• New here? Register here now for access to all the forums, download game torrents, private messages, polls, Sportsbook, etc. Plus, stay connected and follow BP on Instagram @buckeyeplanet and Facebook.

Game Thread Ohio State 3, Southern Cal 35 (Sept. 13)

Jump to latest Follow Reply
plum diamonds buckeye planet
Sort by date Sort by votes
Is it too late to say something witty to make it appear as though this conversation is going well?

Posturing aside, does anyone know of a good website that has all the statistics and data for a given game? I want the yardage yielded from each and every run and pass. CFBstats is good, but doesn't go to quite that depth.

The reason I ask is because this whole conversation has peaked my interest in the topic. The traditional football stats use means, which are very sensitive to statistical outliers (as we've seen in this thread). I'm just curious how sensitive they really are, and would like to figure the variance and do a little math on the numbers.

Getting back to this game, I think the key (for the Bucks) will be defensive line play. Last season it was pretty clear that our linebackers struggled when lineman got to the second level and isolated guys (think the first half of the Penn State game and the MNC game). When our guys were free to flow up and down behind the line and shadow the line of scrimmage they were lights out. As good as McKnight is, I think if our guys don't have a solid hat on them they'll be able to keep him in check. He's going to get some yardage and probably score, but if we can control the line he's going to have a long day.

I know, that's no major insight: control the line of scrimmage. All the same, it's true despite being clich
 
Upvote 0
I agree with your point about the D line. Penn State did a great job of getting a lineman on Laurinaitis and getting Kinlaw in that hole. LSU absolutely put on a clinic in team blocking. I don't know if USC will be able to pull that off with their inexperience on the O line. Usually when O linemen get to linebackers, it involves pulling a guard or some other stunt. USC may not be too effective with that so early in the season. I would hope that our D line will show marked improvement with players maturing and returning from injuries. Last year we were scrambling just to put bodies on the field.

As far as finding play by play stats, I know I've seen sites that have entire games broken down that way. I haven't been able to find one, but if I do, I'll post it. jwinslow listed Mendenhall's big runs a few posts back, you might ask him. I'm hoping he didn't reel that off from memory.
 
Upvote 0
Short version:
Ok, ran some quick numbers. Median for Mendenhall was 3 ypc and population mean confidence interval for 98% (fancy stats way of saying that if he had an infinite number of rushes against USC's defense, in what range would we expect to find the mean ypc 98% of the time) is (-2.19,20.4) ypc. That's ridiculously big, and it's a direct result of the standard deviation being so big (20.01 yards).

The median is a nicer number than the mean because it tells us precisely where the 50% cutoff is. In Mendenhall's game against USC, it tells us that 50% of his runs were for more than 3 yards.

Long version (or "Math is Messy"):
Unfortunately this really doesn't help the issue; the problem is that, statistically speaking, n=17 is a small sample size. I really shouldn't even assume that the stats are roughly normal for this small a sample, and should use a t-test, but I'm lazy and don't know the standard variable values for t-tests off the top of my head.

Just for a basis of comparison, Beanie's game against scUM had a median of 4 ypc, standard deviation of 10.6 yards, and a population mean confidence interval for 98% of (1.7,9.6) ypc. Again, a series of long runs (62, 24, 16, 12, 12) ensure that. The stats are a little better in this case, too, because the number of samples (n=39) is above the rule of thumb cutoff of n=30.

I may go ahead and try to do some more stats for the entire season, as long as I'm at this, but I'm not thrilled about transcribing hundreds of data points. I guess it could be worse though.
 
Upvote 0
generaladm;1185990; said:
As far as finding play by play stats, I know I've seen sites that have entire games broken down that way. I haven't been able to find one, but if I do, I'll post it. jwinslow listed Mendenhall's big runs a few posts back, you might ask him. I'm hoping he didn't reel that off from memory.
Click to expand...

NCAA Site
 
Upvote 0
TheIronColonel;1186036; said:
Median for Mendenhall was 3 ypc.

Beanie's game against scUM had a median of 4 ypc.
Click to expand...

I appreciate the effort in tearing the stats apart.

That's very interesting considering the gap in carry totals. I understand the idea of the median, but this in all its efforts is still confined by the fact it excludes the weakest aspects of the defense/strongest aspects of the offense.

The truly curious cases would be in the more extreme circumstance before becoming relevant. From 9 ypc to 3 ypc by running the median doesn't jump off the page considering the lack of total carries. Indeed it tells more of well rounded account of play by play, but not the degree that tells me Mendenhall was held in check and therefore relegated due to the final score. In fact, its similar to Beanie's 146 on 20 with a 65 yard sprint. The real hole in both games was turnovers, not their unaffectiveness.

Say Rashard had 20 carries; 19 for 5 yards, and one 95 yard run. The medium would therefore be far lower than the 5 ypc that would show from a standard percentage calculation. This is where the decent ypc would be a facade.
 
Upvote 0
BuckeyeGanoosh;1185720; said:
Stoneburner is 6'5", 230, and has been solid all spring. He ain't built like a freshman.

Beyond the fact I would love a buckeye we can call Stoney, his size for a WR jumps off the page. He's built like a TE, played TE, but is listed as a wideout...who would you put on him from USC if we're man to man?
Click to expand...

What about Taylor Mays?
 
Upvote 0
You must log in or register to reply here.
Back
Top