Ben-Jarvis Green-Ellis

[mgteich]

I agree with you entire post, but especially this part.

For me, Woodhead has been a huge plus in replacing the productivity of FAULK. We now need a starting running back, which must be Taylor. I don't think Belichick expected Green-Ellis to be the answer when the season started.

.

Luckily, they picked up Woodhead who has far exceeded anyone's expectations (since most thought he was just a gamesmanship play against the Jets) and he has stepped up in games where BJGE failed. But right now, I think the RB corp is the Pats' biggest weakness at least until Taylor comes back and shows he can be at least close to 100%.

 

[Deus Irae]

Talk about cherry picking. You're including a ton of games when he was in for very few carries, and usually in short yardage. How about this for a cherry pie: let's just take the 8 career games where BJGE had more than 10 carries. He's 132-724 yards for a 5.5ypc with 8 rushing TDs

Or maybe you can just make an attempt at explaining why he's #3 in DVOA *and* DYAR at FO this year -- you're usually so complimentary of their statistical analysis...

You apparently don't know what cherry picking is. What I did was just the opposite of that, since I was responding to those who had cherrypicked. Or, perhaps you just didn't notice that others were pointing to one Maroney game and one BJGE game this season? Then again, you want to eliminate low carry games, so talking to me about Cherrypicking is not only a failure to grasp the definition, but also hypocritical.

As for your question about his rankings, that's easy. The data has been skewed by 2 games: Buffalo and Minnesota. BJGE's production has been extremely inconsistent, with a couple of major highs on one end, and major lows on the other.

One of the weaknesses of 'averaging' stats in small(ish) samples is that there's not enough data for everything to even out.

 

[Rob0729]

C'mon, Rob...At least BJGE has produced in the regular season. Comparing him to Michael Bishop dilutes your entire position.

I only compared him to Bishop because like Bishop his legend far exceeded his production and whenever the starter played less than a great game the masses called to put him in. He isn't Bishop and far better than Biishop was (although that is faint praise), but he was always far more overhyped by a segment of the fan base than he ever produced (and that is similiar to Bishop).

 

[patsfanboy]

He did a great job in the last game but we also have to give credit for the people blocking up front.

 

[Kasmir]

You apparently don't know what cherry picking is. What I did was just the opposite of that, since I was responding to those who had cherrypicked. Or, perhaps you just didn't notice that others were pointing to one Maroney game and one BJGE game this season? Then again, you want to eliminate low carry games, so talking to me about Cherrypicking is not only a failure to grasp the definition, but also hypocritical.

Your cherry picking was partitioning the data set by variance in ypc/game, which is a patently artificial statistic you appear to have invented to make BJGE look bad. You didn't even adduce any comparables for your new stat -- how do we know it correlates with anything interesting at all? My response was intended to illustrate how easy it is to make a case by introducing artificial statistics. I'm actually perfectly content to stay with FO's DVOA and DYAR to make the objective statistical case that he's playing well.

As for your question about his rankings, that's easy. The data has been skewed by 2 games: Buffalo and Minnesota. BJGE's production has been extremely inconsistent, with a couple of major highs on one end, and major lows on the other.

One of the weaknesses of 'averaging' stats in small(ish) samples is that there's not enough data for everything to even out.

Wrong analysis. You're leaving out the Miami and Pittsburgh games where he also ran well. But much more important is that he had a lot more carries in the good games than the bad games; that's why his DVOA is so high. But his DYAR is even more impressive given that he's had so many low-carry games.

Your statistical focus on his poor performances in games where he wasn't featured is both peculiar and uncompelling. If you want to make a case that some kind of unique and importantly high variance applies here, you have a lot of work to do. Perhaps you should start a thread over at FO about this...

 

[Deus Irae]

Your cherry picking was partitioning the data set by ypc/game, which is a patently artificial statistic you appear to have invented to make BJGE look bad. You didn't even adduce any comparables for your new stat -- how do know it correlates with anything interesting at all? My response was intended to illustrate how easy it is to make a case by introducing artificial statistics.

Again, you obviously don't know what cherrypicking is. I'll leave it at that.

As for the YPC,/game, let me use some emphasis here:

I WAS RESPONDING TO PEOPLE WHO HAD USED YPC/GAME, BUT FOR ONLY ONE GAME. I EXPANDED THE YPC/GAME TO INCLUDE EVERY REGULAR SEASON GAME PLAYED BY EACH PLAYER.

If you've got a problem with the ypc/game, take it up with those two people.

Wrong analysis. You're leaving out the Miami and Pittsburgh games where he also ran well.

Nonsense. You asked a question. I gave you an answer.

16 carries at 6.13 ypc
17 carries at 6.59 ypc

Those games have skewed the DVOA and DYAR high. Since that's the question you asked, you should address any problems to yourself for your question.

But much more important is that he had a lot more carries in the good games than the bad games; that's why his DVOA is so high. But his DYAR is even more impressive given that he's had so many low-carry games.

Your focus on his poor performances in games where he wasn't featured is both peculiar and uncompelling.

Since I focused on responding to what the 2 people in the thread had been dealing with (poor performances), I'll simply note that you're complaining to the wrong poster.

 

[Kasmir]

Again, you obviously don't know what cherrypicking is. I'll leave it at that.

I'll bet you think I can't spell "ad hominem", either

As for the YPC,/game, let me use some emphasis here:

I WAS RESPONDING TO PEOPLE WHO HAD USED YPC/GAME, BUT FOR ONLY ONE GAME. I EXPANDED THE YPC/GAME TO INCLUDE EVERY REGULAR SEASON GAME PLAYED BY EACH PLAYER.

If you've got a problem with the ypc/game, take it up with those two people.

What two people? As far as I have read, only you introduced the bizarre idea that variance in YPC/game is somehow more important than YPC -- let alone DVOA or DYAR.

Nonsense. You asked a question. I gave you an answer.

16 carries at 6.13 ypc
17 carries at 6.59 ypc

Those games have skewed the DVOA and DYAR high. Since that's the question you asked, you should address any problems to yourself for your question

Again, if you really think DVOA and DYAR are so skewable by two games out of a 9 game slate, you really should scurry over to FO and inform them of this important result. Then report back. Until then, I'll trust their analysis, which I at least can be sure wasn't invented to make BJGE look good.

 

[Deus Irae]

I'll bet you think I can't spell "ad hominem", either

Hey, it's not my fault that you're calling me out for cherrypicking when I'm doing just the opposite:

...As for Maroney, he is not a great RB, but he might have done better than BJGE consider it is near impossible to run for less than 1.6 YPC like BJGE did in that game.

That would be cherrypicking

Maroney ran for 0.5 yards vs Tennessee.

This would be cherrypicking, albeit in an appropriate manner, since it's a direct response to the above comment, which is basically calling for just one example.

I simply broadened out the data, which runs counter to what's attempted when someone cherrypicks.

Again, if you really think DVOA and DYAR are so skewable by two games out of a 9 game slate, you really should scurry over to FO and inform them of this important result. Then report back. Until then, I'll trust their analysis, which I at least can be sure wasn't invented to make BJGE look good.

FO knows that small data sets are skewable. Anyone who understands "averaging" stats knows that small data sets are skewable.

To use an extreme example: if I have 9 numbers, and they are

3
4
3
3
4
3
4
3
88

The average will be 12.8, which is certainly not indicative of any norm.

 

[Kasmir]

Hey, it's not my fault that you're calling me out for cherrypicking when I'm doing just the opposite:



That would be cherrypicking



This would be cherrypicking, albeit in an appropriate manner, since it's a direct response to the above comment, which is basically calling for just one example.

I simply broadened out the data, which runs counter to what's attempted when someone cherrypicks.

If you had left your criticism at he observation that cherry picked games don't prove anything, I'd not disagree. It's your contrived analysis that I object to.

FO knows that small data sets are skewable. Anyone who understands "averaging" stats knows that small data sets are skewable.

To use an extreme example: if I have 9 numbers, and they are

3
4
3
3
4
3
4
3
88

The average will be 12.8, which is certainly not indicative of any norm.

????. DVOA is per play, not per game, so what you're saying makes no sense. Unless you're making some indirect claim about variance or sample size. FO makes attempts to address sample size. That's why Danny Woodhead isn't the leading DVOA rusher on their list. BJGE is way over that threshold (112 rushes with a cutoff of 80). If it's about the importance of variance in individual DVOA, I have no idea what you might specifically be claiming, and you really should take it up with them. If you don't trust FO's DVOA methodology, just say so, but you appear to be saying you only don't trust it wrt BJGE!

I could point out that DYAR is immune to whatever you're getting at in any case, as it's cumulative, and BJGE is #3 in both DVOA and DYAR. But I suspect we're boring everyone including the FO guys if they're unlucky enough to be reading his.

 

[Kasmir]

Hey, it's not my fault that you're calling me out for cherrypicking when I'm doing just the opposite:



That would be cherrypicking



This would be cherrypicking, albeit in an appropriate manner, since it's a direct response to the above comment, which is basically calling for just one example.

I simply broadened out the data, which runs counter to what's attempted when someone cherrypicks.

If you had left your criticism at he observation that cherry picked games don't prove anything, I'd not disagree. It's your contrived analysis that I object to.

FO knows that small data sets are skewable. Anyone who understands "averaging" stats knows that small data sets are skewable.

To use an extreme example: if I have 9 numbers, and they are

3
4
3
3
4
3
4
3
88

The average will be 12.8, which is certainly not indicative of any norm.

???? DVOA is per play, not per game, so what you're saying makes no sense. Unless you're making some indirect claim about variance or sample size. FO makes attempts to address sample size. That's why Danny Woodhead isn't the leading DVOA rusher on their list. BJGE is way over that threshold (112 rushes with a cutoff of 80). If it's about the importance of variance in individual DVOA, I have no idea what you might specifically be claiming, and you really should take it up with them. If you don't trust FO's DVOA methodology, just say so, but you appear to be saying you only don't trust it wrt BJGE!

I could point out that DYAR is immune to whatever it is you're getting at, as it's cumulative, and BJGE is #3 in both DVOA and DYAR. But I suspect we're boring everyone including the FO guys if they're unlucky enough to be reading this, and we should drop this argument.

 

[Deus Irae]

If you had left your criticism at he observation that cherry picked games don't prove anything, I'd not disagree. It's your contrived analysis that I object to.

There was nothing contrived about the analysis, as you well know. It was, in fact, responsive, basic and clearly significant in the context of the ypc/game argument.

????. DVOA is per play, not per game, so what you're saying makes no sense. Unless you're making some indirect claim about variance or sample size. FO makes attempts to address sample size. That's why Danny Woodhead isn't the leading DVOA rusher on their list. BJGE is way over that threshold (112 rushes with a cutoff of 80). If it's about the importance of variance in individual DVOA, I have no idea what you might specifically be claiming, and you really should take it up with them. If you don't trust FO's DVOA methodology, just say so, but you appear to be saying you only don't trust it wrt BJGE!

I could point out that DYAR is immune to whatever you're getting at in any case, as it's cumulative, and BJGE is #3 in both DVOA and DYAR. But I suspect we're boring everyone including the FO guys if they're unlucky enough to be reading his.

DVOA and DVAR are averaging stats, albeit not a direct average, which is gotten by applying a formula to a data group, as opposed to just a basic data stat (TDs, HRs, Goals, Points). Naturally, a skewed data group will produce a skewed result. FO does its best to limit that, but it can only work with the data it has. There's a reason that the FO stats use a minimum, after all.

 

[Triumph]

Maroney ran for 0.5 yards vs Tennessee.

Hes improving.

Most times he only makes it back to the LOS.

 

[Patspsycho]

Q: Can you talk about BenJarvus Green-Ellis and what he’s been able to accomplish lately. Has he been doing anything differently?

BB: No, I don’t think so. I think Benny’s had three years and he’s been real consistent for us. He’s gotten better at a lot of things: working the passing game and special teams, better understanding of some of our blocking schemes and the running game, more patience. But, that’s kind of the normal progression for any player – any running back. He’s tough. He makes positive yards and has good body lean. He’s been a real dependable player for us. There hasn’t been any one big thing or one day where the light came on, he’s just been a real consistent player that’s improved on a daily basis. [He] really works hard in trying to understand what he needs to do [and] what we’re trying to do, and then does it. You have to give him a lot of credit for that. He’s got a great attitude and a great work ethic. He’s tough and a very dependable player.

Extra Points - New England Patriots News and Analysis - Boston.com

End of story.

 

[buile]

So far, Green-Ellis isn't one of those great 'wow' backs. He's a hard worker, getting the job done. I think the difference between him and Maroney was the situational production, rather than pure avg yards per carry, or per game. In situations where a strong run for a first down would cement the game shut, Maroney would get shut down. Those tough yards in a short yardage situation, it seemed he often got stuffed. Green-Ellis gives the impression of getting some tough yards, getting into the endzone from a couple yards out, or getting those crucial first downs.

Do the pure stats tell that story? No, but I think the differing levels of frustration with the backs say a little something.

 

[mgteich]

??? end of story?

Belichick said nothing to contradict the analysis that Ben-Jarvis is one of the very worst starting running backs on any competitive team.

A healthy Taylor would be a definite upgrade.

Did you expect Belichick to say that he couldn't wait to see Taylor back on the field so that we could see improved production from the running back position?

Extra Points - New England Patriots News and Analysis - Boston.com

End of story.

 

[RayClay]

I think BJ has run well, for the most part. He's not a feature back and the occasional 2.0 ypc games show that. However, he's shown he can get consistent yards when the plays and blocking are with him.

I think we need to use Morris a bit and BJ needs to take it to the next level. If he's going to be the lead back on a championship team, he's got to get out of this "nice guy that tries hard" mode.

He needs to be a killer who executes crisply and decisively every play and gets every yard when in the open field. He caught a pass for about 15 where he almost dropped it om the turn, then stumbled with no one around him and barely kept his feet. Nice play? I don't know, that one might have gone to the house.

If you're going to be the man, you need to be the man. It's not a matter of talent, it's concentration, conditioning, or will but he needs to play as he has for spurts, with that fire and concentration, every play. A little Sammy to spell him might help. Got to face facts, Fred Taylor might not contribute much this year.

 

[Ring 6]

Not to take credit away from BJGE, but this is a perfect example of how 90% of the running game is the OL.
Green-Ellis runs to the hole, gets what is there and what he gets above that is veryt average. He breaks some tackles, not a ton, he has some moves, not a ton. His vision and concept of running is low risk, go where the play tells you to, don't cause a loss trying to make a big play and dont get the big play you need to risk a loss for.
I would like to see him get more carries, and have him and the OL get an opportunity to wea down a defense.
I think if you put BJGE behind a poor OL he won't look real good nor will his stats. If you put him behind a real good OL he will look good and his stats too, and if you put him behind the decent line he is behind on a team that isnt overwhelming interested in running unless they have a mismatch, you get inconsistency.
In the end, I think the deviation between BJGE and 80% of the RBs in the NFL is small, given the same OL and team. I also think his strengths and weaknesses fit our system well, at least as a runner, although I'd like to see him improve as a reciever.

 

[Patspsycho]

??? end of story?

Belichick said nothing to contradict the analysis that Ben-Jarvis is one of the very worst starting running backs on any competitive team.

A healthy Taylor would be a definite upgrade.

Did you expect Belichick to say that he couldn't wait to see Taylor back on the field so that we could see improved production from the running back position?

It is the end of the story to my discussion with rob0279 who was arguing that BJGE was very inconsistent and hence deserving of being labelled "mediocre" and "crappy."

And please spare me the crap that BJGE is one of the very worst RB- 4.2 YPC and 6 TD's are far from "worse."

 

[Kasmir]

Speaking statistically again, Football Outsiders on running back consistency as measured by their system:

The final statistic is Success Rate. This number represents the player's consistency, measured by successful running plays (the definition of success being different based on down and distance) divided by total running plays. A player with higher DVOA and a low success rate mixes long runs with downs getting stuffed at the line of scrimmage. A player with lower DVOA and a high success rate generally gets the yards needed, but doesn't often get more. Success Rate is further explained here. It is not adjusted for opponent.

FOOTBALL OUTSIDERS: Innovative Statistics, Intelligent Analysis | RUNNING BACKS 2010

BJGE is #6 in the league in success rate. His bad games have all been games where he doesn't get many carries. We might also notice that in their stats, BJGE has a very high VOA to DVOA adjustment: he's had to run against some of the better run defenses in the league. Many backs have bad days against good run defenses; BJGE suffers further because he doesn't tend to break the long runs that pull averages up in those circumstances.

Once again: BJGE hasn't had many carries in the games he did poorly in. You can argue cause and effect, but the result is that he's done really well on aggregate.

 

[ctpatsfan77]

Speaking statistically again, Football Outsiders on running back consistency as measured by their system:



FOOTBALL OUTSIDERS: Innovative Statistics, Intelligent Analysis | RUNNING BACKS 2010

BJGE is #6 in the league in success rate. His bad games have all been games where he doesn't get many carries. We might also notice that in their stats, BJGE has a very high VOA to DVOA adjustment: he's had to run against some of the better run defenses in the league. Many backs have bad days against good run defenses; BJGE suffers further because he doesn't tend to break the long runs that pull averages up in those circumstances.

Once again: BJGE hasn't had many carries in the games he did poorly in. You can argue cause and effect, but the result is that he's done really well on aggregate.

In fact, BJGE has the largest adjustment of any RB in the 80+ carries list.

No, he's not CJ2K, but I'd definitely give BJGE a high tender (i.e., one that comes with a draft pick) or extend him long-term. [Ditto for Woodhead, obviously.]