Adjusted Stats through 11/5/11

I’d like to hope we project better than our season stats predict, because it really feels like we’ve turned a corner on both offense and defense these last two games.

Hopefully it’s not just quality of opponent, but it feels real.

The move of Poehlmann to TE and the sudden emergence of Jeffcoat and Okafor have filled in some critical gaps that kept us from being consistently productive.

I’m looking forward to this challenging stretch.

Hey Huck,

Just wanted to remind you that the work you are doing with Adjusted Stats is fucking nails and I greatly appreciate it.

I’m looking for 9 wins. The Aggie game should be a dog fight.

Of course I’m looking for 10 wins but realistically we do have a couple of tough games we could split.

CO -

I agree completely. As an example, our adjusted total rushing per carry rank was in the high 70s two weeks ago and is now #51. That kind of movement this late in the year seems like a big deal to me, at least intuitively.

LHS -

Thank you. Same to you.

As in any sport, there are the intangibles & immeasurables, i.e. the mental factor, confidence, etc. The stats bear that out, but I think if you can add the “II” factor (or maybe call it the “It” factor), their percentages for winning would be higher. This team doesn’t quite have 100% “It” yet, but that metaphysical “It” meter is definitely rising.

Huck, after the massive data dump this summer are these still the stats that most correlate to success on the field?

Texas
1.) Yards Allowed per Play (Defense): 10
2.) Total Passing Allowed per Attempt (Defense): 8
3.) Total Passing per Attempt (Offense): 52
4.) Yards per Play (Offense): 51
5.) Total Rushing Allowed per Carry (Defense): 25
6.) Turnover Margin per Possession: 60
7.) Total Rushing per Carry (Offense): 41

I am taken in by Texas new-found running prowess. The OL seems to have jelled over the past two weeks as well. Texas appears to be developing into an overwhelming behemoth.

I think KSU at our house is the toughest remaining opponent followed by A&M, away. If Texas beats KSU, then I think Texas finishes the regular season 10-2, even though 3 out of the last 4 (all good clubs) are away. Let’s see what we can do in Columbia this weekend against a solidly average Missouri Tiger team.

BTW, Huckleberry, why do you keep the “T” column? I assume that stands for “tie”? I am very impressed that your rankings comport quite accurately with the real-world. I guess you got the weighting factors balanced pretty good through the years.

maninblack -

Yes, those were fairly consistent. All in all, yards per play on either side of the ball is the best correlation.

XOVERX -

The same php script is used to display the ratings for every season from 1869 through the present, so the ties column needs to be displayed.

Great stuff, Huckleberry. Cannot wait to see the post mortem and film analysis from Scipio and LHS.

Would really appreciate the Barkers commenting on whether the accumulation of 800 yards of rushing against KU and TTU is any indication we can control the clock and tempo to the point of beating K State and A&M. We have not beaten both schools in one season since 2003. Winning 10 this year, only losing to loaded Oklahoma schools, is a dream season and speaks of great things in 2012.

According to your ratings, Aggies are the best 5-4 team in the country. For some reason, this amuses me

The thing to remember is these numbers reflect snapshot realities. As y’all may have noticed, the odds of winning each particular game have gone up every week… there are reasons for that. Stop and think about the implications of the odds going up every week if they continue to go up every week.

8-4 is the most likely as of today. Ask again next Sunday, and the Sunday after that, and you may get a different answer. Just sayin’.

I think KSU is our toughest match-up remaining, but I like our D to keep us in the game there. We just can’t have our typical KSU clusterfuck game and turn the ball over.

What I’m most looking forward to seeing the rest of the year is Ash and the offense turn a corner. We found our power running identity and that’s great, but plenty of mediocre teams have identities. We need to find that “swagger”.

The point where Ash stops managing the offense and starts to command the offense. When it feels like the defense is just helpless against being attacked any which way we desire. Most importantly, I just want to be able to stop holding my breath whenever we’re in obvious passing downs.

Maybe we won’t see it until the bowl or next year, but I hope we at least see glimpses of it this year.

Love what you do, Huck – One geek to another.

Here’s a thought: Have you considered any way to account for changes in a team over the course of the year? A basic assumption in your methods is that it’s the same team playing each game, which I think is perfectly fair for ranking teams to reward them (BCS style), but clearly has shortfalls when you consider things like Oregon in 2007, or a team in Texas’s position this season, in which we would expect 12th game Texas to be much better than 1st game Texas (we would also expect Texas to have improved more than the average team improvement over the course of a season, so it’s not washed out by the general improvement during the season of every team).

In cases like these, your model could benefit from some kind of dependence on the d/dn derivative (n is number of games played) of a team’s overall adjusted predictive rating. Actually, I’m not sure about the Oregon example, because that’s a sudden change and just seems like something you can’t quantify in the current system and will have to adjust for subjectively. However, I would hypothesize that most teams improve (or possibly get worse) over the course of a season and at varying rates, which could be quantified and factored in to both long term (season-outlook) projections and short-term (next game) predictions.

In short: I think that, due to the improvement curve we’ve seen from this team, I would be slightly more optimistic than the win-probabilities that the model predicts for the remainder of the season, and I think that, with some careful analysis, there is a potential to quantify it. Have you already considered this?

I would say that the new over/under line for this team is properly set at 9 wins. I agree somewhat with the above noted problem of the team “changing”. However, it would probably depend on a “player weighting” of some sort (think just how much OU’s chances have changed overall now that Broyles is out). Given that 3/4 of our last games are road games, it will be a tall task to achieve a 3-1 record going forward. 2-2 would be an average result, 3-1 excellent, and 4-0 possibly an outlier of expectation. That being said, I’m impressed with the results this year, not because of record per se, but because the new staff truly embarked on changing the team’s “style of play”. I would grade the previous year’s recruiting an A+ and the coaching a B, which is substantially a better grade than last year and even the previous MNC year. Great stuff, keep it up!

We are born to stop the run and none of the remaining four teams play any defense whatsoever. Missouri is a weird team but I still think Texas wins that game by 14 points. James Franklin has a weak arm and decent defenses can force him into becoming a turnover machine. Their defense has also fallen apart somehow which is beyond my comprehension.

I still think Texas goes 9-3 and plays in the Alamo or Cotton Bowl.

Stats are always interesting but can hide certain truths. As noted by Hornbrain and Tower a young team has a steeper learning curve that a seasoned team. But if it begins to climb that curve calculating its future based on past averages is behind the curve.
Also, college football is a highly emotional game. A team that started well (such as a certain top ten ranked team soon to be leaving the Big 12) and then loses significant games, may lose much of its enthusiasm for playing when it faces a new challenge. Or as noted above a team that loses a key player may not perform as well. While a young team that makes a significant win (such as winning more games than last year) or discovers hidden talents on their team such as a great kick-off returner, an incredible second or third string running back, or a better quarterback, a shut down cornerback etc, can face challenges with renewed vigor.
Thus I would predict based on these learning curves we are on are way up while our neighbors to the east may well be on their way down.

Over the last few weeks, we’ve gone from around 7th in various Big 12 “power polls” to 4th. Some of it is us, but more is the opposition. Four losses in seven games has dropped the Aggies a ton . . . Missouri has had several down games . . . Baylor likewise.

Here’s the emerging reality that has me shaking my head 20 hours a day: Tech is now 9th IMO, falling from around 6 to 9 after that egg-laying in Lubbock against Iowa State.

But . . . I still think Blow-U beats OSU, so Sooners will end up No. 1.

With the loss coming to the No. 9 team.

At home.

This seasons feels a lot like 98 . . .both coming off 7 loss seasons . . . both had two early season blow-out losses (well, this year’s OSU was double-digits but maybe not a blowout).

Anyway, the 98 team got better as the season progressed behind a freshman QB (albeit a redshirt freshman) and was better than everybody it faced down the stretch. Still, that team lost a close one to Tech in Lubbock.

Remembering that finish, it’s easy to picture one of these last four going the wrong way.

9-3 seems likely . . . 10-2 with a bowl win would shoot expectations for 2012 through the roof.

Burrito…. if you put that bong down during the game you’ll hold you breath less I figure.
Good stuff Huck… question, as good as the movement on yards per carry on offense over the last 2 games how do we look in the same regard on the defensive side of the ball.
just curious.
Thanks again for the brain food between games.

Huck,

Have you calculated an (AQ conference) expected value for 2-point conversion tries? If so, what’s the EV, and were the influential variables?

“With the loss coming to the No. 9 team.”

edsp:
As a Sooner fan, I think I can say that all 3 phases of the game failed in the OU loss to Tech. Tech didn’t turn the ball over, they ran and played hard, but OU was distracted on offense and defense, and the kicking game has been mediocre all year.

OU will learn a lot about their offense in the next 3 weeks. Broyles-Jones is reminiscent of McCoy-Shipley Part I. Both connections kept the chains moving with throws to the middle of the field.

OU will need Reynolds or someone else to step up in the slot role. If not, I don’t think the Sooners win in Stillwater.

Until Broyles got hurt I thought there was no way OU could lose to OSU. Now I’m not so sure. I tend to think OU’’s D is just that much better, but then there is that Tech game, and OSU is Tech if Tech came to play every week and had some talent of offense.