Some insight into why Philbin was fired and Gase was hired.

[Rich]

And that's why you have to know how to evaluate statistics, so you can know when that's happening, which is precisely what I'm doing here with regard to the PFF subjective data.

The only way to know when the stats are lying is to watch the film, which is exactly what Football Outsiders and PFF do.

[Dolfanalyst]

The only way to know when the stats are lying is to watch the film, which is exactly what Football Outsiders and PFF do.

And how do we know when one's interpretations of the film are lying?

[Rich]

And how do we know when one's interpretations of the film are lying?

We don't because everyone sees something different from the same film. That's the beauty of it.

[Dolfanalyst]

The assigning of the value 1 point to a gain of 45% of the needed yards on first down is subjective.

It's just as subjective as saying a person with a body mass index of 30 or more (an obese person) is at-risk for diabetes.  The application of that assessment of risk to that single individual is indeed subjective, but it's based on objective research.  Likewise, the application of the values you pointed out above are based on similar probabilities of obtaining a first down.

Yards are objective.
Points are objective.
Assigning weighted values to the "success" of a play (outside of the yards it gained/lost, or points it scored) is quite obviously subjective, and it's absurd to argue otherwise.

And when it's done the same for every player, on every team, in every such situation, the results permit objective comparisons among teams.

When it's done differently for every player, on every team, in every situation (as is perhaps the case with PFF subjective grades), we have no idea how objective the comparisons are on the basis of them.

[Dolfanalyst]

We don't because everyone sees something different from the same film. That's the beauty of it.

Well when you're trying to determine the validity of those interpretations of film, you can hardly do that when people are making various interpretations.  There is no validity without reliability.

If your bathroom scale tells you something different every time you step on it, you can hardly call it valid.

[Rich]

Well when you're trying to determine the validity of those interpretations of film, you can hardly do that when people are making various interpretations.  There is no validity without reliability.

If your bathroom scale tells you something different every time you step on it, you can hardly call it valid.

That was a piss poor analogy. Wow.

[Spider-Dan]

It's just as subjective as saying a person with a body mass index of 30 or more (an obese person) is at-risk for diabetes.

So in other words, it's subjective.

Cameron Wake is listed at 6'3", 269 lbs, which gives him a BMI of 33.6.  Does that mean he is obese?  Is he at-risk for diabetes?

Subjective stats don't magically become objective because you really, really like them, nor do they become objective when you apply them to a lot of people.  It would seem that you are aware of this, as you smoothly inserted (without evidence) a claim that PFF does not apply their subjective ratings equally to all players.

Because if they do, it kind of detonates your entire argument; at that point, any stat that anyone can invent is "objective" as long as it is always applied evenly.

[Dolfanalyst]

Because if they do, it kind of detonates your entire argument; at that point, any stat that anyone can invent is "objective" as long as it is always applied evenly.

...and it correlates with winning, which determines its validity.

That's the problem with PFF's grades.  They don't do that nearly as well as DVOA.

[Spider-Dan]

...and it correlates with winning, which determines its validity.

That's the problem with PFF's grades.  They don't do that nearly as well as DVOA.

So essentially, subjective stats are actually great, but only if they are accurate, or something?

Again, you're only making the argument that the subjective stats you like are better than the ones that you don't.  I have no interest in that discussion; that's like, your opinion, man.  If you want to claim the moral high ground on objective statistical analysis, then stick to objective statistics, not Really Awesome Subjective Statistics.

Given your distaste for subjective stats, I look forward to your continued analysis of Tannehill's performance... using only objective stats, of course.

[Dolfanalyst]

So essentially, subjective stats are actually great, but only if they are accurate, or something?

Uh, yes, if they're valid?  Should the statistic not measure what it's supposed to measure?

Again, you're only making the argument that the subjective stats you like are better than the ones that you don't.  I have no interest in that discussion; that's like, your opinion, man.  If you want to claim the moral high ground on objective statistical analysis, then stick to objective statistics, not Really Awesome Subjective Statistics.

Given your distaste for subjective stats, I look forward to your continued analysis of Tannehill's performance... using only objective stats, of course.

I use the statistics that are most strongly correlated with winning, which gives them their validity.  Those happen to be the objective ones at this point in time.

[Spider-Dan]

I have removed the off-topic personal attacks from the thread.  Stay on topic or a lock is inbound.

[Dolfanalyst]

I have removed the off-topic personal attacks from the thread.  Stay on topic or a lock is inbound.

Thank you.  Much appreciated.  Seriously.

[Spider-Dan]

Uh, yes, if they're valid?  Should the statistic not measure what it's supposed to measure?

So you're fine with subjective stats, as long as you feel they are valid?

If we are to correlate PFF's QB evaluation with winning... then why even have advanced stats in the first place?

The best QB is Cam Newton, who won 15 games.
The second-best QB is Carson Palmer, who won 13 games.
The third-best QB is Tom Brady, who won 12 games.

Do we want to know who the best defense was last year?  Why, it was the Carolina Panthers, who won 15 games.  The second-best defense was the Arizona Cardinals, who won 13 games.

Who was the best kicker?  The best tight-end?  etc.

The stats I just gave are 100% correlated with winning, so they are automatically better than any other stat you can provide, right?

[Rich]

So you're fine with subjective stats, as long as you feel they are valid?

If we are to correlate PFF's QB evaluation with winning... then why even have advanced stats in the first place?

The best QB is Cam Newton, who won 15 games.
The second-best QB is Carson Palmer, who won 13 games.
The third-best QB is Tom Brady, who won 12 games.

Do we want to know who the best defense was last year?  Why, it was the Carolina Panthers, who won 15 games.  The second-best defense was the Arizona Cardinals, who won 13 games.

Who was the best kicker?  The best tight-end?  etc.

The stats I just gave are 100% correlated with winning, so they are automatically better than any other stat you can provide, right?

Ryan Tannehill is a better QB than Joe Flacco and Phillip Rivers. He's just as good as Eli Manning.

[Dolfanalyst]

So you're fine with subjective stats, as long as you feel they are valid?

If we are to correlate PFF's QB evaluation with winning... then why even have advanced stats in the first place?

The best QB is Cam Newton, who won 15 games.
The second-best QB is Carson Palmer, who won 13 games.
The third-best QB is Tom Brady, who won 12 games.

Do we want to know who the best defense was last year?  Why, it was the Carolina Panthers, who won 15 games.  The second-best defense was the Arizona Cardinals, who won 13 games.

Who was the best kicker?  The best tight-end?  etc.

The stats I just gave are 100% correlated with winning, so they are automatically better than any other stat you can provide, right?

I'm sure you've heard of the term "straw man"?