Eight weeks ago, the Vikings were undefeated and ranked No. 1 in DVOA. Now they're 6-6 and probably done. What went wrong?
21 Nov 2008
by Bill Connelly
It is quite conceivable that the Big 12 will end up with five 10-win teams this season. Texas and Texas Tech are already there, while Oklahoma, Oklahoma State, and Missouri are just a win away. Of the eight matchups between those five teams, only two remain: Oklahoma at Oklahoma State over Thanksgiving weekend, and Texas Tech at Oklahoma this Saturday night. In the BCS Standings, that's No. 2 at No. 5. Rankings-wise, it's not even the biggest matchup of the year in the Big 12 South (OU-Texas was No. 1 vs No. 5), but we're getting pretty late in the season. The hype and pressure are getting ratcheted up with each new huge game.
In recent weeks, we have gotten into a habit of looking at a "Varsity Numbers Box Score" of a big game from the previous weekend. This time around, let's see what the numbers tell us about OU-Tech in advance.
Both teams have now played six conference games against five common opponents. Meanwhile, OU's non-conference schedule (with matchups against potential Big East champion Cincinnati and BYU-killer TCU) was demonstrably tougher than Tech's. So with that in mind, we will look below at conference-only stats.
(Note: For newcomers, a number of the stats used in Varsity Numbers are now defined in the Football Outsiders glossary.)
|59.7%||40.3%||Field Position %||62.9%||37.1%|
|^ "Close" is defined as a scoring margin of less than 17 points, i.e. within two possessions.
"% Close" refers to the percentage of a team's games that were played under "close game" circumstances.
Oklahoma comes out of the gates faster than Curlin. In six conference games, they have outscored opponents 126-24 in the first quarter. On average, they are already up 17 points when the second quarter begins. If the Red Raiders can avoid an early hole, they can find themselves in a good position. Quite frankly, Oklahoma is not too used to close games, and against the one team continued to respond to them on the scoreboard (Texas), they eventually wilted.
At the same time, if Tech doesn't withstand the early assault, this Game of the Century of the Week could be over quickly.
While the first quarter may not favor Texas Tech, leverage does. Oklahoma finds themselves in quite a few more passing downs than Tech, and as we will see, Tech's passing downs defense is quite good.
Oklahoma games have, on average, about 10 more plays than Tech games. That is impressive considering Tech doesn't exactly plod down the field. OU banks on the fact that they can take advantage of your defense more than you can take advantage of theirs, and for the most part they have been absolutely correct.
In close games, Texas Tech has a distinct advantage in terms of success rates (+10.8% to OU's +3.8%), while OU comes up huge on PPP (+0.23 PPP to Tech's +0.16). In my Win Correlations column, I mentioned that PPP (explosiveness) is more closely tied to wins and losses than success rates (efficiency), which suggests Oklahoma is in good shape, but this does not suggest a large advantage, only a slight one.
I have mentioned this a few different times, but I will once again deliver my stock line on Texas Tech: If they ran the ball 19 straight times, the opposing defense would still be playing the pass on the 20th play. They have used that to their advantage more this year than before, but as we are about to see, they still do not use it a lot. Tech has better rushing numbers, but Oklahoma uses the run in a completely different way. In a third-and-short and other power running situations, Oklahoma is much more proficient.
Michael Crabtree is the star, but thanks to guys like Detron Lewis and Eric Morris, Texas Tech is able to maintain a ridiculously high level of efficiency (as exemplified by a 62.5 percent success rate). OU's success rate isn't as high, but they make up for it somewhat with big plays.
Defensively, OU is more likely to come up with a stop, but only marginally so.
In non-passing downs (first-and-10, second-and-4, etc.), Oklahoma is very much a run-first team, which is unique for a hurry-up, no-huddle attack. They run to set up the pass, and while they can still come up big if the run isn't clicking, they are all but unstoppable when they can run the ball well.
While OU runs the ball almost five of every eight non-passing downs, Tech only rushes about three of every eight. They do it just enough to keep defenses off-balance, which shows in the fact that Graham Harrell gets sacked 0.8 percent of the time he goes back to pass on non-passing downs. Think about that. For every 250 pass attempts in these situations, Harrell gets sacked twice.
One other thing to note: Texas Tech gives up an uncomfortably high success rate on non-passing downs. They take advantage when they get opposing offenses into uncomfortable, high-leverage situations, but they are not the best at getting into those situations.
Two weeks ago, I talked about how Texas' insanely high S&P on passing downs caught up with them when they went to Lubbock and found themselves knocked out of their rhythm by a play-making defense and high-intensity crowd. Well, the shoe is now on the other foot. If Tech can maintain their 57.0 percent passing downs success rate on Owen Field, in a stadium that I said the other day is the loudest place in the country on a big third down, then they deserve every ounce of praise they receive, and then some. They likely will not be able to rely on passing downs success to keep drives going Saturday night. They will need to stay out of those situations as much as humanly possible.
|25.72||79.46||Total T/O Pts||38.44||68.60|
|+53.74||-53.74||T/O Pts Margin/Gm||+30.16||-30.16|
On average, OU benefits about nine points per game from turnovers, while Tech has about a five-point advantage. Taking Owen Field into account, OU stands to possibly gain about 5 to 7 points in this category on Saturday.
First of all, we can ignore fourth-quarter numbers, particularly those of OU. Joey Halzle has played as many or more snaps at quarterback as Sam Bradford in the last six games, and contrary to his reputation, Bob Stoops often does as much as possible not to run up the score (though it's easy to take the high road when you have 50-plus points heading into the fourth quarter, as OU has managed the last three games).
These numbers suggest that the narrative for the game could go like this: OU jumps ahead early with a couple big plays and a quick stop or two, then Tech strikes back and makes it a game in the second quarter. Whoever wins the third quarter, wins control of the game.
Texas Tech's third-down numbers are amazing. On offense, they're simply automatic; on defense, while they still are not great down-to-down, they pick their spots well and come up big on third downs as often as not.
At my Missouri/Big 12 blog, I have been running "+" concept projections based off of Big 12 games; this week I projected a 34.1-34.0 victory for Oklahoma in this game. It's that close. Taking turnovers into account, however (which I do not do in the aforementioned projections), Oklahoma builds about a 7-point advantage. In the end, I think passing downs and turnovers will make the difference. This is Mike Leach's best team at Texas Tech, and if they are ever going to win in Norman, it will be this year. But their disproportionate success on passing downs will likely dry up temporarily, and the turnover bug could strike at an inopportune time.
With an OU win, we would head into Thanksgiving weekend with three different South teams in position to win the South title. That is what I expect to see.
Oklahoma 41, Texas Tech 34.
Finally, this has nothing to do with OU-Tech or football statistics in general; it has to do with pure morbid fascination. With everybody having played at least six conference games, only one ACC team is more than two games out of their division lead.
In the Coastal Division, Georgia Tech is in the clubhouse at 5-3. Miami is 4-3, while North Carolina, Virginia, and Virginia Tech all sit at 3-3. In the Atlantic Division, Maryland is 4-2, Florida State and Wake Forest 4-3, Boston College 3-3, Clemson 3-4, and North Carolina State 2-4.
It's a mess.
What games remain?
Florida State (4-3) at Maryland (4-2)
Boston College (3-3) at Wake Forest (4-3)
North Carolina State (2-4) at North Carolina (3-3)
Clemson (2-4) at Virginia (3-3)
Duke (1-5) at Virginia Tech (3-3)
Maryland (4-2) at Boston College (3-3)
Miami (4-3) at North Carolina State (2-4)
Virginia (3-3) at Virginia Tech (3-3)
North Carolina (3-3) at Duke (1-5)
So if all the home teams but Duke (who looked feisty a month ago but has since faded) hold serve, we end up with something like this:
Virginia Tech 5-3
North Carolina 5-3
Georgia Tech 5-3
Georgia Tech lost to both North Carolina and Virginia Tech, and VT beat UNC, meaning the Hokies would be the Coastal representative.
Wake Forest 5-3
Florida State 4-4
Boston College 4-4
North Carolina State 3-5
Maryland's 26-0 win over Wake Forest on October 18 would give the Terps, who started the season creeping by Delaware and getting dumped by Middle Tennessee State, the Atlantic Division.
But this is all assuming the home teams win out. In this land of extreme parity, that's far from a given.
3 comments, Last at 27 Jul 2009, 1:37am by Malena