Ryan Tannehill

"cbrad wrote:

Just want to point out that the article is a case study in why you need confidence intervals. Let me just repost one graph:

You think it's a coincidence that Mahomes and Jackson have those wild swings that the other QB's don't? That's small sample size right there, and I bet if they estimated confidence intervals most of them would include the league average.

Remember, win probability at any point (e.g., 15% win probability) does not occur that often, so it's sample size at each location we're talking about and it's pretty obvious just looking at the graph that the results aren't reliable for Mahomes and Jackson.

So yeah Mahomes is starting off fantastic, but let's not act like that data is a reliable indicator of Mahomes' and Jackson's performance.

In this graph they compare him to the average QB:

Do you feel like the effect size might overcome the sample size problem there, in terms of significance?

"FinFaninBuffalo wrote:

I prefer the simpler answer. Who did they sign first and who did they lock up longer.

The issue with franchising Tannehill is that his cap hit in 2020 would've been about $32M. If they would've franchised him again in 2021, his cap hit would've been roughly $38.4M. As it is they get him for the same two years with cap hits of $22.5M and $29.5M. The third year involves only $10M of dead money against the cap and $24M in cap savings if they cut him, and so parting with him then wouldn't be all that damaging, assuming they've gotten their cheap (at first) QB of the future in the draft.

Henry on the other hand plays at a far cheaper position, so they were able to franchise him for a much more meager cap hit of about $10.3M in 2020. If they franchise him again in 2021 that figure goes up to only about $12.4M.

Remember that a franchised player gets paid the average of the five largest salaries at the player's position.

"The Guy wrote:

In this graph they compare him to the average QB:

Do you feel like the effect size might overcome the sample size problem there, in terms of significance?

No because I don't think most of those data points are statistically significant, and statistical significance is a necessary condition for any difference to be "meaningful" (in science). So the effect size question wouldn't even be asked here.

As to why effect size measures exist, it's because statistical significance is necessary but not sufficient in many contexts and that's not because of small sample size but because of too large sample size (equally bad actually). You can make anything statistically significant with large enough sample size so people started to come up with effect size measures to try and deal with that. The problem is that most proposed effect size measures either themselves depend on sample size or use arbitrary cutoffs. No easy solution.

"cbrad wrote:

No because I don't think most of those data points are statistically significant, and statistical significance is a necessary condition for any difference to be "meaningful" (in science). So the effect size question wouldn't even be asked here.

As to why effect size measures exist, it's because statistical significance is necessary but not sufficient in many contexts and that's not because of small sample size but because of too large sample size (equally bad actually). You can make anything statistically significant with large enough sample size so people started to come up with effect size measures to try and deal with that. The problem is that most proposed effect size measures either themselves depend on sample size or use arbitrary cutoffs. No easy solution.

Help me understand why you don't think the data point at 10% win probability for example wouldn't be significant. Mahomes is up around 0.58 EPA per dropback, whereas the average QB is down around 0.03, meaning that about half of them are even worse than that. Is it that you think the variation among QBs is that large?

"The Guy wrote:

Help me understand why you don't think the data point at 10% win probability for example wouldn't be significant. Mahomes is up around 0.58 EPA per dropback, whereas the average QB is down around 0.03, meaning that about half of them are even worse than that. Is it that you think the variation among QBs is that large?

Well consider what the sample size there is. Even if win probability at all locations occurred equally often (which is clearly not true.. it's relatively rare to get 10% win probability compared to 50%), how many data points would you have? That's for 2 seasons, so 32 games times about 35 passes per game so 1120 total data points, divided by 80?

That's at the most optimistic 14 data points at a location (that's probably the average or so), and 14 data points is going to have a massive standard deviation. And it's probably more like 5-10 data points outside of the 40-60% range or so. I don't have the data, but my intuition for that is that few if any points will be statistically significant relative to average.

EDIT: btw.. there's one other possibility given that he only goes from 10-90%. Maybe he's using a sliding window so that each data point is really an average over an interval (if the window size is 20 that explains the 10-90%). That would solve the sample size issue and helps explain how smooth the curve is, but the drawback is you don't know anymore how much things vary at any location. The variance even for the league average data could actually be huge but not look that way if a sliding window is used. Who knows.. he doesn't tell you.

"FinFaninBuffalo wrote:

This is exactly what he has been trying to do ever since it became clear that Tannehill was going to continue his torrid pace. For the first few games he kept predicting that he wouldn't keep it up. Going back farther and he made the case for years that even with a good supporting cast, Tannehill would do no better than average. And before that, he kept harping on not performing yet.

So, in summary:

1. He isn't doing it.
2. He'll never do it.
3. He won't keep doing it.
4. He won't do it again.
5. Everybody else is responsible, so he didn't really do it.

6. He is not Patrick Mahomes.

"FinFaninBuffalo wrote:

6. He is not Patrick Mahomes.

The focus is about what the league rules encourage in terms of athleticism at the most important position, and the consequent difficulty in beating a team that may have a prototype player of that nature. Consequently players like Ryan Tannehill may have had a much better shot of winning a Super Bowl years ago than they do now.

This is no different than the difficulty somebody like Isaiah Thomas had in winning an NBA championship when Michael Jordan was in the league, being repeatedly knocked back in conference playoff rounds.

"The Guy wrote:

The focus is about what the league rules encourage in terms of athleticism at the most important position, and the consequent difficulty in beating a team that may have a prototype player of that nature. Consequently players like Ryan Tannehill may have had a much better shot of winning a Super Bowl years ago than they do now.

This is no different than the difficulty somebody like Isaiah Thomas had in winning in NBA championship when Michael Jordan was in the league, being repeatedly knocked back in conference playoff rounds.

If the production is there the style doesn't matter because it's production that correlates with win%, and 1st in passer rating = production.

And the Titans proved in 2019 that their style worked, with Tannehill as QB, at least for one year. The only question remaining is sustainability, but that's a separate question from whether the style works in this new environment favoring dual threat QB's.

It's just harder to make that style work as the league changes, but it worked for the Titans.

The reason it worked with Tannehill is due to his ability to throw pinpoint passes with velocity intermediate and deep. Why wasn't Fiedler (orc whoever the QB was) dominant with Ricky?

Fiedler threw too many interceptions... His arm was average at best... and he floated passes occasionally and they got feasted on.

"Etrius24 wrote:

Fiedler threw too many interceptions... His arm was average at best... and he floated passes occasionally and they got feasted on.

That weak arm hurt the running game. Cornerbacks could cheat to stop the run because Fiedler could not take advantage of it.

It always seemed to me ( I was still pretty young then ) That the interceptions happened at the worst possible time and killed us almost always... ( Though it has been years and I may remember it differently )

"Etrius24 wrote:

Fiedler threw too many interceptions... His arm was average at best... and he floated passes occasionally and they got feasted on.

Exactly. Having a running back going crazy doesn't mean anything unless you can torch them through the air also.

"resnor wrote:

Exactly. Having a running back going crazy doesn't mean anything unless you can torch them through the air also.

Nah Tannehill is still overrated... clearly a product of just luck:001_rolleyes:

"The Guy wrote:

The focus is about what the league rules encourage in terms of athleticism at the most important position, and the consequent difficulty in beating a team that may have a prototype player of that nature. Consequently players like Ryan Tannehill may have had a much better shot of winning a Super Bowl years ago than they do now.

This is no different than the difficulty somebody like Isaiah Thomas had in winning an NBA championship when Michael Jordan was in the league, being repeatedly knocked back in conference playoff rounds.

You make a lot of unsubstantiated claims.

"cbrad wrote:

If the production is there the style doesn't matter because it's production that correlates with win%, and 1st in passer rating = production.

And the Titans proved in 2019 that their style worked, with Tannehill as QB, at least for one year. The only question remaining is sustainability, but that's a separate question from whether the style works in this new environment favoring dual threat QB's.

It's just harder to make that style work as the league changes, but it worked for the Titans.

The league still favors balanced attacks. Fun facts, in the last 20 years:

1. No QB has ever led the league in passing yards and won the SB in the same year.
2. No QB has ever won the regular season MVP and the SB in the same year.
3. Only two QBs have led the league in passer rating and won the SB in the same year.

It is still a team game. Being lopsided (especially in volume passing stats) is a sign of some amount of offensive or defensive disfunction.

Interestingly, during the same span:

1. 5 QBs that led the league in passing yards lost the SB that year.
2. 8 QBs that won the regular season MVP lost the SB that year.
3. 3 QBs that led the league in passer rating lost the SB that year.

And, most amazingly:
4. 2 QBs that won all three lost the SB that year.

"FinFaninBuffalo wrote:

The league still favors balanced attacks.

It’s much less balanced than it used to be. Not at my computer right now but if you plot percentage of plays that are passes for each year since 1966 you see a dramatic decline in run% and a corresponding increase in pass% that’s actually still continuing. I can post a nice graph of that when I get back if you want.

Otherwise, it’s definitely the case that the trend now is towards dual threat QB’s. Just look at who’s been coming out in the last few years. We’re gradually seeing the end of the pocket passer era.

So the trend is definitely towards less balanced offenses and more dual threat QB’s. Tennessee and SF were simply teams that bucked that trend last year.

"The Guy wrote:

This article demonstrates exactly why, as I've said earlier here, Patrick Mahomes will likely be regarded as the "Michael Jordan" of the NFL before long:

You can win with Ryan Tannehill and Derrick Henry, but you'd better be able to stop this guy if you want to win a Super Bowl.

https://www.arrowheadpride.com/2020/1/15/21062877/stacking-the-box-score-patrick-mahomes-is-nfls-best-quarterback-when-trailing

I'll say it again.

Mahomes is incredibly overrated. Take Andy Reid away and he isnt a top 6 QB like you want us to have. Hes the new Donovan McNabb who happens to be playing with better offensive talent and in a passing oriented league.

Oh I think Mahomes is incredibly talented...

He does benefit from a top coach and incredible talent around him.... But he is still a special player.

"cbrad wrote:

So the trend is definitely towards less balanced offenses and more dual threat QB’s. Tennessee and SF were simply teams that bucked that trend last year.

Tannehill is a dual threat QB.

7 of the top 10 rushing teams (total yards rushing) made the playoffs in 2019. 3 of the top 10 passing teams (total yards passing) made the playoffs. 8 of the top 10 most efficient passing teams (passer rating) make the playoffs.

You should be able to run the ball and be efficient passing the ball. This pretty much describes the Titans, Ravens, 49ers, Seahawks, Vikings, and Texans.

"AGuyNamedAlex wrote:

I'll say it again.

Mahomes is incredibly overrated. Take Andy Reid away and he isnt a top 6 QB like you want us to have. Hes the new Donovan McNabb who happens to be playing with better offensive talent and in a passing oriented league.

There have been only three instances since and including 2017 in which a QB finished the season at least two z-scores above the league average in EPA per play during that period, and those have been generated by Mahomes, Brees, and Jackson. Jackson's EPA per play is largely attributable to his rushing performance in 2019, and the z-scores for Brees's other two seasons during that period were 1.43 and 0.38.

In contrast to Brees during that period, in Mahomes's other year as a starter, he was 1.92 z-scores above the league average. The next-highest z-score in the league for those three seasons is 1.70, and the next-highest one after that is 1.49. So things drop quickly from that elevated level.

So Mahomes during his first two years as a starter is essentially outperforming the league by a wide margin. If that's attributable to Andy Reid and the surrounding offensive talent, they're doing an incredible job. They're essentially getting Mahomes to be the best QB in the league over his two years as a starter.

By the way, Ryan Tannehill's EPA per play in 2019 was 0.64 z-scores above the league average from 2017 to 2019. Not even a standard deviation above it. So, certainly nothing special. His EPA per play in 2018 was 1.84 z-scores below the league average.

"FinFaninBuffalo wrote:

7 of the top 10 rushing teams (total yards rushing) made the playoffs in 2019. 3 of the top 10 passing teams (total yards passing) made the playoffs. 8 of the top 10 most efficient passing teams (passer rating) make the playoffs.

You should be able to run the ball and be efficient passing the ball. This pretty much describes the Titans, Ravens, 49ers, Seahawks, Vikings, and Texans.

Rushing yards per game shouldn't be used to measure how well a team rushes the ball because almost the entire difference in rushing yards per game between winning and losing teams is due to differences in the 4th quarter when the team with the lead tries to run out the clock (statistically speaking of course, not every game or every team). In other words the causal relation is reversed.

Try the same thing with rushing Y/A and you'll see how little rushing matters relative to passing. And also for passing you want to look at efficiency not volume as the correlation to points scored and to win% is far higher for passing efficiency than passing yards.

"FinFaninBuffalo wrote:

You make a lot of unsubstantiated claims.

Again, if what we're talking about here, with regard to any topic, rests on the credibility of a nobody on a message board, then the issue is probably so ambiguous and open to interpretation that nobody's credibility really matters anyway.

On the Kansas City Chiefs' message board they can simply call people who think Patrick Mahomes is no good delusional. The issue is so open-and-shut that any one person's credibility hardly matters.

Not so with this topic, apparently.

Guy

Stop doing this man.... It says there is a new post... I come to check it out and you post this blank crap...

Just post if you want to post... You are coming here reading and taking the time to make blank posts... Why?