So it looks like the Texas Longhorns are back to, uh, not being back. At least in terms of contending for the national title this year or the next.
True, only the orangest of bloods thought that the Horns had a shot at the crystal football this year, even with a 6-1 record complete with a signature win over Oklahoma. And now, after falling to underdogs Oklahoma State 38-35 Saturday night, even those who believed this team could shoot the moon are left earthbound.
Gone are the hopes of a shot at the playoff at season’s end, with the consolation that the Horns remain in pole position for the Big 12 title. And should they win their last four games, a New Year’s Six Bowl game would be their likely reward.
But how likely is that scenario? There are two ways to look at it: the old-fashioned human way, and the newfangled computer method.
Judged by time-honored yardsticks— wins and losses of upcoming opponents, past success and failure against common foes, home-field advantage for remaining games, year-after-year trends, polls of coaches and sportswriters, and overall season resume—it’s not very unlikely at all.
At number twelve in the AP poll, West Virginia, this weekend’s opponent, has the best ranking of any team left on the fifteenth-ranked Horns schedule. Sure, the Mountaineers have only one loss on the season versus the Horns’ two, but over the past three-and-a-half seasons, the visiting Mountaineers are a pedestrian 11-12 away from their home stadium in gloomy Morgantown, city of 1,000 burning couches. What’s more, according to various sports books, the Horns are picked to win a close one.
As for the rest of the Horns’ remaining opponents, there are those Cyclones, who are unranked, visiting Austin, and have already been beaten by TCU and Oklahoma, two teams the Horns have defeated this year. Texas Tech is likewise unranked, and they struggled to beat TCU, a team the Horns handled with ease. And there’s also Kansas. Whatever, right? Who loses to Kansas?
So, 4-0, and then it’s on to JerryWorld and the Big 12 Championship Game, where the Horns would face a team they’d already beaten, most likely Oklahoma or West Virginia. Chalk up another win, and the 11-2 Horns are just outside the playoff and headed for a plum of a bowl game.
Not so fast, says Bill Connelly, Podcast Ain’t Played Nobody co-host, author, and college sports editor and analytics director for SB Nation. Connelly is the creator of the increasingly popular S&P+ computer rankings system, and man, has that algorithm ever been down on the Horns this season.
We’ll get to that in a second, but first a little explanation for what S&P+ is and is not. Connelly stresses that is not a yardstick for what a team has done in terms of wins and losses, but what it can be expected to do with the rest of its games. Every play of every game goes into these rankings, which are built on Connelly’s five key factors: efficiency, explosiveness, field position, finishing drives, and turnovers.
Once all those numbers are crunched, each team is assigned an S&P+ number that reflects how many points above or below it would perform in a theoretical game on a neutral site against a perfectly average college football team. As of now, Alabama would beat Generic U at Blah Stadium in Anytown, USA by 26.3 points, while UConn, the worst team in FBS according to Connelly, would lose to Generic U by 21.7. (And Alabama would thrash UConn by 48, which is probably conservative, but whatever.)
Space doesn’t allow for a deep dive into Connelly’s system here; suffice to say that the numbers are intended to factor out old-school emotional thinking and the randomness of lucky bounces.
And it’s scary effective.
Consider this: thanks to the bookie’s take, you have to hit on 52.4 percent of your bets to turn a profit. Once you move that number up to 55 percent, you are in high cotton. If you can do that consistently, you could make a living picking games. Anything above that is considered otherworldly.
This year, Connelly’s number is 55.3 percent against the spread, and 53.4 percent successful with over/unders. That’s picking all the games, not a short list of “locks” some handicappers will try to sell you. As he not so modestly pointed out earlier this week on his podcast, out of dozens of competing computer systems and handicappers, S&P+ is second only to the “Pi Rate” method this year in predicting the outcomes of all college football games this year.
So, what does all of that mean for Texas? Thanks largely to a robust estimate of the Horns recent recruiting classes, the Horns entered the season ranked twenty-seventh according to S&P+, versus fifteenth in the AP poll. After laying an egg against Maryland, the Horns—who have dropped out of and subsequently reentered the AP poll—haven’t risen as high as twenty-seventh again all season in Connelly’s ranking.
That has incensed and bewildered UT fans as the Longhorns—until last week—seemed to be turning it around. How could a team ranked sixth in the nation by the AP poll not even crack Connelly’s top 25?
Well, there was that loss to Maryland. Per S&P+, Texas should have beaten the Terps by twelve points. The next week, 50th-ranked Texas squeaked by 96th-ranked Tulsa at home, and at that point in the season, with games against USC, TCU, Kansas State, and Oklahoma looming ahead, a 1-5 start to the season seemed possible.
After the wins over USC (who’d already lost to a Stanford team that has underperformed both Connelly and the voters all year long) and TCU (ditto), the Horns were only ranked 35th in S&P+. (By this point, Connelly had Texas, TCU and USC all ranked more or less as equals, despite the Horns’ convincing wins over each of the others. Grumbling from the fan base commenced.)
The algorithm insisted that Texas should have beaten 89th ranked Kansas State by 16 points, but the Wildcats scored 14 unanswered second-half points to all-but-overcome a 19-point halftime deficit. Since computers don’t care whether or not a team had not won in Manhattan, Kansas, since 2002, the Horns rose only five notches in the rankings. (And K-State, like USC and TCU, has not played up to preseason expectations.)
And then came the Red River Showdown, won by the Longhorns 48-45. Connelly’s computers spat out a big fat “meh.” Oklahoma remained right where they were, at number six in the rankings, while the Horns actually fell eight spots to number 38. By this point close observers of S&P+ were either baffled or irate. Connelly tried to head off the burnt orange mob with this preliminary explanation:
The Horns had a 9 percent post-game win expectancy against the Sooners, getting out-gained by 2.5 yards per play and allowing a 61 percent success rate. But that one’s at least easier to understand: turnovers, turnovers, turnovers.
Texas had a plus-3 turnover margin when stats suggest about a plus-1.4 margin was more likely. Turnovers are worth about five points in field position lost and gained, so you could say the god of randomness gifted the Horns about eight points in a three-point win.
Or, in layman’s terms, it was a fluke. That did little to silence outraged Horn fans, who took to Twitter to savage both S&P+ as an algorithm and occasionally, Connelly personally. Later that week, Connelly poked the bear by pointing out that Texas was just a weird team all around—they outscored top quality opponents by two touchdowns, while squeaking past bottom 80 teams by little more than a field goal, if they won at all.
The Horns seemed to defy his system. Or, more specifically, Tom Herman did. How one coach can play up to superior competition and down to shoddy teams, year after year, at two different schools? There’s no number you can put on that, but it’s definitely a pattern at this point.
Before the Oklahoma State loss, Connelly wrote that Texas’s season record versus his algorithm’s projections “only makes sense when you realize it is a Herman team.” He laid out Herman’s two up-and-down years at UH— three triumphs over ten-win opponents (including Florida State), another over Lamar Jackson’s Louisville and a loss to 5-5 UConn. (Final S&P+ ranking: 46th, versus number 8 in the AP poll.) The following year, Herman’s Cougars bounded out of the gates with a ten-point season-opening win over third-ranked Oklahoma. Much later in the season, they would dismantle another third-ranked team—another Lamar Jackson-led Louisville bunch—by the score of 36-10.
But those upsets sandwiched losses to Navy and a 22-point drubbing at the hands of lowly SMU, a team Herman had personally taunted a few months before. Despite twice beating the number three team in the nation, and rising as high as sixth in the AP poll, the 2016 Cougars ended up unranked. S&P+ pegged them at number 39, oddly enough, two slots behind the 5-7 Longhorns of Charlie Strong, the man Herman had replaced at Texas before bowl season was over.
“This is the Herman m.o. at this point,” Connelly explained. “You will always get the best version of his team in big games, and that team will turn around and show you as little as possible against everyone else. It makes them almost impossible to project or to trust….S&P+ is basically throwing up its hands at this point.”
Leading up to the Oklahoma State game, Vegas had the Horns as 3.5 favorites, even on the road at night in a homecoming environment. You could see the bookies’ angle— the Pokes reeled into this contest on a three-game losing streak, while Texas was riding high on six straight wins.
But Connelly’s computers saw the Pokes as the superior team. S&P+ predicted a 33-26 win for Oklahoma State, which was (almost) right on the money: he not only beat the spread and predicted the upset, but came reasonably close to forecasting the point totals for each team in the 38-35 win as well.
As of right now, well, don’t click on this week’s rankings if you’re a Longhorns fan. Not only you will be treated to a banner pic of a victorious Oklahoma State lineman flashing the double Horns down, but you will also find Oklahoma still perched in prime position while the Horns are languishing at 42nd, just ahead of USC and TCU, the two paper tigers that lit the kindling under the “Texas is back!” fire.
Herman’s team are once again right behind a Charlie Strong squad—Strong’s previously undefeated (and overrated) South Florida Bulls had been winning even uglier than UT all year long, until they ran into Major Applewhite’s Houston Cougars, who handed them their first loss with a 57-36 shellacking.
And right now, per S&P+, those 7-1 Cougars come in at number 25, the top-ranked team in Texas. They are just a notch ahead of Texas A&M, who can console themselves with the fact that S&P+ ranks the Horns fifth in the state, behind both those two and North Texas and Texas Tech, in that order.
That’s not a typo. Millions of data points have been fed into a really smart guy’s computer and more than halfway through the season, it sees the Mean Green as the Longhorns’ betters.
And there are still some really confounding teams ranked above the Horns. How is 4-4 Missouri ranked coming in at number 27? Is Connelly putting his thumb on the scale for his alma mater? Well, no. According to Connelly’s metrics, with a little luck, the Tigers should be 5-3. And their schedule has been among the toughest in the nation, whether measured by humans or computers. They lost respectably to top-ranked Alabama (only by 29) and fifth-ranked Georgia (by 14). A solid South Carolina and a good Kentucky beat the Tigers by a total of three points. And unlike Texas, Missouri has mauled lesser opponents, and following Purdue’s manhandling of Ohio State, Missouri’s close win over the Boilermakers carries more weight now than it once did. (As for how a 4-4 Memphis team somehow comes in ahead of Texas, I can’t really say. As the games play out, it will almost certainly make more sense.)
Raw S&P+ numbers have Texas as an underdog to West Virginia, Iowa State, and Texas Tech, but Connelly’s latest projected season wins graph has Texas splitting its final four games and finishing 8-4. The losses would come at the hands of West Virginia on Saturday and to Tech later in Lubbock, while per Connelly’s system Iowa State’s slight S&P+ edge is negated by Texas’s home-field advantage. Kansas is close to a slam-dunk per the numbers.
On the other hand, maybe Texas will beat everybody. Except Kansas. After all, this a Tom Herman-coached team.