Revisiting Phil Steele

And I’m sure Huckleberry can have R2D2 spit out an algorithm that has us beating the Saints. There’s no way we finish 11. No. Way. You say his system is based on rankings, therefore quantitative. But aren’t those based off of his qualitative unit/player rankings?

Taylor, you’re right, we need GG to be a star this year, and he will be.

Fair enough. Tack Steele’s predictions to the bulletin board in the locker room.

“But aren’t those based off of his qualitative unit/player rankings?”

No. His qualitative player/unit ratings are a separate product from his forecasts. He doesn’t use them as part of his forecasts.

Ok, thanks for the clarity. However, I refuse to let him point to some 50’s main frame that takes up 75% of his office and blame that as the culprit. Do what the govt. does. Have your answer, then make the numbers work. Scientific method in reverse, imo.

I don’t know why this bothers me so. Blues nailed it. Each toilet in the locker room needs to have a copy of this magazine on the back of it.

Crap, had a response typed out and then the Internet burped.

Suffice it to say that he’s right about our offense scoring more points than the yardage dictated, but our defense also allowed more. That’s not some sort of regressing to the mean situation, IMO, that’s just the sign of a team with a better defense than offense. The offense is leaving the defense with shorter fields to defend so they give up more points than the yardage says they should. And the defense is giving the offense shorter fields than average so they score more points than the yardage dictates.

Anyway, Texas was #30 on offense and #3 on defense in the adjusted yards per play numbers. Oklahoma was #49 and #6. Nebraska was #73 and #4. I think a lot of people’s disbelief is based on the idea that Landry Jones is going to lead a resurgence for the Oklahoma offense. There’s not a lot of room for improvement for their defense over a year ago. So to get from where they were to #1 is going to require a lot of improvement from their offense. And we still don’t know who’s going to block on that subpar OL that lost its best player.

What percentage of points allowed by the Texas defense were in Q4 in 2009? That may skew the analysis some as well.

4. Texas has less experience returning than our peers.

This is arguably a plus at certain positions.

i like Steele’s magazine a lot and i have no problems with his assessment. it does not give a lot of credence to potential, which is logical. it does not measure heart. i think we’ll be fine.

I don’t see how anyone can fault any prediction between 5-15.

All it really comes down to is how many games does he project us to lose, including CCG and bowl game. If we lose 2 we’ll end up in the middle of the top 10. If we lose 3 we’ll likely be just outside the top ten.

Our potential losses are Tech, OU, Nebraska, A&M, Big 12 championship, bowl. Tech and Nebraska don’t really scare me, and I don’t think we’ll lose to both OU and A&M. But you have to think that there might be an unexpected loss somewhere with the young leadership and everyone gunning for us. So my guess is 2 regular season losses, and who knows after that.

In other words, what Scipion said.

*Scipio

Steele is mixing interval and ordinal data and I don’t like it. There are so many components to football. I think it is appropriate to use measurements to analyze certain units or certain phases of the game, but when you put all that together, it’s seldom more useful than the plain-old “eyeball test”.

For example, lots of people had reasons why Ole Miss was going to be good last year, but just looking at Jevan and crew (with the exception of running back) it was pretty obvious they were going to suck. The game has come a long way, but heart still has a lot to do with it.

Scipion sounds cooler. Like a comic book villian.

Shouldn’t all polls be power polls. It’s one thing to predict the championship, even the Top 2, but it’s an excercise in futility to attempt to predict the Top 25 teams. Ranking the teams based on prospective strength is the only way, especially since these “predictive” polls end up as part of the formula for determining a championship (biased and ill-informed Harris pollsters say hi as they open up Steele’s magazine in October).

Scipion sounds like a hip new hybrid my girlfriend would like to drive

also, i think our ranking is about right, we will know who we are by mid October, Little Giants, imo

Scipion,

15?! Are we looking at the same roster? We were one drive away from being preseason #1 or 2 in just about every poll.

I don’t see how anyone could have us out of the preseason top 5.

Wow. I just ran an analysis on Steele’s theory regarding points per yard efficiency and how it should regress to the mean and let’s just say his hypothesis is flawed.

Anecdote: Texas was #5 in the nation in 2008. We went up to #1 in 2009.

Statistic: The correlation between a team’s 2008 number and 2009 number was 0.592, a strong correlation and certainly statistically significant. This is not a figure that should be considered a fluke or pure luck. I was also correct in that correlation between a team’s efficiency value and defensive yards per play allowed value was -0.543 and -0.581 in 2008 and 2009, respectively. Also strong correlations that indicate the more yards your defense gives up the less efficient your offense is at converting yards to points.

And if anyone is curious, there are indeed statistics that don’t correlate from one year to the next, indicating they are based mostly on luck. One of these is forced fumbles per carry, something I didn’t completely expect. Lots of info like this will be in the 2009 Adjusted Stats Year in Review that I’m putting together and I expect to release shortly after the College World Series is over when the focus returns entirely to football.

This is probably the most likely season in forever for Texas to have a truly bad offense. Moving to a type of running game that Davis has never really done very well (aside from ‘98). New, unproven QB. OL lost 3 starters and has no dominant returning players. Allen missed the spring and may not be 100% by fall. Ditto with Walters. Starters at WR and RB look like a bunch of JAG’s or physical freaks with a case of the drops. Only real bright spot there is a track guy that didn’t participate in the spring. And then there’s the luck at TE.

Having said all of that, the Nebraska offense could be even worse.

Huck, does your positronic brain allow you to use contractions? Commander Data’s wouldn’t let him use them. Sorry, kind of random, just wondering.

The Gumps won the NC averaging 32 a game. 32 a game would be considered a shitty O for Texas, I’m sure Gilbert and Co. can avg 32 a game with the feild pos. Booms Boy’s will give them. No worries

Phil is senile, We lack experience? bama lost 8-3yr starters on D alone and 16 total, yet they end up #3 even with the sec scheduling snafu. Honestly this guy has lost it.

In this case I think you should be more wary of Scipion driving your girlfriend, or maybe her older more literate sister.

A look at Phil’s % of #’s for the Big 12 South

% of Letterman returning

Texas Tech 72%
Texas A&M 70.5%
Texas 67.60%
Baylor 67.10%
Oklahoma 67.10%
Oklahoma State 57.9%

% of yards returning

Texas Tech 96.53%
Texas A&M 90.79%
Oklahoma 81.00%
Baylor 72.01%
Texas 40.42%
Oklahoma State 34.81%

% of tackles returning

Texas Tech 72.3%
Texas A&M 70.5%
Texas 67.6%
Baylor 67.1%
Oklahoma 67.1%
Oklahoma State 67.9%

OL career starts returning

Texas 47
Texas A&M 43
Oklahoma 42
Baylor 37
Texas Tech 29
Oklahoma State 12

Mediocre offensive line, running backs, receivers and tight ends; unproven qb. 11th may be high. Oh, and MacWhorter is still employed.

Love the way you “S” types sling the data around. Us “N” types just purr, “C’mon. We’re Texas.”

I saw greatness in young GG last January. He’ll be playing the last game of the season again, if not this coming season, the next.

There’s my ransom!!

I enjoy i when ransom parodies himself. It’s like Mary Shelley book in a book stuff.

Statistical analysis on the projected productivity of young males in a series of athletic competitions months from now.

Pointless.

“Pointless.”

Like everything else @ the B.C., eh?

It’s just what we live for, is all.

I’ve heard that mack and co think gilbert is the complete package and this is part of the reason they didn’t red shirt him(they new they would only get three years out of him because he’s that good). I agree with oldtime horn. This kid will be a stud. Afyer the nc do you think lincoln or the rrs will rattle him? In no way I’m saying he will for sure win those games but his experience and intangibles has to count for something. The nc game was the first time I saw a texas qb audible more than once in one game since major. He’s major but with the physical tools IMO. But I’m drunk… Just a thought

Both Steele’s arguments against Texas performing at the same level as last year along with his arguments that OU will perform better than last year directly contradict his other primary source of reasoning in predicting teams…returning talent.

Texas doesn’t have nearly the same team as last year on offense nor the same intended strategy.
The results of Colt’s offense and whether or not it was statistically unlikely has very little to do with Gilbert’s offense and whether things will bounce back. Why should they?

He uses the same reasoning in explaining that OU lost a lot of close games and statistically should be luckier this year and win some close contests?
Besides the fact that the returning OU team is a lot different than last year’s squad why should they be luckier this year? If you flip a coin and get heads then flip a different coin the odds of flipping a tails instead are still 50%.

None of this is factoring TTRoom’s explanations for why Texas was lucky in yards/points ratios (field position). All that to say, Steele is trying to be way too smart for his own good. Simply evaluating the talent levels at crucial positions along with the expected schedule difficulty would yield a far more well-reasoned prediction.

2. We had only the 3rd best yardage differential in the Big 12, behind OU and NU.

TTR’s take: Valid point, but we had several games where Colt just wasn’t himself, whether due to illness or injury. Since he was the linchpin of the offense, that severely hurt our yardage. Yes, I know we won’t have Colt at all next year.

What is he calling yardage differential? This one is just plain not true as far as I can tell. Whether it’s yards per game, yards per play, yards per possession, and raw or adjusted Texas had the best yardage differential in the Big 12 by all six measures.

Yards per Game Differential -

1. Texas
2. Oklahoma
3. Texas Tech
4. Nebraska

Adjusted Yards per Game Differential -

1. Texas
2. Oklahoma
3. Texas Tech
4. Nebraska

Yards per Play Differential -

1. Texas
2. Texas Tech
3. Nebraska
4. Oklahoma

Adjusted Yards per Play Differential -

1. Texas
2. Oklahoma
3. Texas Tech
4. Nebraska

Yards per Possession Differential -

1. Texas
2. Oklahoma
3. Texas Tech
4. Nebraska

Adjusted Yards per Possession Differential -

1. Texas
2. Oklahoma
3. Texas Tech
4. Nebraska

Just checked. He seems to be counting per game offense, conference games only. He has OU at 133, TT at 107, and Texas at 96 (NU is not 2nd, but lower at -16).

“He uses the same reasoning in explaining that OU lost a lot of close games and statistically should be luckier this year and win some close contests?
Besides the fact that the returning OU team is a lot different than last year’s squad why should they be luckier this year? "

OU started 10 different offensive line combinations last year due to injury. Does every team lose not one, but BOTH, of their best offensive players by halftime of the first game? Even if OU goes to having just slightly bad luck instead of utterly atrocious luck, they should be an 11 win team or better. To compare, UT would’ve had to lose Colt McCoy, Jordan Shipley, and at least 40 combined starts by OL. Imagine having to start Adam Ulatoski at center. That’s how bad it was by the end of the year.

The really surprising part of the forecast is that he has OSU last in the B12 south. Not because Baylor is good, but because he expects OSU to be bad…like maybe 3 – 9 bad. People aren’t used to that, so they will be very impressed when TAMU starts out 4 – 0 with a win in Stillwater.