It was a tough week for the Texas Tech basketball team, as they dropped what seemed to be a winnable game against Iowa State at home and lost to a sharp-shooting Baylor squad in Waco. The Texas Tech offense was particularly bad last week with turnovers, cold shooting spells, and missed free throws ultimately costing them a chance to be 5-1 or 6-0 in Big 12 play at the one-third mark of the conference season.
Let’s look at what the numbers say, starting with how Big 12 teams rank nationally in BPI:
For the first time this season, the conference looks to be distinguishable into tiers of teams. Texas Tech, Kansas, and Iowa State compose the top tier with razor thin margins between each of them (which makes last week’s loss to the Cyclones even tougher to swallow).
TCU, Oklahoma, Texas, Kansas State, and Baylor are the middle tier of teams, all within 11 spots of each other nationally in BPI. West Virginia is a clear ninth, and Oklahoma State the clear bottom team in the conference. Having said all that, I still believe this is the type of conference where anybody can lose on an opponent’s home court.
Surprisingly, Texas Tech’s individual game outlook for the rest of the season hasn’t moved a ton from last week, except for the road matchup in Manhattan tonight. The Red Raiders’ probability of winning tonight took the biggest dip of all the remaining games on the schedule. Most other games didn’t move more than a couple of percentage points.
|Opponent||Chance to Win (%)|
|@ Kansas State||57.1|
|@ Oklahoma State||74.3|
|@ Iowa State||33.5|
Totaling all of this up, the average remaining win probability for Texas Tech is about 66 percent. Here’s how that shakes out in terms of final conference record probabilities (as always, omitting records that have less than a 1 percent chance of happening):
|Big 12 Record||Probability (%)|
My Two Cents…
Here is where I’ll dive into a little commentary and stray from the numbers a bit. When Texas Tech was 4-0 last week, you may remember their chances to win the Big 12 per BPI were roughly 75 percent. As some tend to do with probability, many fans mistook that number of greater than 50 percent for the equivalent of 100 percent. Thus, they’re understandably disappointed at losing two straight games and the odds of a conference title dropping significantly.
Let’s back up to more of a 30,000-foot view of things. Before the season started, Texas Tech’s average Big 12 win probability was just shy of 69 percent. A 4-2 record at this point was the most likely outcome through the first six games. The remaining schedule’s average win probability is 66 percent, projecting to the most likely outcomes of 11, 12, or 13 wins in conference play.
In other words, we’re right back to where we started at the beginning of the year. The model has been pretty darn accurate through six games. Will it remain that way through the final 12 games of the conference slate? I don’t know. That’s the beauty of sports; the unpredictable can always happen, which is why we let them play the games instead of just reading boring spreadsheets.
Important Five-Game Stretch
The next five games make up an important stretch. BPI tells us that Texas Tech “should” win three out of the next five, on average. If they only win two out of the next five, they’d be sitting at 6-5 in conference play and probably out of the Big 12 championship hunt barring something miraculous. If they win four out of the next five somehow, 8-3 likely puts Texas Tech in position to make a serious run down the stretch – especially with a home game against Kansas yet to be played at that point.
Texas Tech might win the Big 12. But it’s more likely that they won’t win it, just as it was before any games were played this year. But they didn’t become a terrible basketball team overnight because of two frustrating losses. Texas Tech dropped three out of four conference games during one stretch in January last season, and they were still very much in the Big 12 title race down the stretch (and would have won it if Keenan’s foot didn’t break). Last year’s team improved from January to March and turned that into a historic tournament run. I’m not giving up on Chris Beard and the gang just yet.
- Texas Tech still leads the country in defensive BPI and KenPom adjusted defensive efficiency.
- Texas Tech is 15-0 when giving up less than 68 points, and 0-3 when giving up 68 or more.
- According to The Prediction Tracker, the opening line in Las Vegas is picking winners this college basketball season at a rate of 76.4 percent. No publicly available model has beaten that. A guy named Erik Forseth has created a model that has predicted winners at a rate of 76.3 percent. The BPI model is close behind at 75.8 percent, but BPI performs better than Forseth’s model against the spread. I’ve done a poor job of tracking the accuracy of different models throughout the season, but I believe ERFunction surpassing BPI in straight up accuracy has been a relatively recent development. Both Forseth and BPI outpace Sagarin and other popular public models, even if it is by a slim margin. Anyway, just thought I’d give y’all some insight to BPI’s overall accuracy and why I reference it.