The Extremes: Can Louisville or UCLA win a title?

December 10, 2016; Los Angeles, CA, USA; UCLA Bruins guard Lonzo Ball (2) moves in to recover the ball after forcing a turnover against the Michigan Wolverines during the first half at Pauley Pavilion. Mandatory Credit: Gary A. Vasquez-USA TODAY Sports
December 10, 2016; Los Angeles, CA, USA; UCLA Bruins guard Lonzo Ball (2) moves in to recover the ball after forcing a turnover against the Michigan Wolverines during the first half at Pauley Pavilion. Mandatory Credit: Gary A. Vasquez-USA TODAY Sports /
facebooktwitterreddit

Taking a closer look at how far your team can go when it dominates on just one side of the floor.

We’re at that point of the season where everyone can start to smell March Madness. Water cooler and Twitter spats break out over ‘quality wins,’ Power 5 fans call for an undefeated Gonzaga as a 4-seed, and the cottage industry of bracketology starts to ratchet up. Even the NCAA is getting in on the fun this year, looking to cash in on our thirst for rampant speculation.

Here, we’ll take the chance to fuel some of that speculation — with cold, hard, stats. This article takes a look at the postseason chances of teams that live on the extremes — those teams with great offenses, or defenses, that can’t quite match them on the opposite side — and how they’ve fared historically in the tournament.¹

Great Offenses

First, a look at teams with great offenses and suspect-at-best defenses. It’ll include a team considered a Final Four contender and one that’d be happy just to get to the second weekend.

UCLA (1st in offensive efficiency, 128th in defensive efficiency)

UCLA’s loss to fellow top-15 team Arizona this past Saturday took a bit of the shine off the Bruins’ strong start. Fueled by the stylings of superstar freshman Lonzo Ball, UCLA charged to a 19-1 start and a top-3 spot in the AP poll. The Bruins are one of the most fun teams in college basketball to watch, playing at a breakneck, hyper-efficient pace on offense while taking and making more 3s than they have in years.

The glow surrounding this team — with the program history, return to relevance, and enjoyable style — has really obscured just how poorly they’ve played on defense. The team is currently ranked as the 125th best adjusted defense in the country, per KenPom, at a time when preseason rankings have fully left the calculation. Their defense has just been bad this year.

And yet, they’re still 19-2 overall and just two games off the pace set by Oregon and Arizona in the Pac-12. Joe Lunardi currently has them as a 3-seed, and it’s tough to see them falling too much farther, with their combination of name recognition, an early season win over Kentucky, and a Pac-12 schedule they’ll probably blow through outside Oregon/Arizona. This is still a team with the nation’s best offense by a wide margin, and just two losses.

Marquette  (8th in offensive efficiency, 131st in defensive efficiency)

While Marquette is more on the bubble than a national title contender — Joey Brackets had them as one of his last four in, as of this writing — Tuesday night’s comeback win over No. 1 Villanova adds a true quality win to their resume, and pushes them closer to a lock.

What a thing perception is, though. Marquette, like UCLA, currently sports a top 10 offense and a defensive efficiency near 125th (they’re 131st). Despite an efficiency profile much closer to UCLA than most Bruins’ fans would probably like to admit, Marquette was unable to cash in on its best opportunities early in the season, and currently sits at 14-6. They’ll likely end up somewhere in the 7 to 12 seed range, barring a series of late season wins over Butler or Xavier.

Granted, Marquette does it with talent that doesn’t quite compare to that on-hand at UCLA, but they get results just the same. Mainly fueled by volume 3-point shooting at a 41.9 percent clip (third-best in the country), Marquette also sports solid inside play from Luke Fischer, and hits their 2-pointers at a rate good for 26th-best in the country. The defense has, obviously, not been great — which has hurt them in many of their contests against top-level competition. Taking down No. 1 might signal a change in their luck there, though.

Outlook

So how can we expect these high-powered offenses to do in the Tournament? Let’s start by taking a look at how similar team archetypes have done in past tournaments. The table below summarizes the results of all teams, in the collected data, that went into the NCAA tournament with both a top 10 offense and a defense ranked outside the top 100.

CountAvg. SeedActual WinsExp. Wins
(Log 5)
DifferenceExp. Wins
(Seed)
Difference
185.442418.685.3222.271.73

Similar team archetypes have been pretty successful (relative to their pre-tournament efficiency, as well as seed) in the tournament within the last 15+ years. They’ve gone 17-4 as the favorite (efficiency-wise, not betting line), and 7-14 as the underdog — certainly outperforming objective expectations for the group.

A few notes of caution, however — most of the positive differential in the above table has been driven by Mike Brey’s Notre Dame in the last two years. As a 3-seed in 2015, and a 6-seed last year, Notre Dame made the Elite 8, despite not being ‘expected’ to make it past the first weekend in either case. And, perhaps worryingly for UCLA, not one team with this profile has advanced past the Elite 8. Of the teams that have earned a top seed (No. 3 or No. 4 — no similar team has been higher yet), three of seven failed to make it out of the first weekend, losing to ‘underdogs’ in each case.

There are some solid indicators a strong offense can carry you through the initial rounds – good news for Marquette. But things clearly begin to break down for weak defensive squads near the final eight — typically, once you get there, the teams you’re playing are good to great on both sides. Just two teams (of 27) with top 10 O’s and defensive efficiencies outside the top 50 (pre-tournament) have made the Final 4 since 2001-2002. Unless UCLA can clean things up on D – which is getting increasingly unlikely — it will be a risky venture to put them in your Final 4.

Great defenses

And not a look at the other side of the coin — the great defenses this year, who just don’t quite have the offense to match (yet).

Louisville (2nd in defensive efficiency, 34th in offensive efficiency)

I’ll be honest, the pitch for this article sounded a lot better before Louisville’s 106-51 dismantling of Pitt on Tuesday night, wherein they scored 1.47 points per possession and jumped from 45th in adjusted offensive efficiency to 35th. And all this without Quentin Snider, who’ll be gone for another week or two with a hip flexor injury. If he reintegrates well and Donovan Mitchell (29 points on 9-13 shooting) continues his recent tear, this could be a moot discussion by the tournament.

And yet, this is kind of who Louisville is. They’ve finished better than 45th in adjusted offensive efficiency just twice in the last six years, but haven’t seen a defense worse than fifth over the same span. They’re a consistently great defensive team that can struggle to score in the halfcourt.

While Snider, Mitchell, and the shift of low-usage all-star Anas Mahmoud into the starting lineup could signal continued improvement for the offense, it wouldn’t be surprising to see it finish around where it currently is — 30th to 40th best in the country. As it’s paired with a defense that’s over three points per possession better than third place Virginia, it’s not exactly a bad place to be.

South Carolina (1st in defensive efficiency, 156th in offensive efficiency)

South Carolina is truly at the extremes in their approach — which, granted, has led them to a surprise 16-4 record and a projected second place finish in the SEC. Currently the nation’s best defense (just edging out Louisville), their offense is just 156th in the country, scoring 105.3 points per 100 possessions.

Their approach is led by PJ Dozier and Sindarius Thornwell — a pair of guard/wings that absolutely hound opposing ballhandlers, then grab the ball and do everything for the team on the other end. Both sport steal rates over four percent — top 40 rates in the country — and offensive usage rates near 30 percent. The degree to which they rely on both was evident during Thornwell’s month-long suspension through December and January — the previously undefeated Gamecocks dropped games to Seton Hall, Clemson, and Memphis as they went 3-3.

Frank Martin’s team prides itself on its defensive effort. They do an excellent job of just about everything on defense outside of rebounding and fouling — they protect the perimeter, force tons of turnovers, and protect the rim well if penetration gets past the wing.

But the offense is…troubling at just a point better than the D-I average. Things may improve with Thornwell back, he’s their best offensive player, but it’s very unlikely they finish the year within the top 100. Lunardi currently has them around a No. 7 seed, which seems fair — despite the first-ranked defense, they’re just 26th in KenPom’s EM metric.

Outlook

Just how far can defense alone take you in the tournament? Let’s take a look.

The first table focuses on the Louisville archetype — offense outside the top 35, with a defense in the top 10.

CountAvg. SeedActual WinsExp. Wins
(Log 5)
DifferenceExp. Wins
(Seed)
Difference
636.176087.74-27.7479.95-19.95

Yikes. Defense might win championships, but these results suggest not bringing a very good offense to the table to complement it might be a problem. Teams with this efficiency profile have gone a poor 51-38 as a favorite, and just 9-25 as an underdog. As the table shows, similar teams underperform in both ‘expected’ wins categories by pretty significant margins.

There is hope for Louisville in that the only team to really succeed with this model is…Louisville. In 2008 (45th in offense), 2009 (54th), 2012 (126th), and 2015 (78th), the Cardinals performed really well in the tournament — combining for three Elite 8’s, and a Final 4. They stand as the only top-4 seed, outside of North Carolina in 2011, to make it past the Sweet 16 with a top 10 defense and an offense outside the top 35. Their defensive style is clearly very effective in the tournament, and they’ve been able to scrounge up enough offense to buck the noted trends consistently. Barring offensive improvement through the rest of the year, a similar run will be the hope for the Louisville faithful.

South Carolina is in another bracket altogether. Here’s a similar chart, except it’s top 10 defenses with offenses outside the top 100.

CountAvg. SeedActual WinsExp. Wins
(Log 5)
DifferenceExp. Wins
(Seed)
Difference
187.841117.40-6.4015.53-4.53

The sample is much smaller, but the results aren’t all that encouraging. Eleven of the 17 teams with a similar efficiency profile scored no wins in their NCAA appearances, and just two made it past opening weekend. While most of the teams are mid-seeds (No. 5 to No. 11), and therefore facing tougher first round opposition than the protected seeds, this archetype has gone just 7-8 as the ‘log5’-EM based favorite, and lost 10 of their 14 games as the underdog.

Next: College Basketball Power Rankings: Week 12

The best case scenario is, of course, Louisville — the 2012 team ranked 126th in offensive efficiency, but made a run to the Final 4. Most teams that look like South Carolina does this year crash and burn early in the tournament, though. Even penciling them in for the Sweet 16 is historically risky.

Clearly, it’s best to just be good at everything but obviously, not everyone can do that. The teams this year that heavily depend on offense or defense certainly have the tools to make deep runs in the NCAA tournament, but history is decidedly against all of them (besides Louisville) making it much farther than the Sweet 16.


¹To accomplish this, I gathered pre-tournament KenPom efficiencies and stats for all available seasons – stretching back to the 2001-2002 season, as well as tournament results and seeds. ‘Expected’ wins for a team were calculated in two ways. First, a simple method that simply averages how many games, on average, the team’s seed (1-16) won over that span. Second – I ran 10,000 simulations for each tournament based on log5 predictions using each team’s pre-tournament efficiencies, and averaged out how many games a team won.