Is there a RB stat for this (Shourright or any other stat guys?)

[MartinTheSNAKE]

Im curious if there is a stat recorded on what a RB's average is on his carries. I dont mean his YPC...

For example.. Lamar Miller's YPC is 4.0 ... But is there a % recorded stat on what he normally gets per attempt? Miller doesnt get a lot of 3-4 yard carries.. Hes more of a 1, 1, 2, -2, 1, 15, -3, -1, 1, 2, 30 type of running back... What im looking for is a stat that tells what amount of yards is he likely to get per attempt. Am i making sense?


If he runs 10 times... runs it for 1 yard (2 times) 2 yards (5 times) -3 (1 time) 5 yards (2 times) For that span he would average 2 yards 50% of the time. 2 yards 20% of the time. 5 yards 20% of the time. -3 10% of the time.

Is there a recorded stat for something like this?

If not, i think this would be a pretty good stat to keep. Would show the consistency of a RB better then his YPC would be.

 

[traptses]

I don't get it. If he averages 4.4 ypc, he's likely to get 4.4 yards per attempt. That's just the way averages work. Maybe what you can do, is look at all his carries. Remove the top 5% longest carries and the lowest 5% carries, and then do the average of the middle 90%

 

[Ruckus45]

You mean more of what he will get as his mode if you broke it out in some sort of obviously skewed bell curve? So if he makes 10 1 yard runs then 1 65 yard run you want to know if he is more likely for the 1 yard run?

Not sure who would carry that stat. You'd have to log every run and outcome. Who am I kidding, someone prob has that.

 

[MartinTheSNAKE]

You mean more of what he will get as his mode if you broke it out in some sort of obviously skewed bell curve? So if he makes 10 1 yard runs then 1 65 yard run you want to know if he is more likely for the 1 yard run?

Not sure who would carry that stat. You'd have to log every run and outcome. Who am I kidding, someone prob has that.

Exactly. 11 carries 75 yards look great on the stat sheet and show and avg ypc of almost 7... But in all actuality that stat is very misleading of that RB performance.

 

[Shouright]

You would need to do some mathematical work to determine what you're looking for here, but the data with which to do it are on this page:

http://www.pro-football-reference.c...margin_max=&order_by=yards&rusher_id=MillLa01

 

[Slim]

You want to 'bucketize' carries. It's an interesting idea. Barry Sanders ran a lot of negative runs, but also broke big ones. His big ones skewed his average, and offset the negatives. You want a stat that exposes that.

Mode and Median would be of interest to you. Suspect a guy like Barry's median would be much lower than his average. Guys with a median/average close together are guys who consistently gain 3-4 but never break the big one.

 

[MartinTheSNAKE]

You want to 'bucketize' carries. It's an interesting idea. Barry Sanders ran a lot of negative runs, but also broke big ones. His big ones skewed his average, and offset the negatives. You want a stat that exposes that.

Mode and Median would be of interest to you. Suspect a guy like Barry's median would be much lower than his average. Guys with a median/average close together are guys who consistently gain 3-4 but never break the big one.

Pretty much. Was hoping they had a stat already for this but i guess not. I dont have the time or will to actually go through and keep this stat myself haha. But yea, i think this would be a very usefull stat to know.

 

[PhinMagic]

Best and fastest thing to do would probably look at the long of the day, remove that and calculate his YPC. like 10 carries for 60 yards looks great but if he had a 30 yard long, then it's more like 9/30, which cuts the average in half.

 

[MartinTheSNAKE]

Best and fastest thing to do would probably look at the long of the day, remove that and calculate his YPC. like 10 carries for 60 yards looks great but if he had a 30 yard long, then it's more like 9/30, which cuts the average in half.

Yea but im not really looking for 1 number... more of a % of what the end result of his runs are.

1 yard (15%) 2 yards (8 %) - yards (8%) 10 yards (4%) etc.

 

[tay0365]

Yea but im not really looking for 1 number... more of a % of what the end result of his runs are.

1 yard (15%) 2 yards (8 %) - yards (8%) 10 yards (4%) etc.

I took a stab at it with the website shouright gave us, as I also became curious.

I broke it up as 8 categories....-0 (Every time caught behind the line of scrimmage), 0 (No gain), 1-2 yards(Considered ineffective run), 3-5 yards (Somewhat effective), 6-10 (Effective run), 11-20(good run), 20 and over (Great run), and Fumble.

------------143 attempts---------

-0 (Behind line of scrimmage)............22 times (15.38% of the time)

0 yards (No gain)..............................14 times (9.79% of the time)

1-2 yards (ineffective run).................33 times (23.07% of the time)

3-5 yards (somewhat effective run)...38 times (26.50% of times)

6-10 yards (effective run).................24 times (16.78% of the time)

11-20 yards (good run)....................7 times (4.8% of the time)

20 and more (Great run)....................4 times (2.70% of the time)

Fumble ...........................................1 time ( .06% of the time)

-----------effectiveness-----------

-0 to 2 yards (Bad run)...........69 times (48.25% of the time)

3 to 10 yards (good run)...........62 times (43.35% of the time)

11 to 20+ (Great run).......11 times (7.69% of the time)

Fumbled....................................1 time (.06% of the time)

 

[MartinTheSNAKE]

I took a stab at it with the website shouright gave us, as I also became curious.

I broke it up as 8 categories....-0 (Every time caught behind the line of scrimmage), 0 (No gain), 1-2 yards(Considered ineffective run), 3-5 yards (Somewhat effective), 6-10 (Effective run), 11-20(good run), 20 and over (Great run), and Fumble.

------------143 attempts---------

-0 (Behind line of scrimmage)............22 times (15.38% of the time)

0 yards (No gain)..............................14 times (9.79% of the time)

1-2 yards (ineffective run).................33 times (23.07% of the time)

3-5 yards (somewhat effective run)...38 times (26.50% of times)

6-10 yards (effective run).................24 times (16.78% of the time)

11-20 yards (good run)....................7 times (4.8% of the time)

20 and more (Great run)....................4 times (2.70% of the time)

Fumble ...........................................1 time ( .06% of the time)

-----------effectiveness-----------

-0 to 2 yards (Bad run)...........69 times (48.25% of the time)

3 to 10 yards (good run)...........62 times (43.35% of the time)

11 to 20+ (Great run).......11 times (7.69% of the time)

Fumbled....................................1 time (.06% of the time)

Awsome job dude. SO basically 50% of the time he is going to have a bad run lol.

 

[Awsi Dooger]

I'm not sure anything qualifies as a bad rush. They are all gorgeous to me. Wonderful football.

The theme of a passing league is exaggerated to such extent that Mike Mayock got caught up in the brainwashing last night. When San Diego rushed for the 42nd time in defeating Denver, Mayock announced that most teams rush only 20 or 21 times per game. He actually believed it. Meanwhile, only Atlanta and Dallas are in that area. We are a couple of notches above. I don't remember anybody finishing below 320 rushes, which is 20 per game. The league median is generally in the 26.5 to 27 range.

 

[nyashfan]

I've mentioned this idea before in a similar context. I suggested looking at nonparametric skewness which is defined as:

(mean - median) / standard deviation

It has a nice natural interpretation as its values lie between -1 and +1.

Values closer to +1 indicate positive skewness, meaning that the distribution of individual yards per carry is right-tailed i.e. dominated by a small number of large runs. For two running backs whose YPC are identical, the one whose skewness is lower would indicate greater consistency.

Note: I'd be shocked if there existed a running back with a decent YPC whose skewness is negative!

 

[Shouright]

I've mentioned this idea before in a similar context. I suggested looking at nonparametric skewness which is defined as:

(mean - median) / standard deviation

It has a nice natural interpretation as its values lie between -1 and +1.

Values closer to +1 indicate positive skewness, meaning that the distribution of individual yards per carry is right-tailed i.e. dominated by a small number of large runs. For two running backs whose YPC are identical, the one whose skewness is lower would indicate greater consistency.

Or perhaps lesser breakaway ability? Which should be more valued, the consistency or the breakaway ability?

 

[nyashfan]

I'm not passing judgment by claiming that positive skewness is undesirable. I was trying to address the OP's question.

Barry Sanders' runs I would assume exhibit the classic positive skewness distribution. He's not the running back you would want in short yardage situations; rather, he's the running back you want to carry it 30 times, knowing that he's likely to break one for a huge gain.

Or perhaps lesser breakaway ability? Which should be more valued, the consistency or the breakaway ability?