Analytics vs. Experts: Teams to Watch The Projections Disagree About

Posted by Will Ezekowitz on November 13th, 2018

As the college basketball season gets under way, everyone wants to get a read on their favorite teams. Does their on-court product match the preseason hype? For some teams, it depends on which preseason hype you considered. College basketball fans can seek either scouting/film-driven projections or analytics/model-driven projections. These two approaches usually tend toward some overlap, but not in every case. I have documented some of the bigger disagreements among top 50 teams below, and it will be interesting to watch which interpretation is borne out during the regular season.

NOTE: For the purposes of this article “Analytics” are mainly KenPom, BartTorvik, John Gasaway and some HoopLens and HoopMath thrown in. The “Experts” will be: the Top 25 polls, preseason conference polls, Matt Norlander’s #1-#353 rankings, and preview articles from NBCSports, ESPN, The Athletic and others, although many of those ranking systems also consider analytics as part of their methodologies.

Group 1: Loved by the Analytics, Hated by the Experts

Miami— KenPom: #20. AP Top 25: #33 (according to “Others Receiving Votes” section)

Jim Larranaga Wonders Why the Polls Don’t Like His Team (USA Today Images)

  • Summary—lots of talent leaving, but hyper-efficient role players taking their place.
  • What the experts say: Bruce Brown and Lonnie Walker are playing in the NBA, Ja’Quan Newton is also gone, and there are no impact newcomers other than graduate transfer Zach Johnson. This one seems simple — a #6 seed lost its three most talented players, so the Hurricanes will take a step back.
  • What the analytics say: Brown, Walker and Newton were all talented, but none of them were that successful last year. To replace them, Jim Larranaga has productive big man Dewan Hernandez (formerly Dewan Heull), a sophomore jump expected from Chris Lykes, and a bunch of efficient role players led by three-point gunner Dejan Vasiljevic. And if you think 20th from KenPom seems aggressive, BartTorvik currently lists the Hurricanes at #8!

CreightonKenPom: #35. AP Top 25: No votes received (52nd or higher)

  • Summary—Four starters are gone, but Martin Krampejl and Greg McDermott are still around.
  • What the experts say: Marcus Foster and Khyri Thomas are now playing professionally in addition to the losses of contributors Toby Hegner and Ronnie Harrell. Some feisty freshmen return, as does big man Krampejl from an ACL tear. However, as Matt Norlander notes while ranking the Bluejays 71st in his rankings, the fact that he’s their best player “slots Creighton as having the least impressive best player of any team in the Big East.” It’s rebuilding time in Omaha.
  • What the analytics say: Norlander should show some respect for Krampelj’s name! Per HoopLens, the Bluejays were an astonishing 0.15 points per possession better defensively when he was playing last year, which is the difference between the fourth- and 180th-best defenses in the country. On the other side of the ball, McDermott has guided his offense into the top 60 nationally every season since 2012. Put the two together, and Creighton would appear poised to make the NCAA Tournament for a third straight year.

Penn StateKenPom: #35. AP Top 25: No votes received (52nd or higher)

Read the rest of this entry »

Share this story

Projecting the 2018 NCAA Tournament Field With 2017 Data

Posted by Will Ezekowitz on January 2nd, 2018

The year 2017 is in the books and we are just beginning the conference season. With so much of a team’s “NCAA Tournament resume” still to be built, it seems like a fool’s errand to compare their relative achievements so far. All we know right now is a somewhat decent idea of which teams are actually good, but can we use resumes to date to project the future? I endeavored to figure just this out, by examining KenPom rankings at New Year’s Day with them to eventual NCAA Tournament seeding for the past five years. My conclusion indicates that, while surprises can certainly happen in the Tournament itself, teams are unlikely to significantly alter their regular season fortunes after the turn of the new year.

KenPom Rankings on New Year’s Day

 

For reference, here are the archived KenPom rankings as we head into 2018. They have been grouped into 20-team cohorts for convenience, and the group was limited to 60 teams because most teams ranked outsidethat range are mid-majors which will need to win their conference tournaments. The focus here will be on teams capable of building solid at-large resumes. Let’s start at the top.

Teams Ranked #1-#20 — 97% make the NCAA Tournament

This year: Villanova, Michigan State, Purdue, Duke, Virginia, Kansas, Texas Tech, Cincinnati, Gonzaga, West Virginia, Texas A&M, North Carolina, Wichita State, Arizona, Oklahoma, Xavier, Kentucky, Miami, Arizona State, Tennessee.

Read the rest of this entry »

Share this story

Three Lessons From PK80 Day One

Posted by RJ Abeytia on November 24th, 2017

My first day at the PK80 Tournament in Portland took place exclusively in the venerable Veterans Memorial Coliseum, where Bill Walton’s Blazers used to run roughshod, where the Showtime Lakers endured many a battle, and where Michael Jordan’s Bulls crushed the hopes of Clyde Drexler’s Blazers. It was amazing to watch a game in what was once considered a state-of-the-art NBA arena but now stands as a relic, but make no mistake: There were lessons to be learned with many future implications when it comes to the here and now in college basketball in The Rose City’s basketball nexus.

Duke is Led by Grayson Allen But Its Most Impressive Attributes are in the Frontcourt (USA Today Images)

  1. Duke’s Frontcourt is Massive. The physical realities of Marvin Bagley and Wendell Carter are by far the most impressive part of watching the Blue Devils play live. Yes, that size was accentuated by the lack thereof on the part of Portland State, but the two combined for 34 points on 13-of-20 shooting and 25 rebounds. Both are very athletic and graceful. Bagley even brought the ball up to help alleviate pressure in the backcourt several times. On the other hand, Bagley’s 6-of-12 from the free throw line certainly warrants monitoring and Grayson Allen’s emotional stability continues to be a coin flip from play to play, but if you are looking for reasons Duke can win the NCAA Tournament before December begins, look no further than the 6’11” 234-lb. Bagley and the 6’10” 259-lb. Carter. Duke isn’t going to face many teams (elite or Portland State-level) that can handle the inside talent the Blue Devils bring to the table.
  2. Shaka Smart is Building at Texas.  After a year two cratering that Smart warned Texas was part of the plan, the Longhorns notched a hard-earned win over mentally-taxing Butler on Thanksgiving. Texas is likely a year away from really competing on the national level, but the Longhorns showcased impressive perimeter talent like Andrew Jones and size from the likes of Mohamed Bamba. Jones had 16 points on efficient 7-of-13 shooting and Bamba logged 12 rebounds and six blocks. The Bulldogs were able to impose its standard low-possession game on Texas, but the Longhorns maximized their transition opportunities to the tune of a 14-2 fast break point advantage that provided the winning margin. Texas has the kind of balance and depth in the frontcourt that make for a very tough draw in Big 12 play and beyond. Assistant coaches scouting from the stands noted some of the finer points as well, like the Longhorns’ help discipline on defense. Texas is a team to watch moving forward, and their brawl with Duke today is a great early litmus test for both teams.
  3. Florida MOVES.  The #7 Gators demolished Stanford with a staggering barrage of 68 percent three-point shooting that featured a scorching 13-of-17 first-half start that included a perfect 5-of-5 from distance by Egor Koulechov. But again, the live impression may actually be more auspicious than the insane shooting performance. Florida rushes the ball upcourt like its hair is on fire. There was one possession where off a made basket, point guard Chris Chiozza already had the Gators in their offense with the shot clock at 29 seconds and an open three look at 26 seconds. Florida’s average possession time was 14 seconds (which KenPom rates as the 12th-fastest in the country) and its blistering 135.0 ORtg over its 80 possessions made for a painful clinic for Stanford. Identity matters in college basketball, and Michael White’s team has already clearly embraced theirs this season.
Share this story

The Models vs. the People: Who Is Right So Far?

Posted by William Ezekowitz on November 21st, 2017

With the rise of KenPom’s preseason rankings and the ratings of other models like it (SI and T-Rank, for example), projection models have become increasingly important in college basketball. But there is still a long way to go before these metrics-based systems replace the good old-fashioned eye test as represented in the national polls. The two varieties of projection mechanisms, both valid in their own right, disagreed about a few teams coming into this year. In this article, we will evaluate the differences on a few relevant teams to determine if we can settle on which method has been accurate so far. We’ll start by analyzing a couple of squads from the Big Ten before considering a couple others.

Minnesota. AP Rank: #15; KenPom Rank: #36

Jordan Murphy has helped Minnesota live up to expectations in the early season (Getty)

  • What the people thought: Minnesota spent the offseason as one of the most hyped teams in college basketball, as Nate Mason received plenty of all-Big Ten buzz and Amir Coffey appeared ready to make a huge leap. Richard Pitino’s Gophers were also expected to play their particular brand of stifling defense, bolstered by possibly the best shot blocker in college basketball, Reggie Lynch. There was a lot to like.
  • What the models saw: Neither Mason nor Coffey were especially efficient for the nation’s 77th-best offense, which meant this year’s outfit was set to improve on that end. The defense, while stifling, was below average in both turnovers forced and defensive rebounding, limiting its potential to become a top-10 unit.
  • Who has been right so far: The people. Jordan Murphy has been unexpectedly dominant through four games, putting up 23 points and 14 rebounds, for example, in a very impressive 12-point victory at Providence. The Gophers are humming along at 18th nationally in offensive efficiency, and if they can stay in that range they will certainly live up to their poll projection as the 15th-best team in the country.

Michigan State. AP Rank #2. KenPom Rank: #10

Read the rest of this entry »

Share this story

On Improving the NCAA Tourney: Part II

Posted by Brad Jenkins (@bradjenk) on March 31st, 2017

This is the second edition of our two-part look at ways to improve the overall structure of the NCAA Tournament. In our Part I piece published earlier this week, we made several recommendations concerning the selection and seeding of the 68 teams in the annual field. Today we will focus on improving the bracketing process. Over the years, the Selection Committee has endeavored to better balance the bracket while easing travel burdens on participating schools’ players and fans. Examples of these modifications include the current “pod” system for the first two rounds; ranking the four top overall seeds so that the top two teams are on opposite sides of the bracket; and, easing the restrictions on where schools from the same conference can be placed in the bracket. Since the last change was implemented, the committee claims that it has not needed to swap any teams off their true seed line. We think there is also an opportunity for further adjustments to that process, including one that is relatively simple and involves future tournament sites.

Black Stars – First/Second Round Sites (2015-17)
Blue Stars – Regional Sites (2015-17)

Go West Young Men — Just Not So Often

Our top suggestion deals with the issue of geographic balance. NCAA member schools’ presidents and chancellors have mandated that the Selection Committee consider traveling distance a priority when placing teams into each season’s bracket. But the committee is somewhat handicapped by the previously established locations that they have available for those placements. In the image above, we can see all of the regional sites for the last three NCAA Tournaments. In almost every year since the NCAA Tournament expanded to a full six-round event in 1985, the western region of the country has hosted two First/Second round sites plus a regional championship. To put it in mathematical terms, 25 percent of the NCAA Tournament prior to the Final Four is typically played in the Pacific and Mountain time zones. Yet, only 19 percent (six of 32) of the nation’s conferences are primarily located in those two zones — and that’s giving full western credit to the WAC even though three of its eight schools are actually in the Central time zone.

Similarly, only about 17 percent of all Division I schools reside in those two time zones. By moving one of those western sites to another part of the country each year, the NCAA Tournament would become much more representative of its membership in a geographic sense. The recommendation here is to only move one of the First/Second round locations out of the west — keeping the regional round located there for the sake of broader balance. As for where to relocate that site, the mid- to deep south needs more representation. Most of the major basketball conferences — the ACC, Big East, Big Ten and Big 12 — have plenty of viable nearby options in most years, but that’s not always the case for SEC schools. It’s not just about the big boys, though — there are a lot more mid-to-lower tier Division I schools that would have improved travel situations as well. There are plenty of options available in this region — cities with NBA arenas like New Orleans, Memphis, Atlanta, Nashville, Orlando and Miami.

Read the rest of this entry »

Share this story

On Improving the NCAA Tourney: Part I

Posted by Brad Jenkins (@bradjenk) on March 28th, 2017

Last June the National Association of Basketball Coaches (NABC) empaneled an ad hoc committee whose stated purpose was to provide the NCAA Division I Men’s Basketball Selection Committee a perspective from men’s basketball coaches and their teams regarding selection, seeding and bracketing for the NCAA Tournament. The NCAA has in recent years become increasingly receptive to considering and making changes, and as this year’s event reaches its climax, we decided to offer some specific recommendations to bolster the best three weeks in sports. Let’s focus today on improvements to the selection and seeding process.

John Calipari is one of the members of the recently created NABC ad hoc committee formed to make recommendations to the NCAA Selection Committee. (Kevin Jairaj / USA TODAY Sports)

John Calipari is one of the members of the recently created NABC ad hoc committee formed to make recommendations to the NCAA Selection Committee. (Kevin Jairaj/USA TODAY Sports)

Bye Bye, RPI

Whenever the subject arises of improving the primary metric that the Selection Committee uses, there is one recurring response: Either dump the RPI altogether, or dramatically limit its influence. The good news is that we may finally be headed in that direction. A month after the NABC formed its committee and began communicating with the NCAA, the following statement was made as part of an update on the current NCAA Selection Committee:

The basketball committee supported in concept revising the current ranking system utilized in the selection and seeding process, and will work collaboratively with select members of the NABC ad hoc group to study a potentially more effective composite ranking system for possible implementation no earlier than the 2017-18 season.

Moving away from the RPI as the primary method for sorting teams into composite tiers would be a huge step toward improving the balance of the field. We have heard committee members for years make the point that a school’s RPI ranking is just one factor of many on its resume. But then the same committee members turn right around and cite that team’s record against the top-50 or top-100 — or its strength of schedule rating — all of which, of course, are derivative of the RPI. That means that the outdated metric is still, even now in an environment of Big Data, a highly significant influence on how teams are judged. The real harm occurs when the RPI results in entire conferences being overrated, which leads to those member institutions likewise being over-seeded. Placing five to seven teams well above their proper seed lines can have a substantial negative impact on the overall balance and corresponding fairness of the entire NCAA Tournament. Here are three recent examples.

Read the rest of this entry »

Share this story

Beware the Polls: Overperforming Teams Regress in NCAA Tournament

Posted by Will Ezekowitz on March 14th, 2017

Preseason rankings may seem irrelevant when it comes to college basketball, but history has shown that their predictive value are surprisingly important. For example, as we have shown in this space before, the preseason poll is in fact just as predictive of a run to the Final Four as the current AP Poll. The reasoning behind this phenomenon is that the preseason polls take into account all of the changes that the mathematical models have more difficulty accounting for (i.e. coaching changes, big recruiting classes, injuries, etc.). The conventional wisdom entering the season is actually fairly sophisticated and smart.

Are You Sure You Want to Choose Michigan? (AP)

Given this intuition, let’s examine the teams that defied that wisdom throughout the season and went from unranked in the preseason to ranked in the final AP poll before the NCAA Tournament. Had the preseason polls gotten something wrong? Or did these over-performing teams regress back to their expectations in the postseason? Using the last 10 years of data that includes seed win expectation data from fivethirtyeight.com, I investigated. Of the 97 teams who qualified for analysis, 32 (33%) of those teams outperformed their win expectations in the NCAA Tournament while 65 (67%) underperformed. As a whole, a group that was expected to win 138.8 games over the 10-year sample won just 107. A statistical T-test found this difference statistically significant. In fact, only one team — Kemba Walker’s 2011 National Championship team, Connecticut — even made the Final Four.

Read the rest of this entry »

Share this story

Is the 2017 Bubble Really the Weakest in Years?

Posted by Shane McNichol on March 2nd, 2017

One of the prevailing narratives that has developed during the second half of this season is the existence of a historically weak crop of bubble teams. The bubble, by its very definition, is a fluid concept where a 68-team field consisting of 37 at-large teams necessarily limits the strength of the group. For whatever reason, though, this season’s bubble dwellers have earned a reputation as a particularly futile bunch. To explore the veracity of that claim, I reviewed the last seven NCAA Tournament bubbles (2011-17). This includes every NCAA Tournament since the 2011 implementation of the First Four, which added three additional teams to the at-large field. For this year’s bubble, I used ESPN bracketologist Joe Lunardi’s Last Four In and his first two out from the bracket released on Monday, February 27 — teams included were USC, Providence, Marquette, Vanderbilt, Georgia Tech and Wake Forest.

There are several clear takeaways here. First, the 2017 bubble does in fact feature the worst aggregate winning percentage and average RPI of the last seven years, along with the second-worst average KenPom ranking. In comparison with the last six years, this group of six bubble teams is statistically weaker than other years relative to the higher levels of automatic qualifiers. The most important finding, though, can be found in the far right column. This season’s bubble teams have all played very difficult schedules, nearly cutting the average bubble member’s strength of schedule rating in half. That’s notable because this season’s six bubble teams are from power conferences, while 19 of the 36 bubble teams from 2011-16 came from the mid-major world. That group included schools like Middle Tennessee, Tulsa, Colorado State, Iona, BYU (twice), Boise State (twice) and Oral Roberts.

Read the rest of this entry »

Share this story

How Tournament-Proof Are the Nation’s Top Five Offenses?

Posted by RJ Abeytia on February 18th, 2017

This year multiple coaches across the country have conceded publicly that a team’s offense is the biggest factor in its ability to maintain a defense. “Defense wins championships” may still be a treasured maxim, but the truth is that offense is the fuel in college basketball. The question then becomes one of how vulnerable the best offenses in college basketball are to a one-game slump? Since only a single bad night is all it takes to be sent home from the NCAA Tournament, it’s worth investigating the nation’s top five offenses to set some criteria for evaluating the rest of the field. Per KenPom, here are the top five offenses nationally based on adjusted offensive efficiency, along with their corresponding adjusted tempo.

Team Adj. ORtg Adj. Tempo
1. UCLA 124.5 14.1 (6)
2. Oklahoma State 123.9 16.5 (91)
3. North Carolina 122.2 15 (16)
4. Gonzaga 122.2 15.7 (33)
5. Villanova 121.7 18.8 (314)


As the tempo column shows, teams can play at both warp speed (UCLA, North Carolina, Gonzaga) or at a relative crawl (Villanova) and still be extremely effective. That said, to the extent that the game slows somewhat in the NCAA Tournament, it is reasonable to suggest that some of these teams may face more trouble than others. 
The Bruins, Tar Heels and Bulldogs all use a healthy dose of tempo when they play. This is not to say that any of those three teams cannot also win a low-possession game, but their opponents would certainly be better-suited to impose a slowdown game on them to the extent possible. Villanova has already proven its favored pace can win championships. The next question then becomes which of the faster teams are most poised to handle a grind-it-out half-court game?

Read the rest of this entry »

Share this story

Finding a Path For Four Teams on the Wrong Side of the Bubble

Posted by Shane McNichol on February 17th, 2017

The end of the regular season is creeping up on us. With just a little more than three weeks remaining until Selection Sunday, teams around the country are prepping for their final stretch of conference games. For schools sitting precariously on the bubble, chances to grab signature wins are dwindling. For teams on the outside looking in, the home stretch represents a do-or-die opportunity to make an at-large bid a reality. The four teams we examine today all have pathways to March Madness ahead of them, but nothing will come easy and their odds are diminishing daily.

Wake Forest

It’ll likely come right down to the wire for the Wake Forest faithful. Getty)

The Demon Deacons are 6-8 in ACC play but a losing conference record won’t be quite as damaging this season thanks to a soft mid-major bubble. Still, even if Wake manages to get to 8-10, some conference losing records will be viewed much better than others. The biggest issue for Danny Manning’s team is that it has yet to beat any top-tier ACC squads this season. The Deacs have a favorable RPI at #36, but close losses against Duke and Notre Dame simply aren’t enough. With four games remaining, Wake Forest must top Pittsburgh and Virginia Tech to remain on the bubble, but to truly make some waves, the Deacs need a win at Duke or against Louisville. Stealing either of those games would result in a 9-9 ACC record and a legitimate, perhaps even likely, chance at an at-large bid. Read the rest of this entry »

Share this story