The impact of momentum on tourney performance is a tricky thing

Stop me if you’ve heard this one before: “To succeed in March, you need to be hot coming into the tourney.” College basketball analysts say the darnedest things. The problem is that most of them are either half-truths…or patently false.

The value of pre-tournament momentum isn’t as clear-cut as the pundits would lead you to believe. For one thing, it’s hard to define who’s hot and who’s not coming into the dance. For another, the dynamics of the tournament have changed in one very significant way over the last decade. And finally, momentum can have a bigger or smaller impact depending on the nature of the team. Let me explain.

There are a couple of ways I measure momentum in my Bracket Science database. For all 1792 teams in the 64-team bracket era, I’ve tracked the number of wins tallied in the last ten pre-tourney games. This includes games played in conference tournaments. The other way I measure momentum is by tracking each team’s winning streak coming into the dance.

If you just look at the overall numbers for “wins in last 10,” you’ll see that the impact of momentum on tourney performance is mixed. I calculated the performance against seed expectations (PASE) of one through twelve seeds with succeeding numbers of wins in their last 10 games, and here’s what I found:

  • Fewer than six wins in last 10: +.072 PASE
  • More than five wins in last 10: -.003 PASE
  • More than six wins in last 10: +.004 PASE
  • More than seven wins in last 10: +.021 PASE
  • More than eight wins in last 10: +.091 PASE
  • Undefeated in last 10: +.015

The highest class of overachievers is teams with 9-10 wins in their last 10 pre-tourney tilts (+.091). But they don’t beat seed expectations that much more than the coldest teams with fewer than six wins in their last ten (+.072). And every other class plays pretty much to seed expectations. That includes the hottest teams, the ones undefeated in their last 10 dance tune-ups (just +.015).

Of course, looking at the overall “last 10” numbers masks an important distinction between teams. The numbers for Big Six conference teams (ACC, Big 10, Big 12, Big East, Pac-10 and SEC) are significantly different than those of Mid-Major and Small conference teams. Just look at this chart:

13_momentum_last10

The impact of momentum for Mid-Majors and Smalls is much less dramatic than it is for Big Six teams. And it’s a little more intuitive. While the hotter teams don’t play that much better than seed expectations, the colder ones are certainly the biggest underachievers. Not so with Big Six teams. Squads from the Power conferences that limp into the dance beat expectations at a +.101 PASE rate. By my way of thinking, these teams play in tough conferences with some of the best teams in the country, so four or five wins in the final 10 games may not necessarily be a sign that they’re playing bad basketball. On the other hand, the biggest overachievers by far are Big Six teams with nine or ten wins. That would seem to support the notion that hot teams are the best tourney performers.

Now let’s look at the impact of winning streaks on performance in the dance. Here, the numbers are somewhat similar to the chart we just saw. I broke down Big Six and Mid-Majors/Smalls into five streak categories: a losing streak of two or more games, a single loss before the dance, a winning streak between one and three games, a streak between four and seven, and a streak of eight or more wins. Here’s how the PASE numbers broke down:

13_momentum_streak

Just as with “wins in last 10,” the worst Mid-Major and Small conference teams are the coldest heading into the dance. If you have any sort of losing streak heading into the tourney, you’re not likely to live up to seed expectations. But being hot doesn’t seem to confer particular advantage. In fact, teams with 4-7 straight wins are actually slight underachievers.

The numbers are much different for Big Six teams. Teams with a losing streak of two or more beat expectations, as do teams with a single loss. But think about it: most Big Six teams come into the dance on the heels of one loss—the ousting from their conference tourney. There’s no shame in that…and it certainly doesn’t suggest that the team is cold. What’s interesting though is that teams with 1-3 straight wins are the biggest underachievers (-.195 PASE) among Power schools. What gives? Well, many of these teams won those three games in their conference tourney. So they peaked after the regular season. But this short spurt of success didn’t carry over onto the national stage. Curious. It’s the 4-7 win streak teams that are the big overachievers (+.194 PASE). They carried strong late regular season play through their conference tourney and into the national dance. It seems, however, that too long a win streak might subject Big Six to the law of averages. They’re pretty significant underperformers (-.137 PASE).

The win streak numbers for Big Six teams mask a big change that occurred in 2002. That’s the year that the Big Ten and Pac-10 joined the rest of the Power conferences in having a post-season tourney. The numbers before and after this year are markedly different. Take a look:

13_momentum_post2002

Before every Big Six conference had their own tourney, coming to the national dance with a losing streak wasn’t necessarily a sign of an underachiever—probably more an indication of a tough late-season schedule. Now, Big Six teams with two or more losses are massive underachievers. (Single-loss teams still play pretty much to expectations.) And now every Power team that has 1-3 straight victories did so by getting hot in their conference tournament. But those 18 teams are big-time underperformers in the dance. And the 4-7 win Power teams—the ones that were hot going into their conference tourney and carried that momentum through—are even bigger overachievers. In fact, they’re the squads to keep an eye on.

So what’s the bottom line on the impact of momentum? While hot Mid-Majors and Smalls don’t necessarily beat seed expectations, hot Big Six teams are certainly overperformers. But it isn’t enough for a Power conference team just to get hot in time to win their tourney (think Syracuse in 2005 or Vanderbilt last year); they need to have carried their momentum from late in the regular season through their conference tournament. In short, be selective in the teams you credit for having momentum.

This entry was posted in Tourney Trends. Bookmark the permalink.

5 Responses to The impact of momentum on tourney performance is a tricky thing

  1. Matt says:

    Any thought on the role of the timing of the conference tournaments as well as when they start NCAA Tournament play? Most of the Big Six Conference Tournaments end on the Saturday and Sunday prior to the NCAA tournament, whereas the Mid-major and small conference tourneys end earlier. Does the longer time off between the end of the conference tournament and the start of the NCAA perhaps contribute somewhat to tempering the momentum of those smaller schools?

    Similarly, do we see a difference in performance for Big Six conference winners if they play opening games on Thursday vs. Friday of the NCAA Tournament? The Big Ten usually is the last conference tournament to finish. Do those winners do better if they have that extra day of rest by starting Friday, vs., say, having the opening game on Thursday?

    • ptiernan says:

      Matt, Unfortunately my data doesn’t track this. It would be interesting to measure though. I’ve often wondered if teams that fall early in their conference tourney do better in the national dance because they can rest up. And I wonder about all those teams that play on Sunday. It’s not just Big Six conferences either. A-10 and others finish on Selection Sunday.

  2. Andy M says:

    Hey Pete,

    Can you still be contacted at ptiernan@bracketscience.com?

    Stoked for another year of bracketscience,

    andy

  3. P.H. says:

    Awesome work. I’ve tried to use streaks and such in my pool picks without much luck. This breakdown is exactly what I needed to be thinking all along!

Leave a Reply

Your email address will not be published. Required fields are marked *

* Copy This Password *

* Type Or Paste Password Here *

You may use these HTML tags and attributes: <a href="" title=""> <abbr title=""> <acronym title=""> <b> <blockquote cite=""> <cite> <code> <del datetime=""> <em> <i> <q cite=""> <strike> <strong>