Sunday, September 19, 2010

Revisiting the Bobo Assessment

I was urged by a reader to look deeper into the statistics I gathered, and so I've done so. I'm sure that when I complete the research I mentioned to him, even more will be revealed. For now, though, still sticking with the statistics I have already compiled, I show the following:

2006 - UGA's offense under-performed offensively in one out of the two games Bobo called.
- 1/2 = 50%

2007 - UGA under-performed offensively in three games, losing two.
- 3/12 = 25%

2008 - UGA under-performed offensively in one game.
- 1/12 = 8%

2009 - UGA under-performed offensively in three games, losing all three.
- 3/12 = 25%

Now, these totals seem far better than what I had in my original assessment. Part of this is because I went back and looked at games where only one area of the offense was "lacking." When I did that, I eliminated that game as an "under-performance" if we made up for that lacking by increasing the other phase of the game by close to the same amount or greater. For instance, in the 2008 game against Georgia Tech, our running game was sub-par by 40 yards, but our passing game was 214 yards better than what teams usually accomplished against them. So, that's a wash - apparently a game-plan difference.

Using THAT analysis, we see that Bobo's offenses in 2008 actually performed much better against the competition than their average opponent's offenses than the 2007 versions. However, people think of 2007 as being a much better season. Now, add to that the fact that in all three of our losses, our offense performed by-and-large BETTER against the teams who beat us than the average opponent did, and you get a different idea about 2008 altogether. Admittedly, of course, these statistics do not take into account SITUATION. Alabama was up on us 30-0 at the half. We did most of our damage in the last quarter and a half. The same is very true in the Florida game. Against Georgia Tech, our defense fell apart in the second half, and we couldn't keep pace with Tech's option. The funniest part? None of these games are the game that our offense under-performed. That was against South Carolina, where we were in the red in all four categories. In our two blowout losses, we were never in those games, and that's what causes the frustration on the part of the fanbase. When we faced the best teams, we didn't look competitive.

The question was posed to me - if these success percentages are too low, what am I looking for? Well...I want to see our offense perform better against our competition than the average offense does EVERY SINGLE WEEK. Why? Because I want our team to be the best. If our team is the best, we won't perform below AVERAGE. I will be looking into how the last four National Champions performed using this same analysis, because my hypothesis is that they did much better than we did. If I'm wrong, I'm wrong. We'll see.


Also - if anyone knows a good way to set up a table online so that you guys can have access to my raw data, let me know. I'd love for you to be able to look at the same thing I am, but there's no way I'm "html"ing this many cells.


Go Dawgs.

5 comments:

Cool Cal said...

google docs - you can make public and upload excel or openOffice to it.

Ben Dukes said...

Thanks Cal! I'll definitely have to do that!

Gawdsports said...

I used part of your blog for my piece. http://leatherhelmetblog.com/2010/09/22/guest-post-in-depth-look-at-offense-leadership/

I want to apologize that the qoutes didn't format well w html environment. I sent it in done correctly, but between my hands and actual posting by ecdawg things went hay wire a little bit.

As to your Q: regarding codes ecdawg gives a possible work around. He saved the xls graphs as a pic then copy and pasted the pic. The code thing is too complex for me; I need something more GUI as well.

If you would like the "stats" behind my work you can email me @ jimfrankli@gmail.com

Holla said...

Ben,

This is interesting, and thanks for the work. It always gets more interesting the more layers you peel back (so, again, thanks for the work, since I'm not the one doing the peeling, only the enjoying).

For clarity: you are now only counting a game as an offensive "underperformance" if UGA's offense did worse than the average opponent in BOTH rushing and passing yards, right?

That is clearly better than simply counting any shortfall as an underpformance. But, I wonder why we don't simply use total yards, and not worry about passing vs. rushing at all?

The point is, taking into account our differences in game plan, talent, available personnel, and opportunistic "taking what the defense gives us," how well did we move the ball on their D compared to average. Right? But that is a really total offense issue, period.

If we rush for more than average but pass for less, it's still not clear whether we have outperformed the average opponent that D plays. We need total yards to know that.

Also, Georgia Tech must outperform in virtually every game, even when they are nowhere near a national champtionship caliber team, simply because they will virtually ALWAYS get more rushing yards than the average opponent. Not having run these numbers, I hedge my bets by saying "virtually," but if I had to bet on it I would guess that they have NEVER, under Paul Johnson, underperformed in rushing yards. (What is a bad game for Tech's rushing O? 200 yards? Maybe that will underperform, depending on the D they are playing. But in the ACC, probably not.) Or, heck, how about Johnson's Navy teams?

I think making one part of the offensive attack sufficient to say you "outperformed" is too LENIENT a measure of the O's success.

Also, ANY team with a QB who runs often is going to have a leg up on this statistical measure. As in, Florida with Tebow. Sure, after he bangs out his 80 yards and largely avoids sacks by baby-rhinoing his was back to scrimmage even on a broken play by the OL, Florida is going to have a very good rushing total. (And, of course, Florida's O the last three years is deservedly going to have HIGH numbers anyway...they were a very, very impressive offense. But, that 08 teams spent the first month of the season with a national meme hanging around their necks about their offense struggling. But, if all it takes to "outperform" the average is to do so in one category, regardless of total yards, then I'll bet UF's number comes out looking better than it should. Though, again, they are going to have a very good number regardless.)

One minor clarification about the 08 blowouts. The Bama game definitely saw most of our offensive production come in garbage time. But in the Florida game, we moved the ball throughout. Turnovers killed us in that game. We would both drive the length of the field, but they would get 7 every time while we would get 3 or even 0 (missed a FG up pretty close as I recall in the 1st quarter that would have given us a quick lead in that game, and of course any time you turn it over you get 0). Seriously, I think we had MORE yards than Florida in that game. (We definitely cleared 400). It's just a minor point, though, made for sake of accuracy.

Holla said...

Hey, one other thing I forgot.

As a matter of simple statistics, polynomial distribution and all that, it is a little crazy to say that you expect us to NEVER have a single performance that is below the average. No team clicks on all cylinders every time out, and any time you don't you will risk falling below the average for the opponents of the D you are playing against, even if that D is relatively bad. In fact, the worse the D is, the more likely it is that your own shoot-in-the-foot performance on a particular day will leave you short. You know this, but it is one of the big challenges (and frustrations) of football, right? Getting eleven guys to move as a coordinated whole is difficult stuff. Some days the timing and execution just aren't quite where they usually are.

(Michael Jordan made 80% of his free throws over his career, right? But he certainly fell below that average on individual nights, and sometimes surely he had such a poor night (1/2...50%?) that he was below the league average for free throw percentage. To say that being the best means you can NEVER have a day like that seems naively utopian to me. If I may suggest so delicately. :-)