren't good for runners.

"resnor wrote:

Dude, you watched the games right? Teams were routinely rushing only 4, and getting 2 or 3 guys through to the QB. So teams were getting significant pressure WITHOUT blitzing. Which then allows there to be more people out in coverage, which then causes they're to be no open receivers, especially when they aren't good for runners.

That happened sometimes yes. All that shows is we had a bad OL. That is in no way evidence that there was a higher percentage of "jailbreak" pressure than "non-jailbreak" pressure, just that our OL gave up more pressure in general (which implies more jailbreak pressure). Again, you see a lot of this "jailbreak" pressure in other games too. We're not arguing about whether we had a bad OL here.

"cbrad wrote:

OK.. so let's say we take your third scenario where the coach attempts to cover up for OL weaknesses, and as long as the game situation permits this leads to equal amounts of overall pressure (by your assumption or just hypothetically). So you're saying that you'd see more "jailbreak" pressure when the coach can no longer cover up for OL weaknesses. I agree.

But how is it possible that in that case the OL doesn't give up more pressure in general? Again, if you game plan so that you reduce or increase the probability any player on the OL gives up pressure, then the conditional probability that 2 players give up pressure, or that 3 or more give up pressure is mathematically determined. So you can change the total amount of pressure but not the relative frequency of different types of pressure.. unless of course you get rid of "random variation" in how players play, which is untenable IMO.

I feel like you nailed my point in the first paragraph, but I'm not sure I follow on your 2nd.

In my 3rd scenario, the OL would have given up more pressures then normal, but the strategy prevented the pressure from registering. What I'm saying is if you reduce the number of chances the defense gets at getting pressure, but of those chances they get more jailbreaks, you would have the scenario I described. For example if we give them 10 drop backs that they get to go against the O-line with no shielding, and they pressure on 9 of them but 6 of them are jailbreak styles, 9 for the course of a game isn't a lot, but 6 of them being jailbreaks is probably way above average for that scenario.

Left to their own devices the OL would have given up tons of pressure, and because we had multiple bad O-line when we finally leave it up to the O-line to protect, the defense would take advantage and get the jailbreak pressure we are talking about, however, due to the relative few times they get that opportunity they don't show a larger then normal overall number of pressures.

Just to redefine the argument, I was saying that it was possible to get more jail break pressures, while still not giving up more overall pressures.

My scenario is predicated on the fact that an O-line as a unit is bad enough that it would give up a higher % of jailbreak pressures in relation to the overall number of pressures, but mitigation tactics is the reason that the overall number of pressure is not extremely high.

And again I was mainly pointing this out as a scenario where this metrics can exist in the NFL, not that specifically the Dolphins did exactly as I said. I THINK the Dolphins might have done this, but I can only say that from an eye test and as you have pointed out before, that's not always reliable.

"cbrad wrote:

Of course if you remove all the bad games you end up with better stats. Here's another interesting result using ranks for the moment: Tannehill was ranked #12 by passer rating among all starting QB's at the end of 2016. He was also ranked #12 when you ONLY consider that "best consecutive 8 game stretch" from games 6-13 if you compare to all other QB's "best consecutive 8 game stretches".

Just pointing out what the data look like from a statistician's perspective. It's consistent with random variation. So sure one can try to explain everything by looking for an explanation, but it's also consistent with the type of random variation you generally see for most QB's across the league. Anyway.. we'll see what happens this year.

I’m not trying to remove the bad games the first 5 weeks but even his own coach said he was one of tbe most consistent players on the team during that stretch. So there is that part of it.

As for your best 8 game numbers being consistent with 12th best assuming that’s all accurate and I will give you the benefit of the doubt in that as your statistical analysis seems genuine what’s exactly wrong with that? Top 12 but trending up when he got injured I mean how is that a negative thing?

"Hoops wrote:

I’m not trying to remove the bad games the first 5 weeks but even his own coach said he was one of tbe most consistent players on the team during that stretch. So there is that part of it.

As for your best 8 game numbers being consistent with 12th best assuming that’s all accurate and I will give you the benefit of the doubt in that as your statistical analysis seems genuine what’s exactly wrong with that? Top 12 but trending up when he got injured I mean how is that a negative thing?

It's not a negative. Just wished he would have performed like that more consistently.

"cbrad wrote:

That happened sometimes yes. All that shows is we had a bad OL. That is in no way evidence that there was a higher percentage of "jailbreak" pressure than "non-jailbreak" pressure, just that our OL gave up more pressure in general (which implies more jailbreak pressure). Again, you see a lot of this "jailbreak" pressure in other games too. We're not arguing about whether we had a bad OL here.

Remember it was just 5 or 6 years ago that completion% was one of the go to stats for QBs.

Now we know air yards thrown, time to throw, the amount of separation the receiver has and what % of similar passes are caught by down and distance. In 2019 raw completion% is seen as a dinosaur stat and the higher level analysts adjust it by the factors listed above. It isn’t too hard to think that in a few years in addition to pressure% there will also data about number of rushers, number of players providing pressure, whether the pressure is from blocked or unblocked pass rushers and so on.

I know that after the 2013 season I saw you tube clips of all the sacks allowed by the Dolphins and all the sacks Cam Wake got. In 2013 Tannehill was regularly getting murdered on jailbreak sacks where the OL completely whiffed on their blocks, yet Cam Wake didn’t have a single jailbreak sack, he had to fight through a block for every single one.. so I do believe their is something to the jailbreak sack issue.

"resnor wrote:

And the league has a problem with oline play.

If the whole league has a problem with offensive line play, then it can’t possibly be a determinant of variation in quarterback play. Quarterback play would be completely independent of offensive line play, if in fact the entire league has a problem with it.

If there is no variation in offensive line play, how can it possibly cause variation across the league in quarterback play?

"Phins_to_Win wrote:

I feel like you nailed my point in the first paragraph, but I'm not sure I follow on your 2nd.

In my 3rd scenario, the OL would have given up more pressures then normal, but the strategy prevented the pressure from registering. What I'm saying is if you reduce the number of chances the defense gets at getting pressure, but of those chances they get more jailbreaks, you would have the scenario I described. For example if we give them 10 drop backs that they get to go against the O-line with no shielding, and they pressure on 9 of them but 6 of them are jailbreak styles, 9 for the course of a game isn't a lot, but 6 of them being jailbreaks is probably way above average for that scenario.

Yeah, there's value here to starting off with what we know (or is highly likely to be true) rather than what might be true. We know there is random variation in how individual players play. In practice "random" in this case means that two different players on the OL are not trying to work together to deliberately let 2 rushers through as opposed to just 1 for example. That is, we are assuming that each player is trying their best to prevent pressure independent of what that other player does, regardless of the situation he is in.

If you make that assumption you're almost forced into the argument I was making. Since both aren't working together to deliberately allow more "jailbreak" pressure than "non-jailbreak" pressure, the probabilities are independent of each other. So if there is a 50% probability of player 1 allowing pressure and a 20% probability of player 2 allowing pressure in a given situation, then the probability they both allow pressure is simply 0.5*0.2 = 0.1 = 10%. That's the "joint probability" of two independent event occurring.

Of course, different situations can change each individual's probability of allowing pressure. You could create situations where that 50% for player 1 might substantially decrease or that 20% for player 2 might increase, but as long as each is trying to prevent pressure independently, then the relative frequencies are determined by joint probability.