Offensive Review: Giants

[hooskins]

Gtripp, always great stuff. I really hope your Olineman stat works out.

I would be very interested in seein the result at the end of the year. Even Samuels sample size isnt high enough. The rule of thumb is 30, but in this case it depends on N(the number of time Samuels will run block in his career) to calculate a decent sample size to potentially make an accurate statement about each lineman. Yeah, I took a ton of stat classes.

If you are interested in football stats, look into take some multiple regression/econometrics classes. You could compile a bunch of stats, try to isolate vars and reach some meaningful conclusions.

[GTripp0012]

Originally Posted by hooskins

Gtripp, always great stuff. I really hope your Olineman stat works out.

I would be very interested in seein the result at the end of the year. Even Samuels sample size isnt high enough. The rule of thumb is 30, but in this case it depends on N(the number of time Samuels will run block in his career) to calculate a decent sample size to potentially make an accurate statement about each lineman. Yeah, I took a ton of stat classes.

If you are interested in football stats, look into take some multiple regression/econometrics classes. You could compile a bunch of stats, try to isolate vars and reach some meaningful conclusions.

Yeah, it takes about 3 games before I'll have any semblance of a meaningful sample on the left side of the line, and probably about twice that for anything on the right side of the line.

I also might run into a problem where Guard numbers are just higher than tackle numbers, figuring that if a running back can ever find a hole off guard, he probably won't run to the tackle.

In the end, when I have a significant sample size, I'll probably adjust the numbers so that the mean lineman yards figure equals the yards per carry average for the Redskins backs. Right now, it's just a very raw average without any context.

[hooskins]

Originally Posted by GTripp0012

Yeah, it takes about 3 games before I'll have any semblance of a meaningful sample on the left side of the line, and probably about twice that for anything on the right side of the line.

I also might run into a problem where Guard numbers are just higher than tackle numbers, figuring that if a running back can ever find a hole off guard, he probably won't run to the tackle.

In the end, when I have a significant sample size, I'll probably adjust the numbers so that the mean lineman yards figure equals the yards per carry average for the Redskins backs. Right now, it's just a very raw average without any context.

Yeah, but it's great stuff nonetheless. I guess overall it wouldn't make sense to compare the number for a G vs an OT just straight up. Rather we could use other factors, or just use the raw stats and compare tackles to tackles, etc.


Imagine at one point we might even have a formula, with which we could calculate how effectively(on avg) the Oline could be. I am thinking a y= Ax+Bx+Cx+....obviously not just linear, and even vars to calculate each of the dlineman, and their impacts on the olineman too. Obviously this is all theory but is it nerdy to get excited over this? lol.......

[GTripp0012]

Originally Posted by hooskins

Yeah, but it's great stuff nonetheless. I guess overall it wouldn't make sense to compare the number for a G vs an OT just straight up. Rather we could use other factors, or just use the raw stats and compare tackles to tackles, etc.


Imagine at one point we might even have a formula, with which we could calculate how effectively(on avg) the Oline could be. I am thinking a y= Ax+Bx+Cx+....obviously not just linear, and even vars to calculate each of the dlineman, and their impacts on the olineman too. Obviously this is all theory but is it nerdy to get excited over this? lol.......

Not until it's proven pointless!