Step Away From the Ledge

One of the things that's hard for fans to do is see the forest from the trees. Every game feels like the Championship of the Free World. Losses feel terrible, and when the losses pile up, it gets hard to see the big picture.

In an effort to see where UConn stands in the world of college hoops, I picked 12 “predictive” computer ratings, most of which you should recognize, but a few random ones (without any research to their methods) just to get more diversity. Through Sunday’s games, the consensus is that the Huskies are one of the best 30ish teams in the country:

Is being one of the 30 best teams good enough? It depends on your expectations. Considering this is a team that, a year ago, lost in the first round of the NIT, I think a reasonable expectation this season would have been one of the nation’s best 20-25 teams – potentially top 15 with some luck. So I think it’s fair to say UConn has been a little disappointing. But keep in mind Amida Brimah missed 11 games with a broken finger, and UConn unquestionably suffered.

Looking at the big picture, let’s not lose sight of the fact that the Huskies – despite how they’ve looked at times – are a pretty good team, if not great, and have more than a reasonable chance of getting into the NCAA Tournament. Team Rankings estimates an 88 percent chance, and they’re listed on more than 90 percent of brackets at The Bracket Matrix.

But really what I wanted to do was talk about the past two games – both late collapses after UConn seemingly had the game in hand. Last Thursday, the Huskies were outscored 18-2 over the final 5 minutes of a 63-58 loss at Temple. They won the next one over Tulsa, but nearly blew a 20-point lead in the process, leading to even more frustration.

I think most sports fans inherently understand the law of large numbers, but the basic idea is: The larger your sample of trials, the more “true” the average will be. In basketball, every possession is an opportunity to score and prevent the other team from scoring. As these possessions add up, we get a truer and truer picture of a team’s quality. Using box score data, we can estimate that the Huskies have had 1,682 possessions this season – a good amount, though certainly not large enough to be considered perfectly correct.

But the point here is that the collapses were two very small samples of a larger body of work. The 18-2 collapse at Temple took just 8 possessions. If we pull it out a bit further and use the final 7:48 of the game, Temple closed with a 23-8 run over 14 possessions.

Yes, the Huskies lost. Yes, these were a bad 14 possessions, but that was just 24 percent of the game. For the other 76 percent, the Huskies were really good! Right before Temple got hot, UConn was on a 21-9 run of its own. It built a double-digit lead. This should count for more than zero.

The Huskies won against Tulsa, so it’s easier to swallow, but from an analysis perspective, that game was more troubling. Leading 48-29 with 15 minutes to go, UConn proceeded to allow Tulsa to score 1.69 points per trip over the next 26 possessions. This wasn’t just the other team getting hot for a few minutes, this was 42 percent of the game. But in the other 58 percent, UConn was really good! It held Tulsa to 0.78 points a trip. It built a 20-point lead. This should count for more than zero.

Combining the two games, that gives us 40 possessions that, on the aggregate, UConn was awful in, but 80 where it was really good. Taking it further, the two bad stretches combine for only 2 percent of UConn’s season. We have lots of other data (1,642 possessions), much of which is good! At times, they’ve even been excellent. This should count for more than zero.

After the past two games, it’s easy to say that these collapses have become a trend. After all, it’s two games in a row, and the Huskies have done it a few other times this season. But there are also a lot of games that show the opposite. UConn led by 18 against ECU with 10 minutes to go… and won by 18. One game before that, UConn made Memphis look awful down the stretch, expanding an 11-point lead to a 20-point win in the final 10 minutes. Going back further, against Georgetown, the Huskies closed on a 12-2 run (7-0 over the final 6 minutes) to win 68-62. They outscored Texas 15-13 over the final 7 minutes in Austin.

Early in the season, it was the opposite complaint. It wasn’t that the Huskies couldn’t close, it was that they were bad at the start of games and kept falling behind. But I argued then– as I do now – that this was just random and that “struggling in the first half” was not an actual problem. As the season went on, these first-half struggles have dissipated.

Similarly, I don’t think these recent 40 possessions are a symptom of a larger problem. It’s probably just random. UConn is not the very best team in the world and, like every other team on the planet, will struggle at times. Sometimes it happens at the beginning, sometimes at the end.

What we can say with absolute certainty is that the Huskies, on the whole, haven’t been very good offensively. Instead of talking about “heart” and “leadership,” we can actually look at the data and see inconsistent point guard play and Daniel Hamilton’s shooting woes. (Of course, there have been signs of life lately, so we’ll see.)

UConn probably won’t win the national title. In fact, we could say that even if they were undefeated. But depending on how the bracket shakes out, the Huskies could have a reasonable chance at making the Sweet 16, a puncher’s chance at a Final Four, and a nonzero chance at a title in a year with no great teams. Given where this team stood last year – and the stockpile of talent Kevin Ollie is bringing in the next few years – this has been far from a disaster of a season.