Making NFL history

[Bob Sacamano]

Wrong.

The coaches who lost those games in the AFC were Marty Schottenheimer, Bill Cowher, Mike Shanahan, Jeff Fisher, and Tony Dungy.

Not a bad group of coaches.

Homer to the end.

 

[windward]

Homer to the end.

8-8= homer?

Doubt it.

 

[Bob Sacamano]

8-8= homer?

Doubt it.

Whatever, homerward.

 

[AdamJT13]

I think the stats verify what we all know. We move the ball, commit dumb penalties, bad turnovers in bad times and don't create turnovers.

They report effects, not cause. The question is "Why does this team commit more penalties?" and "Why do we not get turnovers?" This is likely the most subjective answer that gets (caoching as a catch all).

If you ignore stats, you wouldn't know that the team commits more penalties than it should or doesn't get as many turnovers as it should. And if you think they're meaningless, then you wouldn't pay attention to the stats about penalties and turnovers anyway.

The stats quantify what happens on the field. And they report both effects and cause, because one stat can affect another, which affects another, which affects another. And collectively, they cause us to lose.

 

[JBS]

If you ignore stats, you wouldn't know that the team commits more penalties than it should or doesn't get as many turnovers as it should. And if you think they're meaningless, then you wouldn't pay attention to the stats about penalties and turnovers anyway.

The stats quantify what happens on the field. And they report both effects and cause, because one stat can affect another, which affects another, which affects another. And collectively, they cause us to lose.

With all of this being said, what is your take on how we should approach fixing it?

 

[AdamJT13]

Great thread Adam. I have a question for you since so many of the participants in this thread see this as further evidence of our poor coaching.

Has our coaching staff contributed to that 118 per game yard difference?

Or is that all a function of our superior talent?

And only the coaches are responsible for the penalty difference and poor special teams play?

If you want my opinion, the coaches have a hand in almost everything that happens, both good and bad. And when players don't do something properly, the coaches have a way to remedy it. But it's never ALL coaching or ALL talent.

 

[sonnyboy]

If you want my opinion, the coaches have a hand in almost everything that happens, both good and bad. And when players don't do something properly, the coaches have a way to remedy it. But it's never ALL coaching or ALL talent.

Figured you'd say his. Was obviously looking to make a point to the rest in the thread.

 

[TheDude]

If you ignore stats, you wouldn't know that the team commits more penalties than it should or doesn't get as many turnovers as it should. And if you think they're meaningless, then you wouldn't pay attention to the stats about penalties and turnovers anyway.

The stats quantify what happens on the field. And they report both effects and cause, because one stat can affect another, which affects another, which affects another. And collectively, they cause us to lose.

I agree stats are valuable. That is what I do as a real gig. But You have to define all independent variables for the dependent variables (i.e. wins).

As you may be able to prove by yds against and yds for, Dallas is an outlier (reject the null). meaning with an alpha of 5%, we can say that with 95% certainty that these 2 variables are not influencing wins. We can prove it but it is empirical evidence to.

So you can figure in penalties (which you did a good job). but if penalties do not explainit all, it could be net penalty yds, lost yds, missed fgs, plus special teams, 3rd down success, etc.

Ultimately, you could come up with a multiple regression of significant variables that may show yds have nothing to do with it when you factor in others.

you may end up with a mx regression parametric equation: y=b+m1x1+m2x2+...mixi.

i thought about trying all this, but I don't have the time.

Even when you get to the solution, it is a predictive model, but its based on historical data. you can get a good idea, but why Barron IMMEDIATELY grabsand holds on the last play of the game could throw all of the other varaibales away (assuming it is the only penalty of the game).

Stats will get you some insight, but some times its obvious without the work, sometimes you lose man to man battles and it is a domino effect.

I think you have done a good job, but there is always a random variability.

I measure refinance/prepayment behavior. you think that lower rates and low refinance costs would entice borrowers to refi, but there are things like appraisal, and burnout that make it an inexact science. Stats can forecast to the best of the current data, but it cannot predict causes of extreme future events.

The questions is can the outliers produce new data to use as a dependent variable.

 

[tupperware]

I agree stats are valuable. That is what I do as a real gig. But You have to define all independent variables for the dependent variables (i.e. wins).

As you may be able to prove by yds against and yds for, Dallas is an outlier (reject the null). meaning with an alpha of 5%, we can say that with 95% certainty that these 2 variables are not influencing wins. We can prove it but it is empirical evidence to.

So you can figure in penalties (which you did a good job). but if penalties do not explainit all, it could be net penalty yds, lost yds, missed fgs, plus special teams, 3rd down success, etc.

Ultimately, you could come up with a multiple regression of significant variables that may show yds have nothing to do with it when you factor in others.

you may end up with a mx regression parametric equation: y=b+m1x1+m2x2+...mixi.

i thought about trying all this, but I don't have the time.

Even when you get to the solution, it is a predictive model, but its based on historical data. you can get a good idea, but why Barron IMMEDIATELY grabsand holds on the last play of the game could throw all of the other varaibales away (assuming it is the only penalty of the game).

Stats will get you some insight, but some times its obvious without the work, sometimes you lose man to man battles and it is a domino effect.

I think you have done a good job, but there is always a random variability.

I measure refinance/prepayment behavior. you think that lower rates and low refinance costs would entice borrowers to refi, but there are things like appraisal, and burnout that make it an inexact science. Stats can forecast to the best of the current data, but it cannot predict causes of extreme future events.

The questions is can the outliers produce new data to use as a dependent variable.

This whole thing kind of reminds me of this:

[youtube]PYhM5m5ZbJg[/youtube]

 

[TheDude]

This whole thing kind of reminds me of this:

[youtube]PYhM5m5ZbJg[/youtube]

same concept in many ways. just think of it as each bank thinks it finds one better variable to price a security that is different from the previous. If this proves correct, investors leverage to be part of the new better predictive model.

What tends to happen is that the PHds and Quants get caught up in formula/stats and forget about common sense.

Don't think housing always go up, dont think yds and penalties are the entire cause.

sometimes it is coaching, talent, random corrective action.

 

[tupperware]

Yeah the part about using things you can quantify reminded me the most of it. There are various reasons why a certain individual may underperform on any given day that has nothing to do with his typical level of play. It probably averages out over time but has the possibility to deviate from the norm at any given time. Which could mean Jenkins might normally shut WR X down, but on a particular Sunday he might have the flu and negatively impact his own play and statistics, which isn't tracked at all by the statistics. I mean the things that cause the dropoffs aren't tracked.

 

[theogt]

Wrong.

The coaches who lost those games in the AFC were Marty Schottenheimer, Bill Cowher, Mike Shanahan, Jeff Fisher, and Tony Dungy.

Not a bad group of coaches.

Eh...I wouldn't hire any of them.

 

[TheDude]

Yeah the part about using things you can quantify reminded me the most of it. There are various reasons why a certain individual may underperform on any given day that has nothing to do with his typical level of play. It probably averages out over time but has the possibility to deviate from the norm at any given time. Which could mean Jenkins might normally shut WR X down, but on a particular Sunday he might have the flu and negatively impact his own play and statistics, which isn't tracked at all by the statistics.

Exactly, or take injuries. They play a huge roll. if Jenkins is hurt, that will affect his play. Maybe you lose Phillips and you feel it in the running game and passing game. Perhaps he protects better and lets romo have more time to throw 9i.e. Romo didn't get sacked either way, but maybe bennet gets him hit to miss a throw for a TD and Phillips lets him have time to complete it.)

That's why i didn't bother doing a full in depth study because there are so many more moving parts than say in Baseball. that is a perfect sport to measure.

you can do similar in football, but the confounding variables are vast and subjective

 

[Don Corleone]

Teams that get no turnovers, have bad penalties, and no kick return yardage have to drive the ball further for points, and generally rack up lots of yards.

On the flip side, we are quite generous in all of the above categories, so our opponents get a short field.

...we then lose.

 

[tupperware]

Exactly, or take injuries. They play a huge roll. if Jenkins is hurt, that will affect his play. Maybe you lose Phillips and you feel it in the running game and passing game. Perhaps he protects better and lets romo have more time to throw 9i.e. Romo didn't get sacked either way, but maybe bennet gets him hit to miss a throw for a TD and Phillips lets him have time to complete it.)

That's why i didn't bother doing a full in depth study because there are so many more moving parts than say in Baseball. that is a perfect sport to measure.

you can do similar in football, but the confounding variables are vast and subjective

Wouldn't you have to score like every individual linemens average play, then score the oppositions play across their games. I'd guess you'd have to come up with a system for scoring every position, really. How much is getting beat worth, were you beat on grass or turf at home or on the road and other variables I'm just not thinking about. Then, in order to accurately score a QB's play on any Sunday you'd need to score the entire offensive line and every position to see if they were in line with their typical performances and adjust them accordingly up or down considering opposition. Then after doing all of that, you can grade out your QB's play for the day.

You actually should do it, it'd be interesting

 

[TheDude]

Teams that get no turnovers, have bad penalties, and no kick return yardage have to drive the ball further for points, and generally rack up lots of yards.

On the flip side, we are quite generous in all of the above categories, so our opponents get a short field.

...we then lose.

Sound assessment. It's pretty apparent.

You can test that in stats, but what it won't tell you is what is causing each individual issue. Personell, technique, consiency, mistake, etc.

What I do know is that this team doesn't get enough turnovers. I think that is culture,scheme and coaching. You just don't see the Cowboys actively go for the ball like the Jets and steelers. plus they tackle with better technique.

You can measure Pitts turnovers vs us, but why do they get more. Just watch them play

 

[tupperware]

Sound assessment. It's pretty apparent.

You can test that in stats, but what it won't tell you is what is causing each individual issue. Personell, technique, consiency, mistake, etc.

What I do know is that this team doesn't get enough turnovers. I think that is culture,scheme and coaching. You just don't see the Cowboys actively go for the ball like the Jets and steelers. plus they tackle with better technique.

You can measure Pitts turnovers vs us, but why do they get more. Just watch them play

They play confident, we don't. There are times guys in our secondary have to direct each other on where to go.

 

[TheDude]

Wouldn't you have to score like every individual linemens average play, then score the oppositions play across their games. I'd guess you'd have to come up with a system for scoring every position, really. How much is getting beat worth, were you beat on grass or turf at home or on the road and other variables I'm just not thinking about. Then, in order to accurately score a QB's play on any Sunday you'd need to score the entire offensive line and every position to see if they were in line with their typical performances and adjust them accordingly up or down considering opposition. Then after doing all of that, you can grade out your QB's play for the day.

You actually should do it, it'd be interesting

It would be interesting, but i don't know where to get all that data. I had a coworker want to start a business that recorded all of this to play the line's in vegas. he had 15 years of data for every play of each team and team match up. but it was a maintenance nightmare. Injuries, coaching changes, weighting schemas, etc.

You can run this through Matlab and get a predictive model, but it still can't capture strange decisions like Felix getting no tuches in Washington in 08 and few in 2010. It is real hard compared to baseball.

 

[TheDude]

I think all of adam's data is great, it is real interesting to statisitcally show how ridiculous the team is perfoming. It does indicate that the team is killing itself to a degree, but ultimately I am curious about the subjective.

I think some of this is talent related. i think romo and Miles and Dez and RW11 are putting up yards despite a lack of talent on the oline.

I think the defense is a sound defense, but it is not an imposing defense, which it should be with Ware, Spencer, Newman, Jenkins, Ratliff, etc. Maybe that is completely safety related, but i doubt it - there just aren't any big hitters there.

I don't know what to think about the ST. If Beuhler can't get TBs and do FGs, he is failry worthless as he isn't accurate enaough to be a FT FG kicker now. That is a coaching decision, which leads me down my biggest issue with this team.

can't prove it with math though

 

[sonnyboy]

Teams that get no turnovers, have bad penalties, and no kick return yardage have to drive the ball further for points, and generally rack up lots of yards.

On the flip side, we are quite generous in all of the above categories, so our opponents get a short field.

...we then lose.

Fantastic point. And it is a contributin factor. But nowhere close to the difference in net yards from scrimage.

Special teams hurt us against the Viks and Titans, but not so much to make up the net difference in yards from scrimage.

We clearly outplayed the Vikings at the LOS. 3 sacks and twice as many hits on Farve while not surrendering one sack. They had pressure, but I don't think one fan thought we had a shot to get out of Min without giving up more than one sack. Even in a win.