Rushing statistics and their correlation to sucess

[Idgit]

The misuse of statistics in this scenario reminds me of the story about the 3 guys who drove over the cliff because the GPS told them to...

There is a lot of use for statistics, and I used to use it all the time.
But this situation is far too complex for a simple regression with 2 parameters.

I guess I'm unclear which statistic you think is getting misused in this case.

 

[waldoputty]

People don't have message board careers, visionary. They just get together because they're interested in a common topic and they discuss things. Or they take pot shots and leave.

But again, because this somehow always gets confused for no reason in these threads, the argument is not and has never been that 'the running game is not important.' The discussion is regarding whether or not it makes sense to spend limited resources on being able to run the ball more effectively when the team has other specific personnel limitations that actually do have a much greater impact on who wins and who loses a game that we could be addressing, instead.

Actually this was exactly the statement made: "The point is that overall rushing efficiency is virtually meaningless when it comes to whether you win or lose in the NFL."
This is silly.
It is a misuse of statistics, which I used to use often for work.

It makes sense to discuss the best allocation of resources.
Totally reasonable topic.

 

[waldoputty]

I guess I'm unclear which statistic you think is getting misused in this case.

lol, it was only used metaphorically
my point is driving off a cliff due to misuse and over-reliance on technology

 

[AdamJT13]

The point I was trying to make is that the statistics YPC is very flawed as a statistic parameter for making predictions about winning.

Who's talking about predictions?

So trying to use a flawed statistic to make a point is not meaningful.

Again, you're saying the same thing that has been said for years. Rushing efficiency is not meaningful when it comes to winning or losing.

A further point is that the ability to run (not the silly statistics) is very important for passing efficiency.

So, "the ability to run" is very important for passing efficiency, but not in any way that can be measured? Do we just take a vote to determine whether a team can run the ball well that day? Or do we just pretend that a team can run well when it suits our purpose? ("We passed well, so that means we had the ability to run." "We didn't pass well, so that means we didn't run well enough." "We didn't stop the pass, so that means the opponent ran well." "We stopped the pass, so that means the opponent couldn't run well." Never mind what actually happens in the game.)

 

[T-RO]

You have to remember that you are talking to a poster whose football message board "career" hinges on telling us how the running game is not important so don't expect any sense

There is an apt expression here:
Sometimes people don't know...what they don't know.

 

[waldoputty]

You have to remember that you are talking to a poster whose football message board "career" hinges on telling us how the running game is not important so don't expect any sense

I think the correct use of statistics is really good.
I think it is silly that posters are 'afraid' of posting about the importance of the running game due to being attacks based on flawed statistics.

 

[Idgit]

There is an apt expression here:
Sometimes people don't know...what they don't know.

Since visionary's talking about me in that quote there...what is is that that I"m supposed to not know I don't know?

If I'm going to learn something from this thread, I'm looking forward to finding out what it is. So far, it sort of feels like a rehash of what this topic always becomes: 'we're not sure why rushing the ball more effectively doesn't correlate with winning better, but we believe it that it does because, gee whiz, it sure seems like it ought to.'

 

[waldoputty]

Who's talking about predictions?

Again, you're saying the same thing that has been said for years. Rushing efficiency is not meaningful when it comes to winning or losing.

So, "the ability to run" is very important for passing efficiency, but not in any way that can be measured? Do we just take a vote to determine whether a team can run the ball well that day? Or do we just pretend that a team can run well when it suits our purpose? ("We passed well, so that means we had the ability to run." "We didn't pass well, so that means we didn't run well enough." "We didn't stop the pass, so that means the opponent ran well." "We stopped the pass, so that means the opponent couldn't run well." Never mind what actually happens in the game.)

You are only talking about rushing efficiency as a statistical parameter.
The problem is that the parameter is very flawed.
So what is the meaning of regression correlation (or lack thereof) based on the flawed parameter?
Not much.
My point is the ability to run is not reflected accurately by a statistical parameter that I am aware of for this.
May be there are some advanced stats that are better, but I am not privy to those.

Just because something cannot be tested does not make it irrelevant.
There are lots of things that science cannot answer - do we simply deny they exist?

My point is that the rushing attack situation is too complex to model.
Lets look at the anecdotal evidence, does drafting Zeke make a difference to winning?

 

[waldoputty]

Since visionary's talking about me in that quote there...what is is that that I"m supposed to not know I don't know?

If I'm going to learn something from this thread, I'm looking forward to finding out what it is. So far, it sort of feels like a rehash of what this topic always becomes: 'we're not sure why rushing the ball more effectively doesn't correlate with winning better, but we believe it that it does because, gee whiz, it sure seems like it ought to.'

I think he is talking about Adam.
I wrote the article to exactly discuss this issue - and my conclusion is that the statistic variable is no good.
So the correlation is no good because you are regressing a flawed parameter.

 

[AdamJT13]

You are only talking about rushing efficiency as a statistical parameter.
The problem is that the parameter is very flawed.
So what is the meaning of regression correlation (or lack thereof) based on the flawed parameter?
Not much.
My point is the ability to run - there is no good statistical parameter that I am aware of for this.
May be there are some advanced stats that are better, but I am not privy to those.

Just because something cannot be tested does not make it irrelevant.
There are lots of things that science cannot answer - do we simply deny they exist?

My point is that the rushing attack situation is too complex to model.

Again, you're merely agreeing that rushing efficiency is essentially meaningless when it comes to winning or losing in the NFL.

 

[waldoputty]

I guess I'm unclear which statistic you think is getting misused in this case.

One common statistic to use is yards per carry.
That parameter is extremely flawed because a lot of game variables are rolled up into it.

For example, if passing efficiency is bad, the yards per carry could be lowered as teams will just load up on the run.
Unlike the pass, the run can be quite effectively stopped by loading the box.

Thus, statistical significance of the 'ability to run' is washed away because the statistic (yards per carry) in not useful in many regression analysis.
To try to draw conclusions from a regression of yards per carry and points scored is not useful.

 

[waldoputty]

Again, you're merely agreeing that rushing efficiency is essentially meaningless when it comes to winning or losing in the NFL.

Actually I am agreeing rushing efficiency is not a useful variable for this case, and to draw conclusions of whether "running is important" based on it is a misuse of statistics.

 

[Idgit]

One common statistic to use is yards per carry.
That parameter is extremely flawed because a lot of game variables are rolled up into it.

For example, if passing efficiency is bad, the yards per carry could be lowered as teams will just load up on the run.
Unlike the pass, the run can be quite effectively stopped by loading the box.

Thus, statistical significance of the 'ability to run' is washed away because the statistic (yards per carry) in not useful in many regression analysis.
To try to draw conclusions from a regression of yards per carry and points scored is not useful.

Yah, but the argument's never been built around YPC as a measure of rushing effectiveness.

 

[waldoputty]

Yah, but the argument's never been built around YPC as a measure of rushing effectiveness.

ok, what is your measure then?

 

[AdamJT13]

Actually I am agreeing rushing efficiency is not a useful variable for this case, and to draw conclusions of whether "running is important" based on it is a misuse of statistics.

The only conclusion to be drawn is that rushing efficiency (either offensively or defensively) is not important to winning or losing. I've never argued anything but that.

If you're pretending there are other conclusions to be drawn, you're just setting up a straw man.

 

[Sydla]

It's a undeniable fact a strong ground game increases your chances at winning. Every football expert in the country will agree.If you watched games broacasted by Summerall and Madden they preached it like the gospel. .
On the other hand your defense still has to do its job.

I ask because I don't know the actual facts. But is there statistical evidence to what you claim here?

 

[T-RO]

I think the correct use of statistics is really good.
I think it is silly that posters are 'afraid' of posting about the importance of the running game due to being attacks based on flawed statistics.

The world is all about analytics. There is tremendous value in statistics and GMs, managers and coaches who have been resistant to them have been losing their jobs right and left.

Everyone who knows anything about statistics know that context is important and that there are many compound relationships. People who study statistics are looking at correlations and tendencies and probabilities. Not absolutes.

 

[waldoputty]

The only conclusion to be drawn is that rushing efficiency (either offensively or defensively) is not important to winning or losing. I've never argued anything but that.

If you're pretending there are other conclusions to be drawn, you're just setting up a straw man.

You agreed that rushing efficiency is flawed as a variable.
Yet you insist on making conclusions based on it.
Does not make sense.

Based on your statistic analysis, you probably think draft Zeke would not make a difference.
Is that right?

 

[waldoputty]

The world is all about analytics. There is tremendous value in statistics and GMs, managers and coaches who have been resistant to them have been losing their jobs right and left.

Everyone who knows anything about statistics know that context is important and that there are many compound relationships. People who study statistics are looking at correlations and tendencies and probabilities. Not absolutes.

Cowboys have analytics.
Do you think they would have drafted Zeke if they relied on a simple statistical regression of YPC vs. score?
That is a total misuse of regression analysis.
The rushing situation is far too complex to model, as least with the tools I see available.
Just because something is too hard to model accurate, it does not make it untrue.

 

[AdamJT13]

One common statistic to use is yards per carry.
That parameter is extremely flawed because a lot of game variables are rolled up into it.

Again, you can say the same thing about any stat. Your conclusion must be that all stats are flawed and meaningless.

For example, if passing efficiency is bad, the yards per carry could be lowered as teams will just load up on the run.

So wouldn't that create a correlation between bad passing and bad rushing? And wouldn't the opposite be true -- that if the pass efficiency was high, teams couldn't load up against the run, thereby increasing rushing efficiency?

Unlike the pass, the run can be quite effectively stopped by loading the box.

So you're saying there's no way to effectively stop the pass? And that using extra defensive backs isn't helpful against the pass?