Comparing Tannehill to Luck,

"roy_miami wrote:

Not sure what your point is. Every QB with a low overall passer rating in the 15+ category would be considered not great QBs. Your Jay Cutlers, Matt Staffords and worse. Its not a coincidence.

How many of them have low ratings when their team is blowing out the other team? If your team is getting blown out, you should expect a low QB rating. Like, it's not a fair comparison, if your comparing QB A, who has had 8 15+ games, where 7 of them his team was ahead 15+, and QB B has had 16 15+ games, and 8 of them his team has been behind by 15+. One of those QBs will most likely have a much worse rating. The guy playing from behind has a much more difficult time, as they have to force passes deep to try to catch up before time runs out, while the guy up by 15+ can throw to whoever is open, regardless of yardage.

This really isn't a hard concept. I think your trying to do more with the stat than you should be. You're trying to make it something it isn't.

Yeah, the 15+ ratings depend on opponent, team and QB, so you should see correlations with each of those three.

A simple example: Rivers is generally considered one of the better QB's and he has a career 109.4 rating for 15+ games: http://www.pro-football-reference.com/players/R/RivePh00/splits/

However, last year when he played on a crappy team that went 4-12 (Rivers averaged 9.8 wins per season before that with the worst being 7 wins), his 15+ rating was 84.4: http://www.nfl.com/player/philiprivers/2506121/situationalstats

Oh, and just to show it does depend on +15 or -15, in 2015 Rivers had a 102.9 rating in his only +15 game (against the sorry Dolphins!) but 65.7 and 81.2 ratings in his two -15 games.

So that shows the effect of the team (and to a lesser degree the opponent) pretty clearly. Of course, the better the QB, the less likely you will be down by 15+ points, so it's true it's no coincidence that good QB's tend to have good 15+ ratings even when not separating by + or - 15.

"resnor wrote:

How many of them have low ratings when their team is blowing out the other team? If your team is getting blown out, you should expect a low QB rating. Like, it's not a fair comparison, if your comparing QB A, who has had 8 15+ games, where 7 of them his team was ahead 15+, and QB B has had 16 15+ games, and 8 of them his team has been behind by 15+. One of those QBs will most likely have a much worse rating. The guy playing from behind has a much more difficult time, as they have to force passes deep to try to catch up before time runs out, while the guy up by 15+ can throw to whoever is open, regardless of yardage.

This really isn't a hard concept. I think your trying to do more with the stat than you should be. You're trying to make it something it isn't.

If you want to do something else with the stats you can fill your boots, but the fact of the matter is most, if not all good QBs have a good career passer rating in combined 15+ games. Its very simple, and effective.

"cbrad wrote:

"consistent with an outlier coming back to a mean", yes, but what is the mean? As far as we know the mean is his overall record in such games, which is 20-7. Who knows if that was luck or Luck (probably both) but that's the data we have. If that was real (it was Luck), then that is more than enough reason to argue Luck contributed to the 11-5 record in an important way.

So what happened in 2014 and 2015? Well first of all, he was 4-2 in 2014 and 2-3 in 2015. 2014 fits perfectly in with the idea Luck's abilities in close games is above average. In 2015, he was 1-3 AFTER the injury, which is a good reason why one shouldn't compare 2012-2013 vs. 2014-2015, but should instead compare pre-injury vs. post-injury. Pre-injury Luck is 19-4 in one-score games which is astounding and probably reflective of something real because that's over a 3 year span.

In any case, note that the 14-2 alone in 2012-2013 could explain the 11-5 record over an 8-8 record because you have 3 extra wins per season due to that.

To me this sort of sounds like a prosecutor's fallacy.

Whilst the odds of Andrew Luck in particular randomly getting a massively positive W-L record in 0-7 point games is extremely low, there have been enough QBs playing the game over the last 80 something years that makes it almost inevitable that at least one QB would generate such a record simply by blind chance.

A QB's wins in their record in 0-7 games is going to end up very similar to their GWD stats. So GWD and record in 0-7 games aren't measuring two separate things they're measuring very similar things.

I still want to see more evidence, maybe it requires examining individual game logs, that Luck has been the cause of Indy's record.

"roy_miami wrote:

If you want to do something else with the stats you can fill your boots, but the fact of the matter is most, if not all good QBs have a good career passer rating in combined 15+ games. Its very simple, and effective.

I don't want to do anything with it. YOU brought this stat up, as if it meant something. Grouping wins and losses by 15+ together tells us nothing. Show me a QB who has a consistently good rating in games where his team is getting blown out, I think you probably have a decent QB.

All this being said, if a team is routinely getting beat by 15+, then that is a poor team, and one should further expect the QBs rating to be lower. He has dog waste around him.

Again, the stat really doesn't really show anything. I'd anything, it's more of a coincidence, at least as you're presenting. Not to mention, it starts to steer back to wins/losses being on the QB (at least as you're using it), not the team, which I highly disagree with.

"cbrad wrote:

I'd think it's the opposite. Imagine the NE Patriots playing a season against only weak college teams. How many game winning drives do you think they would have?

None!

GWD's have to occur in the 4th quarter or OT by definition.

So the fact Luck had so many GWD's is even MORE impressive given that the AFC South was so weak. He got most against non-division teams: http://www.pro-football-reference.com/play-index/comeback.cgi?player=LuckAn00

Going by their records.
Combined 98-126
5 against teams with winning records
9 against teams with losing records.
0 against 8-8 teams.

Basically in line with what I predicted about it easier to get GWDs against teams with Bad D, BAd QB or both.

Also, did you compile stats on every QB, or just a few? How many years did you go back? You compiled these stats yourself, correct?

"Pauly wrote:

Whilst the odds of Andrew Luck in particular randomly getting a massively positive W-L record in 0-7 point games is extremely low, there have been enough QBs playing the game over the last 80 something years that makes it almost inevitable that at least one QB would generate such a record simply by blind chance.

I still want to see more evidence, maybe it requires examining individual game logs, that Luck has been the cause of Indy's record.

Well, let's calculate the probability of one of these 181 QB's having made a 19-4 run (or better), going with the pre-injury stat. These QB's all satisfy certain minimum conditions so it's a good list:
http://www.pro-football-reference.com/leaders/pass_rating_career.htm

Back to the binomial theorem. Assuming the league average probability for close games is 0.5, then the probability of any single QB making a 19-4 run or better is 0.13%, and the probability of at least one of 181 QB's making a 19-4 run or better is 20.98%. So it's pretty unlikely for any QB over the entire history of the NFL to have done that (4:1 odds against).

Actually, the real odds are arguably worse than those I just calculated because the calculation assumes you just randomly choose 23 close games and that 19 of them are wins. The math doesn't include any assumption about whether those games were consecutive games. Of course we all know that if you can do something consistently for 3 years in a row it's a lot harder than if you don't.

Anyway, 21% probability at least one QB did what Luck did in NFL history.. I think that's decent statistical evidence what he did isn't a fluke.

"resnor wrote:

Also, did you compile stats on every QB, or just a few? How many years did you go back? You compiled these stats yourself, correct?

pro football reference and excel have been my friends this off season.

It isn't a systematic attempt to analyse the NFL, just digging up some interesting rabbit holes.

"Pauly wrote:

Going by their records.
Combined 98-126
5 against teams with winning records
9 against teams with losing records.
0 against 8-8 teams.

Basically in line with what I predicted about it easier to get GWDs against teams with Bad D, BAd QB or both.

Yeah, but let's be fair. For teams that are slightly worse than your team, I think you are right that it's easier to get GWD's. But for teams that are MUCH worse than yours, you are less likely to get GWD's, and the fact Luck had so few against AFC South teams (your original argument) supports that claim.

To insert some stats into the argument (for the moment just from the AFC South), from 2012-2014 Indy had 8 blowouts and 2 close (one-score) game against teams in the AFC South that won at most 4 games. But against teams in the AFC South that won at least 6 games, that stat was 2 blowouts and 6 close games. Point is, GWD's are much less likely against bottom-feeder teams.

"Pauly wrote:

pro football reference and excel have been my friends this off season.

It isn't a systematic attempt to analyse the NFL, just digging up some interesting rabbit holes.

I guess what I'm getting at, is that without looking at a number of years, for all QBs in 15+ blowouts, he has no way of knowing if his data is accurate at all.

It's actually much more suspect since he didn't do multiple years for all QBs. He might find bad QBs who put up good ratings in blowouts.

"resnor wrote:

I don't want to do anything with it. YOU brought this stat up, as if it meant something. Grouping wins and losses by 15+ together tells us nothing. Show me a QB who has a consistently good rating in games where his team is getting blown out, I think you probably have a decent QB.

All this being said, if a team is routinely getting beat by 15+, then that is a poor team, and one should further expect the QBs rating to be lower. He has dog waste around him.

Again, the stat really doesn't really show anything. I'd anything, it's more of a coincidence, at least as you're presenting. Not to mention, it starts to steer back to wins/losses being on the QB (at least as you're using it), not the team, which I highly disagree with.

It shows you nothing because you refuse to open your eyes and see anything that might show Tannehill in a negative light. If something is shown to be true a high percentage of the time most normal people would see the usefulness. Not Tannehill fans though. Its called correlation.

These are the stats for the 15+ split for all active QBs with at least 48 starts, nothing seems too out of place. Dalton might be high, Eli might be low but other than that its a believable ranking (besides Tannehill not being ranked top 5 LDO)

rank
name
wins
losses
win %
rating


1
Russell Wilson
24
0
100.0%
119


2
Tom Brady
73
11
86.9%
112


3
Philip Rivers
37
10
78.7%
109


4
Aaron Rodgers
32
9
78.0%
117


5
Ben Roethlisberger
41
12
77.4%
105


6
Andy Dalton
18
6
75.0%
96


7
Joe Flacco
27
9
75.0%
94


8
Drew Brees
51
25
67.1%
106


9
Tony Romo
30
17
63.8%
99


10
Matt Ryan
19
11
63.3%
96


11
Andrew Luck
13
8
61.9%
90


12
Cam Newton
18
12
60.0%
93


13
Alex Smith
26
20
56.5%
86


14
Michael Vick
25
20
55.6%
77


15
Matt Hasselbeck
44
42
51.2%
80


16
Carson Palmer
21
21
50.0%
84


17
Mark Sanchez
16
16
50.0%
75


18
Matthew Stafford
10
10
50.0%
84


19
Matt Cassel
24
25
49.0%
77


20
Eli Manning
32
35
47.8%
77


21
Jay Cutler
21
24
46.7%
84


22
Matt Schaub
21
25
45.7%
81


23
Ryan Fitzpatrick
16
22
42.1%
78


24
Ryan Tannehill
8
14
36.4%
82


25
Sam Bradford
5
14
26.3%
73


26
Josh McCown
7
24
22.6%
79


27
Chad Henne
3
20
13.0%
71

Dude, just stop with that garbage. Stop trying to start crap.

I mean, look at your list. Sanchez is higher than Tannehill. It's meaningless. Why 48 starts? This isn't a list where you should be excluding stuff. Look at your top 10. Almost all, if not actually all, are QBs on good TEAMS. Further, how many of the "good"ones were in more blowout losses than blowout wins?

You're so caught up on proving Tannehill isn't good, that you are blind to what I'm saying. I don't care if Tannehill is or isn't the answer. What I've always argued against, and I've said it many times, are the criticisms of Tannehill for stuff that isn't is fault, or the double standard people on here use when discussing Tannehill as compared to other QBs. Your list is meaningless unless you actually show blowout wins as compared to blowout losses.

Like I said, an hypothetical QB who is on a bad team and only has blowout losses is almost always going to look worse than a QB who is in a good team always in blowout wins. I wouldn't be in the least surprised if average QBs on a good/great team would look better than good/great QBs on bad teams, based on how your stats are compiled.

For instance, what is Tannehill rated in blowout wins? What is he rated in blowout losses? How many blowout wins has he been in? Losses? Is it easier for your number one, Wilson, to have blowout wins, as compared to Tannehill, when his defense has been holding teams to an average of 16 points for almost his entire career?

I'm not disputing your list because I think Tannehill is awesome. I'm disputing it because I think your methodology and premise are faulty.

"resnor wrote:

For instance, what is Tannehill rated in blowout wins? What is he rated in blowout losses? How many blowout wins has he been in? Losses? Is it easier for your number one, Wilson, to have blowout wins, as compared to Tannehill, when his defense has been holding teams to an average of 16 points for almost his entire career?

I'm not disputing your list because I think Tannehill is awesome. I'm disputing it because I think your methodology and premise are faulty.

You only think its faulty because Sanchez is higher than Tannehill and Wilson is at the top. Why don't you show some actual evidence as to why its faulty besides your biased opinion. Besides, if you filter for rating Tannehill would be a tier higher and above Sanchez.

No, dude. I used Sanchez as an example. I also used Wilson as an example. Your little chart doesn't factor in the team aspect of the game. Again, it's QB responsible for wins and losses, which is total bull****.

You don't like Tannehill, never have. You've created a chart, which you believe to be indefensible for those who don't have an issue with Tannehill, but the problem is, your methodology is flawed.

You absolutely need to differentiate between wins and losses. It's very telling that instead of addressing the issues I bring up, you resort to essentially calling me a Tannehill lover, and accuse me of being blind.

LMAO

"resnor wrote:

No, dude. I used Sanchez as an example. I also used Wilson as an example. Your little chart doesn't factor in the team aspect of the game. Again, it's QB responsible for wins and losses, which is total bull****.

You don't like Tannehill, never have. You've created a chart, which you believe to be indefensible for those who don't have an issue with Tannehill, but the problem is, your methodology is flawed.

You absolutely need to differentiate between wins and losses. It's very telling that instead of addressing the issues I bring up, you resort to essentially calling me a Tannehill lover, and accuse me of being blind.

LMAO

I don't understand your issues. Whenever a stat makes Tannehill look less than great its "a team stat," or "its somebody else's fault," or "I don't like those results, try a different method."

I accomplished my goal and using only one obscure stat I think I created what many would consider to be a pretty accurate QB ranking. I have no idea why you think splitting wins and losses would help me accomplish what I want to accomplish but if you think it would create a more accurate ranking then do it yourself and we can compare.

Using nothing but a silly stat sorted for passer rating it spits out this for tiers:

100+


Russell Wilson


Aaron Rodgers


Tom Brady


Philip Rivers


Drew Brees


Ben Roethlisberger

90-100


Tony Romo


Andy Dalton


Matt Ryan


Joe Flacco


Cam Newton


Andrew Luck

80-90


Alex Smith


Carson Palmer


Matthew Stafford


Jay Cutler


Ryan Tannehill


Matt Schaub


Matt Hasselbeck

70-80


Josh McCown


Ryan Fitzpatrick


Michael Vick


Matt Cassel


Eli Manning


Mark Sanchez


Sam Bradford


Chad Henne

Tannehill fan: worthless stat...

"roy_miami wrote:

I don't understand your issues. Whenever a stat makes Tannehill look less than great its "a team stat," or "its somebody else's fault," or "I don't like those results, try a different method."

I accomplished my goal and using only one obscure stat I think I created what many would consider to be a pretty accurate QB ranking. I have no idea why you think splitting wins and losses would help me accomplish what I want to accomplish but if you think it would create a more accurate ranking then do it yourself and we can compare.

I hope at this point you basically knew this before you posted your stats.....not like these guys are going to give any credence to any post that doesn't "make it okay" that Ryan has been a subpar QB thus far.

"roy_miami wrote:

I don't understand your issues. Whenever a stat makes Tannehill look less than great its "a team stat," or "its somebody else's fault," or "I don't like those results, try a different method."

I accomplished my goal and using only one obscure stat I think I created what many would consider to be a pretty accurate QB ranking. I have no idea why you think splitting wins and losses would help me accomplish what I want to accomplish but if you think it would create a more accurate ranking then do it yourself and we can compare.

Roy what you've done is pretty helpful, although it to me kind of highlights why the rating in 14+ games is kind of meaningless stat. There is a strong correlation between rating and the win%.
Also as discussed in other threads there is a positive correlation between QB rating and your game position i.e. you can play safe football with a lead and you have to play risky football when behind.

Tannehill's rating is in line with what you'd expect. Although I'd like to plot the trend line. My eyeball of the figures is that Tannehill's passer rating is about 5 -7 points higher than the hypothetical average NFL QB would have with a 36 win% in blowouts.

"Finster wrote:

I think what he's pointing out, is that is very likely the case with Luck as well, as it was with Marino, and any team with a bad defense, iow, it's not just a QB17 stat.

I didn't say nor imply it was a Thill stat. I said the 4th quarter comebacks numbers aren't really indicative of what they are being used for, because a defense can lose a game. I then gave an option that would more accurately show what people are looking for and the person I was talking too, agreed.

"roy_miami wrote:

I don't understand your issues. Whenever a stat makes Tannehill look less than great its "a team stat," or "its somebody else's fault," or "I don't like those results, try a different method."

I accomplished my goal and using only one obscure stat I think I created what many would consider to be a pretty accurate QB ranking. I have no idea why you think splitting wins and losses would help me accomplish what I want to accomplish but if you think it would create a more accurate ranking then do it yourself and we can compare.

It's.
Not.
About.
Tannehill.

For you, it may be. I don't know, I'm not inside your head. FOR ME, your chart is fundamentally flawed. I've given you my reasons already, but you continue to act like my only reason is that Tannehill is to low on the chart.

1. Does not differentiate between wins and losses.
2. Does not take into account team (the top 10 of your list are all on good/great teams).
3. Did not look at all QBs, and didn't go for multiple years.

Those are my issues with it. It has zero to do with where you rank Tannehill. I'm not going to rank anyone. You posted your made up chart, then got upset when I pointed out flaws. Frankly, of course it does what you wanted. You had a goal, then you restricted the data in certain ways to get what you wanted. Not very scientific.

"Pauly wrote:

Roy what you've done is pretty helpful, although it to me kind of highlights why the rating in 14+ games is kind of meaningless stat. There is a strong correlation between rating and the win%.
Also as discussed in other threads there is a positive correlation between QB rating and your game position i.e. you can play safe football with a lead and you have to play risky football when behind.
.

I understand all that but the bottom line is the stat produces an accurate representation of how your average non-biased guru would rank the QBs in the league today, especially if you exclude Wilson and Tannehill from the ranking.

I've seen many an "expert" say that according to their expert eyes Wilson is nothing special, as late as the first half of last season even. The stat disagreed. The stat was right. So whats more useless, the stat or the eyes?

Last offseason many had Luck in their top 5 while Wilson was generally in the 10 to 15 range. That made no sense to me and felt it should have been reversed. The stat agreed with me. The stat (and I) was right. I think the stat has Luck exactly where he should be, next to Matt Ryan, Cam Newton. I think the stat has Tannehill is in the right grouping with Cutler and Stafford.

The stat hates Eli. Incidentally I think Eli is Tannehill's ceiling, which is a player you can win a championship with (as long as he's as good as Eli in the postseason) but overall it will be a frustrating, up and down tenure.

"resnor wrote:

It's.
Not.
About.
Tannehill.

For you, it may be. I don't know, I'm not inside your head. FOR ME, your chart is fundamentally flawed. I've given you my reasons already, but you continue to act like my only reason is that Tannehill is to low on the chart.

1. Does not differentiate between wins and losses.
2. Does not take into account team (the top 10 of your list are all on good/great teams).
3. Did not look at all QBs, and didn't go for multiple years.

Those are my issues with it. It has zero to do with where you rank Tannehill. I'm not going to rank anyone. You posted your made up chart, then got upset when I pointed out flaws. Frankly, of course it does what you wanted. You had a goal, then you restricted the data in certain ways to get what you wanted. Not very scientific.

Of course all of the top QBs also coincidentally have the top teams. I forgot that in your world all QBs are equal and their play never deviates from week to week.

I've said this before and it still speaks volumes.

Ryan has very little to do with the teams W's and L's. This can be sustainable with a better supporting cast. He does not elevate his team nor does he let them down. He is just a plain ole avg guy to the definition at this point.

Adam Gase is the most important man in Ryan's life right now, if he can elevate him then he will be our QB for a long time. If he is just average ole Ryan again this year?? Time to move on.