Before I start looking at the impact of injured or suspended players, I want to talk a little about losing streaks. (This piggy-backs nicely on Ken Pomeroy’s recent writing on the predictive power of wins.) This weekend we heard a lot about Louisville’s three-game losing streak and Minnesota’s four-game losing streak.

The point I want to make is that not all losing streaks are created equally. When Illinois went on a recent losing streak against Minnesota, Wisconsin and Northwestern, John Groce’s team played distressingly poor basketball. They played basketball roughly equivalent to the 229th best team in the nation. That was the type of losing streak that correctly caused people to adjust their expectations. Even if Illinois had a few nice wins early in the year, they were NOT a Top-25 team.

But Minnesota and Louisville’s recent losing streaks have been much less distressing. While losing four in a row, Minnesota has still played roughly equivalent to the 32nd best team in the nation. And while Louisville has lost three in a row, the Cardinals have been roughly equivalent to the 55th best team in the nation. Both losing streaks could have easily been stopped with a few bounces. Had Minnesota’s Rodney Williams made a free throw in a one point loss, or had Georgetown’s seldom used Aaron Bowen not tipped in a circus shot against Louisville, we wouldn’t be talking about long losing streaks at all.

That’s not to say that those two teams are playing well right now. Both team’s offenses and defenses have fallen off in recent games. But neither team’s performance is remarkably distressing. In fact, I’m much more distressed by how Kentucky is playing in the SEC this season. The Wildcats have gone 4-2, but given how poor the SEC is this year, Kentucky has actually been playing worse basketball than Louisville during the losing streak.

*Does not include Sunday’s game.

Splits replicate the Adjusted Offense and Defense calculation on Kenpom.com which controls for opponent quality and venue. These measure how many points the team would score on a neutral floor against an average D1 team based on the team’s performance in the sample of games.

All this suggests that Minnesota and Louisville will be relevant at the end of the year, while I can’t be quite as certain about Kentucky.

For the record, I am a little nervous about Russ Smith’s play the last two games. Louisville’s national-player-of-the-year candidate has posted ORtgs below 100 in back-to-back games. (From the sideline, I can tell you Georgetown freshman D’Vauntes Smith-Rivera was very much bothering Russ Smith on Saturday.) Louisville hasn’t quite played a juggernaut schedule yet, and I’m worried if Smith might revert to his old inconsistency under the grind of Big East play.

Unfortunately, sometimes a breakout performance is just a hot-streak. As an example, look at Florida St.’s Michael Snaer. Snaer posted a career high 110 ORtg last year, and everyone thought he would be an ACC superstar this year. But Snaer has always had turnover problems prior to last season, and after a year of cutting down the turnovers, Snaer’s turnovers are back with a vengeance this season. Snaer’s ORtg has suffered because of it.

Similarly, Wisconsin's Ryan Evans is making last year look like the fluke. Here are Evan’s ORtgs over the last four years:

2009: 92

2010: 82

2011: 102

2012: 92

Sometimes when inefficient players suddenly look efficient, it really is just a temporary fluke. Russ Smith has clearly played brilliantly this season, but until we get to March, I am always going to wonder if the inconsistent Russ Smith, who falls in love with impossible shots, isn’t still around.

Of course, even if I believe in Minnesota and Louisville, these losing streaks will matter to the NCAA selection committee. And personally, I believe they should matter. A lot of people advocate for seeding the NCAA tournament based on team quality (read: Margin of Victory), not based on team accomplishment (read: Quality Wins). And I understand the arguments. Certainly, when you don’t seed by team quality, you run into situations where a 1-seed gets a horrible draw. And the NCAA committee is instructed to pick the BEST teams for the tournament.

But I view it this way. You can either win an NCAA title by over-achieving in the regular season or over-achieving in the post-season.

No one wins the NCAA tournament without performing above expectations. If you look pre-tournament, no team is favored to win more than four games. But every year someone wins 6 games and over-achieves. Similarly, some teams over-achieve in the regular season. They earn surprising wins and earn better seeds, even though they have to squeak by with a series of close wins. But why condemn over-achievement in the regular season when we don’t condemn over-achievement in the tournament?

Louisville and Minnesota are under-achieving. That just means they have a harder road to post-season glory. It doesn’t mean they are bad teams.

Injury Splits, Part 1

Over the next two days, I’m going to talk about where injuries or suspensions may be skewing our evaluation of various teams. I won’t be talking about all of this year’s critical injuries. For example, Wisconsin’s Josh Gasser, Missouri’s Michael Dixon, and Tennessee’s Jeronne Maymon have missed the entire season. While those injuries have clearly hurt their teams, because they didn’t play a minute this year, the Pomeroy and Sagarin ratings for those teams accurately reflect their future expectations.

But when players are out for a period of time (like Duke’s Ryan Kelly), it can take awhile for the rankings to catch up. Duke is now playing worse basketball, and we may want to look at how much worse the Blue Devils are playing without Kelly.

Of course not every player who is injured matters. For that reason I focus on players who play at least half their team’s minutes. And often we are looking at very small samples. Random noise may certainly explain some of the deviations from the norm. But I do think it is informative to look at how teams have performed without their missing stars.

A lot of smart folks have written about how Duke will be a much worse team without Ryan Kelly. I wasn’t quite as convinced because I happen to have high expectations for Duke’s Amile Jefferson. But through four games, the numbers are clearly worse without Kelly in the lineup. Duke’s offense has slid, and their defense has fallen off a cliff.

That’s probably too big a drop off to be permanent, and Duke’s horrific performance at Miami felt like a once-per-season collapse, not a permanent sign of bad things to come. But I think it is informative how Mike Krzyzewski is allocating playing time with Kelly out. While Amile Jefferson has seen his percentage of minutes increase from 21 percent to 58 percent in the four games Kelly has been out, the second biggest beneficiary of playing time is actually Mason Plumlee. And this worries me a little bit if I’m Duke. Plumlee has been playing 96 percent of Duke’s minutes since Ryan Kelly has been out, and Krzyzewski seems hesitant to ever take him out. I worry that all those minutes are having a negative impact on Plumlee’s energy level. Plumlee’s ORtg was 115 prior to Kelly going down, and has been just 95 in the four games since Kelly went down. Some of that is due to the tougher ACC defenses Plumlee has faced, but you have to wonder if the lack of rest time is hurting Plumlee’s overall performance.

However, I honestly think Duke may be better in the long-run if Kelly can come back. That’s because Jefferson has thrived as an offensive player now that he is getting more playing time. Jefferson has seen his ORtg increase from 104 before the Kelly injury to 119 after the injury. Now that he finally has a chance to get a rhythm in games, Jefferson is showing his offensive talents. If Jefferson can eventually improve his defense, his development will only be an asset for Duke come tournament time.

The second most important injury split in this column might be the Wyoming split listed above. Ever since Luke Martinez was suspended for his role in a bar room brawl, Wyoming’s offense has fallen off a cliff. You simply cannot lose such a potent scorer and expect to replace him with other rotation players. Larry Shyatt has done a masterful job keeping Wyoming playing elite defense without Martinez. And that defense will keep Wyoming competitive in the MWC this year. But Wyoming simply lacks offense without Martinez.

Lastly, Marquette’s offense has improved since Todd Mayo joined the team mid-semster. (Mayo was suspended for the first semester.) Mayo’s ORtg hasn’t been fantastic, but he has been aggressive and has attracted some offensive attention. The part I find more interesting is that Marquette’s defense has sagged since Mayo joined the team. In fact, in the last game against Providence, Mayo received an unusual DNP-Coach’s Decision.  Buzz Williams explained his decision not to play Mayo by saying he didn’t have anyone for Mayo to defend. But it certainly seemed puzzling for Mayo to go from over 20 minutes per game to riding the bench for a non-disciplinary reason.

Long Island was the media’s pick for the NEC title this year because they returned almost their entire rotation from last year’s tournament team. But Long Island’s season was off to a disappointing start, and then super-scorer Julian Boyd went down. And while a recent winning streak has improved expectations somewhat, this remains a heart-breaking season for fans of the team.

Former South Florida guard LaVonte Dority joined Valparaiso mid-season, and the aggressive offensive player has helped boost the team’s overall performance. He is attracting a ton of attention and making his teammates better.

Finally, Gregg Marshall has to be a national coach-of-the-year candidate. He lost virtually his entire rotation to graduation, but in November and December his team continued to play at a very high level. Then on Dec. 20th he faced a situation where three of his key rotation players were out. Carl Hall, Ron Baker, and Evan Wessel were all injured and missing in action. And yet since that time, Wichita St. has actually played better basketball. Carl Hall has returned for the last four of those games, but Gregg Marshall’s ability to find new players and stick them in the lineup has been nothing short of amazing.

Arguably, Missouri has never been at full strength (since Michael Dixon left the team), but for two games in December they had everyone else active. They crushed South Carolina St. by 50 and beat an Illinois team that was playing well at that time.

The rest of the season has seen key player’s missing and the team’s performance has suffered because of it. Bowers injury has clearly been the worst, but Missouri wasn’t exactly playing elite basketball before Jabari Brown became eligible either.

I’m not going to show the Kentucky or Louisville splits (since I discussed those teams at length earlier), but surprisingly, their struggles cannot really be tied to Willie Cauley-Stein’s injury, Ryan Harrow’s early absence, or Gorgui Dieng’s absence.

Click here for Part 2: James Southerland, Greg Whittington, Taylor Braun, CJ McCollum, Mike Moser, Kris Dunn and more.

Follow @DanHanner

The older I get, the more I see that one of the things I love most about sports is the variety of it, the diversity of it and the CHARACTERS. Men’s tennis is at its best in many years because, for the first time in a long time, the top three or four players all have wildly different styles. The Tim Tebow story was fun on so many levels, but one of those levels was that he was just SO DIFFERENT in how he played — I’d say we are entering a great time for quarterbacks, because Tom Brady and Aaron Rodgers and Eli Manning and Drew Brees and Michael Vick and Cam Newton and Tebow and others are not really alike at all.

-- Joe Posnanski

As a basketball fan, I’ve never understood the division that exists between fans of the NBA and the NCAA. While the NBA has the best basketball players in the world, March Madness is compelling in its own right and as entertaining as anything that happens on the professional level.

In the NBA, the owners of the 30 franchises consider turning a profit and getting an equal shot at the top players a right, regardless of how well (or how poorly) they run their organization and the respective size of their fan-bases. Since every losing team is a few ping pong balls from the rights to a LeBron James, Kevin Durant or Dwight Howard, personnel determines scheme in the NBA.

In contrast, the vast majority of the 344 Division I programs in college basketball have little chance of ever receiving a commitment from a McDonald’s All-American. But instead of petulantly trying to sabotage the sport in a misguided effort to legislate fairness, schools try many creative ways of leveraging the talents of the players they can recruit. As a result, scheme determines personnel in the NCAA.

At Syracuse, Jim Boeheim has made a Hall of Fame career out of running a contrarian scheme, in his case an aggressive 2-3 zone. The Orange traditionally have rosters full of “1.5’s”, 6’3+ combo guards lacking the quickness to defend elite PG’s and the size to defend SG’s, and “3.5’s”, 6’8+ combo forwards lacking the quickness to defend elite SF’s and the size to defend PF’s. However, because Syracuse never plays man defense, the athletic deficiencies of their players are minimized.

So while nearly every NBA team runs a fairly similar system of isolations, pick-and-rolls and man defense, an incredibly diverse array of styles can be found in the college game. On one end of the spectrum, teams like Missouri play four guards and pressure the ball 94 feet for 48 minutes, on the other, teams like Wisconsin run a deliberate motion offense, trying to minimize the number of possessions and shoot at the very end of the shot-clock.

In the NBA, the players are too good for the “40 Minutes of Hell” system (which Mike Anderson has brought to Missouri and Arkansas in the last few years) to be successful. Like Mike Leach’s bizarre pass-happy offense in college football, Anderson’s system, which he learned as a member of Nolan Richardson’s staff in Arkansas in the 1990’s, has philosophical holes that professional athletes can exploit. Nevertheless, that doesn’t make them any less entertaining on the collegiate level.

And with 68 teams set to compete in the NCAA Tournament, there are a lot more surprises in the college game. Even programs ranked in the top-15 like Murray State have barely been on national TV this season.

We have a pretty good idea of how teams like the Pacers and the 76ers match up with the top of the Eastern Conference but not whether an undersized Murray State squad can handle the size of an elite team from a Power Six conference. It’s an open question how Isaiah Canaan’s speed and athleticism translates outside of the Ohio Valley Conference. Non-conference play in college basketball generally ends in late December, so it’s almost impossible to gauge how younger teams like Texas, Washington and Tennessee who have found their groove in the last two months will fare in March.

In the NBA, it’s hard to envision a scenario where Chicago, Miami and Oklahoma City aren’t three of the final four teams left in the playoffs. In the NCAA, as many as two dozen teams have a legitimate shot at making a run at the Final Four.

Of course, in terms of entertainment, none of this makes the NCAA necessarily better or worse than the NBA, just different. But, as Posnanski writes, there’s something to be said for the concept of “different” in the modern sports world. Basketball fans of all stripes should enjoy March Madness; the NBA will still be here in a few weeks.

This weekend Michigan St. beat Ohio St. by forcing Jared Sullinger to commit 10 turnovers, Texas Tech beat Oklahoma to become the final BCS team to win a conference game, Alabama head coach Anthony Grant suspended his four best players for a violation of team rules in a loss to LSU, and Texas had a huge comeback to beat Kansas St.

There were plenty of exciting games, from Louisville’s comeback against West Virginia to UNLV holding off San Diego St. on a last second steal. But the heart of the season is sort of an awkward mix. There aren’t very many surprises any more. (I already knew Florida and Creighton didn’t play defense.) And we are too far from Selection Sunday to make blanket statements. (I’m pretty sure Pittsburgh’s loss at Seton Hall means their NCAA hopes are over, but I’ll feel a lot more certain in a couple of weeks.) So rather than focus on what happened this weekend, let me spend a little time talking about some player data.

Understanding Breakout Offensive Players 

College basketball fans might think that the major contribution of Dean Oliver is the creation of tempo free player statistics. But the “tempo free” part is actually a minor part of his ORtg formula. His key contribution was figuring out how to give players credit for assists and offensive rebounds. Clearly on a possession, the player who scores the basket might not be the only player who deserves credit for the points.

Let’s take North Carolina as an example. In this table ORB=Offensive Rebounds, AST=Assists, PTS=Points Scored, and PP=Points Produced.

(Through Wed. Feb 8^th)

In terms of actual points scored, Kendall Marshall is a relatively quiet scorer in North Carolina’s rotation. But because Kendall Marshall assists on so many baskets, in Dean Oliver’s formula, Marshall gets a lot of credit. When you look at the PP (points produced) column, Marshall gets credit for almost 100 more points on the season in Oliver’s formula. Basically everyone else has to share credit for their points with Marshall. So everyone else has lower PP relative to PTS. Offensive rebounding numbers work the same way. Because Zeller grabs more offensive rebounds than Harrison Barnes, he gets some credit back.

ORtg just takes PP and divides by the number of Possessions each player uses. If you produce 115 points and use 100 possessions, your ORtg is 115.

If you want to try and back out the “weights” for assists and offensive rebounds in Oliver’s formula, you cannot. Oliver’s formula isn’t linear. Also, if PP isn’t complicated enough, Possessions also must give credit for things like assists and offensive rebounds. If you check out Dean Oliver’s book “Basketball on Paper” from the library, you’ll see the full formula fills up about a page of text and takes a chapter to explain. It is ridiculously complicated, but it does do an incredible job assigning credit based on the stats in the box score.

If you’ve been following college basketball for awhile, you probably know who the breakout players are this season. But my goal today is to tell you why those players are breakout stars. At first, I thought about generating these types of tables:

Biggest Jump in Percentage of Possessions Used

(Table includes only players with at least five games and a minimum of 25 possessions in both seasons in a top 18 conference.) 

When 80% of Penn St.’s production graduated, Tim Frazier became a high volume shooter. One of Ken Pomeroy’s classic observations is that role players rarely become high volume shooters, but when teams lose high volume shooters to graduation, it sometimes happens. I could also create this type of table:

Biggest Jump in ORtg

(Table includes only players with at least 5 games and a minimum of 25 possessions in both seasons in a top 18 conference.)

Jaquon Parker’s improvement in efficiency for Cincinnati seems very important, but has anyone even heard of Fred Gulley? (Hint: He transferred out of Oklahoma St. mid-year to attend Arkansas.) But because Gulley rarely played in either season, his explosion in ORtg isn’t terribly important. Now, I could make the list look a little more sensible with some strategic cut-offs. For example, I could require that players have over 40% of the team’s minutes in both seasons, ect. But that can be a bit of a puzzle because often breakout players go from few minutes to major minutes. 

And that’s where the PP concept from the beginning of this article comes into play. Why not take PP per Game and see how and why that has changed for various players. (Remember this isn’t literally Points Per Game, this is the Dean Oliver concept of Points Produced per Game.)

Taking these numbers, who are the Top 20 breakout players this year? I not only identify the breakout players, I can also explain why their stats have improved:

ChangePPG: Total Change in Points Produced Per Game (with credit re-assigned for assists and offensive rebounds.)  This has three components:

ChangeORTG: Given the players current role, how does the change in ORtg from last year impact the player’s Points Produced Per Game?

ChPossUsed: How does the change in percentage of shots taken (really possessions used) impact the player’s Points Produced Per Game?

ChPT: How does the player’s playing time (and team’s tempo) impact the Points Produced Per Game?

Top 20  Breakout Players

(Table is limited to Top 18 conferences.)

-With so many players graduating from San Diego St., Jamaal Franklin shoots a lot more often, and he plays a lot more minutes. His ORtg has slipped slightly, but that has only minimally held back his stats.

-Meyers Leonard isn’t necessarily being all that more aggressive on the floor, but he is converting at a higher rate, and playing more than last year.

-Thomas Robinson is no more efficient than last year, and while he is a little more aggressive in his shot selection, his main improvement in PP has come from additional playing time. (In other words, imagine what he could have done last year if he wasn’t stuck so low on the depth chart.)

Now, besides looking at the top overall improvements, we can also identify players where an improvement in ORtg has been the most critical:

Critical Improvements in Efficiency

(Table is limited to Top 18 conferences.)

Just to emphasize what this list means, this isn’t the players with the biggest improvements in ORtg, this is the list of players with big improvements in ORtg who also use a lot of possessions for their team. 

Indiana’s Matt Roth improved his ORtg from 131 to 165 (which is mind-boggling), but he plays and shoots so rarely that his offensive improvement is only worth about 0.8 PPG to Indiana. Instead, Will Barton’s improvement in efficiency is much more important. Will Barton’s ORtg has improved from 99 to 114 and given his role in Memphis’ offense, that translates to about 2.4 PPG.

More to come on Thursday!