Outliers, YPC, and the Cowboys running game

[big dog cowboy]

Dallas lost 5 games last year.
In 1 of them Romo was injured.
In another he was out.
In the playoff game a league rule of epic football stupidity cost us a long passing TD.

Even in a historic rushing year our limited failures were tied to Tony Romo's health and the passing game.

Teams VERY seldom pay RB big money. Those that do VERY seldom win.
Adrian Peterson has played in 4 playoff games and he is a generational RB.
Barry Sanders has often been called the best back of all-time... he played 6 playoff games, and WON ONLY 1.

Asking why teams do things is pointless because it assumes they can't make mistakes yet they obviously do.
Minnesota traded us the world for Herschel Walker and Ditka traded the world for Ricky Williams. Were those smart moves?

http://img.***BLOCKED***/albums/v628/cowboyszone/thread_stuff/applause.gif

 

[jterrell]

My advice is to beware of articles written by those that are too proud of the C- they got in "Statistics for Journalism Majors"

I'm giving Kiper the benefit of the doubt here.

Yes, take away Murray's top nine runs and his average per carry becomes 4.14.

Now do the same for those RB's with an average carry around 4.14.

Chris Ivory drops to 3.4.

Anthony Dixon drops to 2.8.

Take away every RB's top nine runs and guess what?

Murray is elite again.

but not as elite joseph randle right?

 

[jterrell]

FWIW, running backs are best judged via game tape by their own coaches.
Stats are very dependant upon other factors. Almost no two runs are exactly the same.

The coaches see each play and judge how many yards were left on the field or how many the RB stole.
Dallas had more info on Murray than anyone and offered him deals starting at 16m over years.
They liked Murray and wanted him back as their starter but were not willing to pay him like an 1800+ yard guy.

That was probably smart but we will see soon enough.
Same as we will see one years stats mean exactly zero for the next year.

 

[BlindFaith]

I watched a couple of Dolphin games early and Miller didnt stand out. I am really curious about that one. He did have a 97yrd TD, but excluding that was still pretty darn good

Have you ever analyzed the following:

http://www.footballoutsiders.com/stats/ol
http://www.footballoutsiders.com/stats/rb

I'd be interested on your take to see how valid these guys are.

 

[Doomsday101]

Nope, it was not.
NFL teams are often stupid.
In most trades one team wins and another loses.

But the Commanders were correct to identify needing a QB as primary to winning.
They just didn't go about getting one the right way or insuring it was the right QB.

One thing they do have going for them is RG3 is still a QB and he doesn't have a 4 year shelf life via position (tho he may due to injury).
So if he ever stops sucking and gets healthy he can make a very real impact on that franchise either via wins, in trade or both.
They have had plenty of success with a bottom of the draft RB.

Wonder why they aren't winning more?

Yes they got away from Morris when Washington won the division in 2012 they were handing it off to Alfred Morris and he was a main weapon for them and took the pressure off of RGIII. Much like Seahawks do with Lynch, they don't ask Russell Wilson to carry the load they use their running attack and use passes off of the running game. Again there is more than 1 way to win in the NFL. You do not have to be a top passing team. SF same way when they made it to the SB they did it off a strong running game both SF and Sea were dead last in passing attempts yet lead the league in rushing.

 

[Doomsday101]

Frankly I think there are some teams like Det, Denver and Indy that if they ran the ball better they would be much harder teams to play. Peyton and Denver want to throw it like crazy great Peyton will make the HOF and Denver will continue to come up short.

 

[AbeBeta]

So now, nothing. I'm not saying anything different here than I said previously.

And there's all sorts of stuff wrong with this example. It's not meant to be an exhaustive study of anything. It's a anecdote involving three runners and their relative position to the league average. And how does that justify your blind guess in any way? You're missing the forest for the trees if you can't understand the premise of the argument at this point. 8 plays or not, when you're talking about 400+ carries in a season, the percentage of total runs is not very significant, which is the point whether you want to accept it or not.

How does that justify my blind guess? You noted that Murray, outliers removed had a YPC that dropped about 0.6. If Murray's production is representative of the normal number of outlier runs then you would expect the league average to drop by a similar amount. The exact number is irrelevant - but what is clear is that when you take out outliers, the league average will go down a good bit. So comparing Murray's average without outliers to the league average with outliers is a garbage argument.

 

[TheDude]

Have you ever analyzed the following:

http://www.footballoutsiders.com/stats/ol
http://www.footballoutsiders.com/stats/rb

I'd be interested on your take to see how valid these guys are.

I'll take a look. My day job is basically picking apart why a model is wrong. All models are wrong, its just to what degree and is it statistically significant - but this doesnt mean they aren't useful to fill in a picture.. Bank's know that all of there mortgage will prepay at a 10-30% clip per year based on product and rate and a couple of other factors. These models are pretty accurate based on yesterdays (history) observation, but are horrendous on predicting a single loan or what underlying factor in a group of loans may jump from minor to large. .

The problem I have seen with many advanced stats and football is that there is inherently some subjectivity for which you must account. Here's where I would look at the Oline. On run blocking, while it is somewhat easy to tell where the run "went" (i.e. left , middle, right/middle, right, etc.), it is not always known if the run was supposed to go there. The RB has sight adjustments, im sure the Oline has adjustments. Its possible that these may get washed away in large numbers, but not always. For sacks, do you count the score between team, do you adjust for how long the QB held the ball, did a receiver run the wrong route? Football just appears to be too dynamic for 1 person to catalogue all of these variables without some subjectivity. Then perhaps you get different bias as no one person can watch every player, every play, every situation, etc. That said, Im not going to belittle FO or Brian Burke, I think many of their conclusions are sound and I need to refresh my understanding of their metric definition

I know many of these guys have access to good data and try to normalize, ensuring the data is most objective and clean is priority #1, then that assumptions are applied consistently, etc.

 

[Idgit]

How does that justify my blind guess? You noted that Murray, outliers removed had a YPC that dropped about 0.6. If Murray's production is representative of the normal number of outlier runs then you would expect the league average to drop by a similar amount. The exact number is irrelevant - but what is clear is that when you take out outliers, the league average will go down a good bit. So comparing Murray's average without outliers to the league average with outliers is a garbage argument.

A more reasonable assumption is that the most productive backs in the league have outliers that actually outlie, so that adjusting the other backs' totals would have less of an effect on their averages.

But that's still just not the point. The only point I've tried to make in the entire thread is that the vast bulk of even an exceptional RBs carries conform significantly to a league average for production. And because our line is far above average, there's a good chance we're going to see consistent production from our collection of RBs something like 98% of the time.

Again, yes, I know this is anecdotal and not necessarily scientific. I posted it because it gives context to how an NFL running game actually behaves, and that context is different from how it's often perceived by a fan base that's worried about how defenses are going to change they way they play us with Murray out of the lineup.

 

[Hoofbite]

Frankly I think there are some teams like Det, Denver and Indy that if they ran the ball better they would be much harder teams to play. Peyton and Denver want to throw it like crazy great Peyton will make the HOF and Denver will continue to come up short.

Detroit probably, but I don't think Denver is really in the same situation. Detroit has Stafford who's capable of being very good, but more often than not he's just above average.

 

[Doomsday101]

Detroit probably, but I don't think Denver is really in the same situation. Detroit has Stafford who's capable of being very good, but more often than not he's just above average.

I think running game could help a lot of teams just as it did Aikman. I think Rodgers is the best QB in the NFL I also think GB being a bit more balanced could put together consecutive SB. I have never stated passing was not important I think it is very important I also believe teams who can also run effectively puts defense in a very bad position.

 

[AbeBeta]

A more reasonable assumption is that the most productive backs in the league have outliers that actually outlie, so that adjusting the other backs' totals would have less of an effect on their averages.

But that's still just not the point. The only point I've tried to make in the entire thread is that the vast bulk of even an exceptional RBs carries conform significantly to a league average for production. And because our line is far above average, there's a good chance we're going to see consistent production from our collection of RBs something like 98% of the time.

Again, yes, I know this is anecdotal and not necessarily scientific. I posted it because it gives context to how an NFL running game actually behaves, and that context is different from how it's often perceived by a fan base that's worried about how defenses are going to change they way they play us with Murray out of the lineup.

Really?

Is Joseph Randle one of the most productive backs in the league? If you take out his thee outlier runs, the team's average goes from 4.63 YPC to 4.38 YPC. Poor old Joe's average plummets from 6.7 down to 4.2. I didn't even have to look outside our team for an example of how other backs are going to show some huge drops when you take out the extreme runs.

The idea that other backs' totals would have less an effect on their average and the league average in general just doesn't have any basis in reality.

 

[jterrell]

Yes they got away from Morris when Washington won the division in 2012 they were handing it off to Alfred Morris and he was a main weapon for them and took the pressure off of RGIII. Much like Seahawks do with Lynch, they don't ask Russell Wilson to carry the load they use their running attack and use passes off of the running game. Again there is more than 1 way to win in the NFL. You do not have to be a top passing team. SF same way when they made it to the SB they did it off a strong running game both SF and Sea were dead last in passing attempts yet lead the league in rushing.

They led the league in rushing by QBs too. Those QB rushing yards were key to their success because many came in passing sets and situations off pass plays.
That is effectively part of the passing game because you have to account for the QB.
And as RG3 showed that spread option attack is a double edged sword because it puts the QB at greater risk for injury.

The Commanders didn't get away from Morris so much as they couldn't run their offense when RG3 got injured.
And minus the spread option looks Morris has fallen off in performance.
Morris has averaged more carries across 3 seasons than anyone in football.
He's just seen his ypc drop from 4.8 to 4.6 to a very pedestrian 4.1.

 

[FuzzyLumpkins]

Really?

Is Joseph Randle one of the most productive backs in the league? If you take out his thee outlier runs, the team's average goes from 4.63 YPC to 4.38 YPC. Poor old Joe's average plummets from 6.7 down to 4.2. I didn't even have to look outside our team for an example of how other backs are going to show some huge drops when you take out the extreme runs.

The idea that other backs' totals would have less an effect on their average and the league average in general just doesn't have any basis in reality.

Now making lots of big plays is a negative.

 

[jterrell]

I think running game could help a lot of teams just as it did Aikman. I think Rodgers is the best QB in the NFL I also think GB being a bit more balanced could put together consecutive SB. I have never stated passing was not important I think it is very important I also believe teams who can also run effectively puts defense in a very bad position.

huh? Green Bay has already won Super Bowls with less effective rushing than they get from Lacy.
They were 11th in the league in rushing last year.

 

[Doomsday101]

They led the league in rushing by QBs too. Those QB rushing yards were key to their success because many came in passing sets and situations off pass plays.
That is effectively part of the passing game because you have to account for the QB.
And as RG3 showed that spread option attack is a double edged sword because it puts the QB at greater risk for injury.

The Commanders didn't get away from Morris so much as they couldn't run their offense when RG3 got injured.
And minus the spread option looks Morris has fallen off in performance.
Morris has averaged more carries across 3 seasons than anyone in football.
He's just seen his ypc drop from 4.8 to 4.6 to a very pedestrian 4.1.

Running game was big for Washington you can deny it all you want but fact is they ran him more and they won. I'm not against the pass I think passing is important has been since the pass cam into effect but this notion the run game means nothing and has nothing to do with winning is flat out stupid and I have heard pro coaches say how important that run game in including Jason Garrett.

For those who think passing wins game running is nothing but a waste of time fine think as you choose but I have seen Dallas try to be a passing team and we down right sucked. I have seen us run a balanced attack even with a bad defense and win and not just vs bad team but vs top teams. It is not that hard to figure out the run matters

 

[Doomsday101]

huh? Green Bay has already won Super Bowls with less effective rushing than they get from Lacy.
They were 11th in the league in rushing last year.

Yes GB won a SB they have the best QB and I think if they could run better then they would have more than 1. Aikman was great just as well and has 3 an Dallas running game played a big part. Troy may not have been putting up the same numbers as Young, Favre and Kelly but we won championships.

 

[Hoofbite]

I think running game could help a lot of teams just as it did Aikman. I think Rodgers is the best QB in the NFL I also think GB being a bit more balanced could put together consecutive SB. I have never stated passing was not important I think it is very important I also believe teams who can also run effectively puts defense in a very bad position.

Sure, it can help. The issue with Denver is that teams are already in a bad position when they play them because they're going to score 30 points. Under just about any circumstance 30 points is good enough to win. A better running game may help them, but maybe the reason they're falling short is something other than their offensive production.

I don't think a better running game necessarily does much for them because they should already be winning with what they have. Win by a little more maybe? Control the clock better? They could already do that with what they have. They had 8 fewer rushing first downs than Dallas did last year, and they were 6th overall in total first downs. In terms of TOP, they were 13th. They could control the clock to a greater extent if they wanted.

 

[AbeBeta]

Now making lots of big plays is a negative.

Says the OP - that's what I was arguing against - read the whole exchange

 

[FuzzyLumpkins]

Says the OP - that's what I was arguing against - read the whole exchange

No he was trying to get a context of what a RB accomplished down in and down out. What you present here is a gross oversimplification.