Blog | Sportemind's College Basketball

Coaching Merry-Go-Round

3/8/19

Sports may be the closest thing to a meritocracy in society. Not only are there clear goals, but the goals are an easily defined metric - wins. This metric can be used to compare one's program to other programs - and to its past - to determine current success pretty easily. If a program isn't successful, then its "employees" must change. There's trades, the draft, and free agency in professional sports and recruiting and transferring at the college level. If a player can't play, he won't continue to play. The same goes for a coach... but to a lesser extent.

Using my coaches database (some of which you can see in the misc tab), I did some statistical research for hiring and firing of coaches. It was a bit disappointing for meritocrats. Success (and lack thereof) was definitely a factor, but the correlation between success and a new job was not as high as I'd hoped.

It should be noted, there is a lot of noise in the sample. A coach could get fired for player mistreatment, off-court scandals, cheating, etc. Also, many coaches resign which could mean they saw the writing on the door, or they have health problems, got a new job (eg NBA coach or school's AD), want to spend time with the family, etc. Along with that, coaches retire, and you have to ask yourself, "did Bob Williams really retire?" And then there is moving up the ladder in college basketball. High major programs hiring mid major coaches. But sometimes the line is gray. Cuonzo Martin, Jamie Dixon, Kevin Stallings, Frank Martin, Buzz Williams, etc. Also, it could just be because of a connection (Lavall Jordan to Butler anyone?).

Nevertheless, there are trends. Here's some things I found out...

Firing

Six points. If you are six points worse than the program average, that's when you should start sweating.
Two/All years. For long tenures (7+), past success doesn't matter too much. If you are below average two years in a row, you could be gone. Whereas shorter tenures, they look at the whole body of work. A few below average years w/ a bad one gets you the hook.
4th year. That's when most firings take place. It is true, ADs like to give a coach a full recruiting cycle; you have to be very bad or there be extenuating circumstances to get fired before then.
Previous coach. There's a shorter leash if the coach before you had success.

(Up)Hiring - a program hiring the head coach of a worse program

Six points. If you are six points better than the program average, then you'll get looks.
Four(ish) years. For all tenure lengths, it's about an upward trend over the last four or so years.
2nd-4th year. Most uphires impress in a short time. Either hitting the ground running or turning a program around in four years.

With this information, I'm going to highlight some coaches who may get fired or hired. Now, with 353 teams I am not going to go over everyone here I think has a chance, so if your team's coach as the bane of your existence and he isn't on here, don't worry, he might get fired without my input.

Super-Twos

Bob Richey (Furman) - This is Furman's best season in their 100 yrs existence. And possibly more important, they made headlines. They beat the defending national champs, had a long undefeated run, and were ranked for the first time. They might not need a bid.
Dustin Kerns (Presbyterian) - Presby didn't have nearly the publicity, and they aren't making the NCAAs, but this was the 2nd worst team in basketball two years ago. Now, they're above .500 in conference-play for the first time ever.

Terrible Twos

Wyking Jones (California) - They've been terrible.
Jean Prioleau (San Jose St) - They were finally looking respectable, then Doug Wojcik resigned for health reasons and star Brandon Clarke transferred. Now, they've won two conference games in two seasons.

Three-mendous

Byron Smith (Prairie View A&M) - He's been on fire since he too over midseason as an interim coach for a 1-18 team. Looking for that tourney berth.
USF, UTSA, UMBC, UTRGV, JSU all have good coaches

Three-pathetic

Mike Dunleavy Sr (Tulane) & Terry Porter (Portland) - Both winless in conference. College ain't the NBA.
Maurice Joseph (George Washington) & Rodney Billups (Denver) - Trending badly in the wrong way. Both followed mildly successful coaches.
Allen Edwards (Wyoming) & Kyle Keller (Stephen F Austin) & Matt Lollitch (Valparaiso) - They're following up Larry Shyatt, Brad Underwood, and Bryce Drew. Interesting to see if they are treated the same way as group before.
Bryce Drew (Vanderbilt) - He is winless in conference which usually means a firing. However, he got two 5-star recruits this past year, and he is Bryce Drew. Should be interesting.

Fantastic Fours

Eric Musselman (Nevada) - Obviously amazing
Nate Oats (Buffalo) - Markedly awesome
Ritchie McKay (Liberty) - He's been very good. He has also been at Liberty before. He was the guy who got Seth Curry. But, after two years, he left to be the assistant coach to Bennett at Virginia.

Fatal Fours

Surprisingly, no clear-cut cuts.

Fast Fives

Kelvin Sampson (Houston) - Time forgives.

Furious Fives

Ernie Kent (Washington St) - ... I would draft Robert Franks.
Danny Manning (Wake Forest) - Victim of success. Could have had John Collins, Doral Moore, and Bryant Crawford as seniors, but he coached too well and they left early for pro careers.
Andre Payne (Mississippi Valley St) - They've been very bad since going 17-1 in conference.

Sweet Six...teen

Casey Alexander (Lipscomb) - Now is the time to leave, but maybe he'll wait for Belmont or Vanderbilt to open up.
Joe Golding (Abilene Christian) - Another example of a good transition from DII.
Dan Majerle (Grand Canyon) - He might be waiting for Arizona State to come calling, but if Sean Miller does eventually go down to the feds, I would expect University of Arizona to call.

Sour Sixes

Chris Casey (Niagara) - Six years later, he is still not the next Joe Mihalich.
Murray Garvin (South Carolina St) - He is not the next Cy Alexander... we're getting pretty niche now.

Aged Like Wine

Wes Miller (UNCG) - He's had more success there than Fran McCaffery and Mike Dement.
John Becker (Vermont) - He's had more success there than Mike Lonergan and Tom Brennan.
Tod Kowalczyk (Toledo) - He's had two recruiting cycles now and twice shown he can produce a top team in the MAC.
Russell Turner (UC Irvine) - He's made Irvine the goliath of the Big West w/ and without Mamadou Ndiaye.
Mike Young (Wofford) - I expect him to follow Bob McKillop's path and stay where he's at.
Bill Coen (Northeastern) - The Colonial no longer has VCU, George Mason, or Old Dominion, but it's still impressive Coen has made Northeastern the top program in the conference.

Might Be Put to Pasture

Steve Masiello (Manhattan) - Oh, how far he has fallen since the undergraduate degree debacle.
Sydney Johnson (Fairfield) - Remember when Fairfield was good?
Steve Hawkins (Western Michigan) - After a 2-16 year in conference, he might "resign."

The Oklahoma-Kruger Effect

2/18/19

The Oklahoma-Kruger Effect is a cognitive bias in which people mistakenly assess teams based on a lack of success in conference. One notable example of this phenomenon is apparent in college football where many believe a team should not be allowed to play for a national championship if said team did not win their own conference. A seemingly logical view, it suffers due to SSS (small sample size) and artificially constraining the amount of games a team has played. Before moving on to the task at hand, it should be noted that on January 8th, 2018, Alabama - a team that did not win its conference - won the national championship proving the physical possibility of such a situation.

In college basketball, there is much disdain for the below .500 conference team. The old adage says, "a losing record may burst your bubble, but never venturing from home may do just the same." There are a couple ideas involved here, but let's look at the first where a team that is below .500 is said to have their bubble burst, a euphemism for not making the tournament. So, it is understandable that there is concern of teams gaining at-large bids despite a losing record in conference.

Oklahoma basketball is currently 4-9 in conference. In spite of this, OU is safely in the tournament field as they currently sport a 16-10 record against a very high difficulty schedule of opponents. The computer rankings herein rate Oklahoma as a 9 seed. The human rankings at bracketmatrix believe Oklahoma to be a 10 seed with a staggering 104 out of 111 having OU as an at-large.

Why is there, currently, no significant evidence of the Oklahoma-Kruger Effect?

This observer posits the time of year is a direct factor in human beings suffering from this cognitive bias. It is known, no one took Kentucky out of their respective brackets when Kentucky went 0-1 in conference play. Why is this? Because the human knows there are more conference games to play, and Kentucky can get to a .500 record and, in fact, surpass it.

Oklahoma has 5 games left. It would not be surprising to see them finish 6-12. At this point, with their record etched in stone, the masses will begin to notice. Lon Kruger's team will be put through a crucible, metaphorically, on TV. One should remember, Trae Young is no longer a member of OU, and ESPN will have no mercy.

The Southern Conference & Getting Four Bids

2/14/19

The Southern Conference is going to do something unprecedented. They are going to be a multiple bid conference. This has never happened for the SC - Steph Curry's Davidson couldn't even get an at-large. That is impressive in and of itself, but the SC can do even more. And that's what I want to discuss, the opportunity of getting four bids.

Before I get to the four teams, let's talk about their competition for bids - it is low. Be prepared for pundits talking about the weak bubble. Still, the ACC, B10, B12, and SEC are doing well, maybe even better than normal... but then there's the fifth power conference, the Pac-12. The Conference of Champions blows. You know this, I know this, everyone knows this but Bill Walton. Washington is solid thanks to Mike Hopkins's 2-3 zone, and then there's... Arizona State. Bobby Hurley needs to stop throwing away games if he wants an at-large, but the hole might be too deep already. The Pac-12 may only get one bid. That's pathetic.

However, they aren't the only conference struggling. The BEast is also a power conference (in basketball). They actually have more bids over the last five years (ie since they lost their football schools) than the Pac-12. This is by far their worst year with only Villanova and Marquette looking like locks - it is possible St. John's or someone else goes on a run and gets an at-large.

And it doesn't start their. The next tier is struggling too. The Mountain West and West Coast Conference can usually muster a 2nd bid, but that doesn't seem to be the case this year. Although, it will happen if Nevada or Gonzaga don't win their respective conference tourneys. The American Athletic looks like a two bid conference, though the same caveats apply. Also, to note, the American Athletic has constantly underachieved, but the hopes are that Gregg Marshall, Penny Hardaway, and Danny Hurley should make this a perennial five bid league. And last, but certainly not least, the Atlantic 10. They haven't gotten fewer than three bids since 2007. They have been a victim of attrition/realignment, but this is bad. They don't have a single at-large team, at the moment.

That's a lot of open at-large bids. Someone has to take advantage. Why not the SC?

The biggest name is Furman. They started off 12-0 with wins at Villanova and Loyola-Chicago. Those would have been phenomenal wins last year, but they're still pretty good this year. They've lost some games since then. Four losses away from home to top-100 teams doesn't hurt too much, but a loss at home to Samford does. Going to the line up three points with 19 seconds left has to be a win, but it's not. Because of this, Furman still has some work to do. According to my tournament projections model (if you don't know, it's based on the theoretical 46th best team playing your schedule), Furman needs to go 4-1 to finish out the regular season to stay on the right side of the bubble which means they have to take two of three of vs Wofford, vs UNCG, and @Samford.

The best team is Wofford. They're 13-0 in conference. They lost four times in non-conference, but the worst one was @Oklahoma. Plus, they did win @SCarolina. That's a solid non-conference. Fletcher Magee may be the best shooter in college basketball, and Cameron Jackson may be the best player in the SC, though he did get outplayed by Furman's Matt Rafferty. It's impressive what Bob Richey has done in two years at Furman, but it's also impressive what Mike Young has done in about 100 years at Wofford. If you had to pick a best program in the SC, they're it. I have them needing to go 3-2 the rest of the way to feel like they're in. At Furman is a tough game, but they're favored in all five games.

The best resume actually belongs to UNC Greensboro. Losing @Kentucky, @LSU, and vs Wofford and beating everyone else is a good place to be. Wes Miller, the former Tar Heel, has done way more at UNCG than Fran McCaffery and should get hired by a bigger program any day now. They need to go just 3-3 to keep their head above water. Their next two games are @Furman and @Wofford, so they may need to take care of business down the stretch, but winning one of those two would give them their best win of the season and an inside track at a bid.

Now, this is all good and dandy, but there is a problem - the world is systematically against the proletariat, err, I mean conference tournaments are systematically against mid-majors. Either you win your tournament and get the auto bid, or you add a bad loss to your resume. The conference tournament is an at-large killer... for some. If you are in a major conference, you don't need to avoid bad losses because there aren't hardly any. Plus, you can pick up a good win or two away from home. You can put yourself into the field without winning the tourney. This is impossible to do as a mid-major. Conference tournaments are mid-major killers.

However, the SC still has a (small) chance of getting a fourth bid for the SC has an ace up it's sleeve, East Tennessee State.

Losing a neutral game (ie a tournament game) to ETSU is not a good loss, but it's palatable. ETSU is about as good as Northwestern or Colorado - bad, but not terrible, power conference teams. A neutral loss to ETSU is about -0.65 wins. I project both Wofford and UNCG to have that much cushion. Given that ETSU will play one of them in the semis (for Wofford and UNCG will be the top two seeds), it shouldn't cost the SC an at-large if ETSU wins the tourney and gets the automatic bid. And, as a tangent, this would really help Steve Forbes. He's been wonderful at ETSU, relying on down-transfers early in his tenure, but now his team is comprised mostly of players who have only know the old navy blue and gold.

Still, it's a pipe dream. The committee would never do it, give a mid major conference three at-large bids. But it's good to dream.

Is Virginia's Defense this Good?

2/9/19

Tony Bennett's calling card is becoming an annual post... but I ain't burying the lede, Virginia is getting lucky.

They're shooting very well offensively. It's hard to shoot 76.1% from the line. It's hard to shoot 39.4% from three. (That puts them in the top 20 teams in both metrics.) So, conservative projections would say they would regress (downwards) in both. However, free throw shooting does correlate with three-point shooting (and vice versa), so they probably shouldn't regress too much.

Then, there's their defense. (If I could've just fit a "they're" in there.) It isn't as good as last year's when they were getting an unprecedented amount of turnovers for a pack-line team. But, they are making up for it so far by holding their opponents to 24.7% from deep and 62.5% from the line.

Those aren't gonna hold.

Let's start off with the freebie - Virginia doesn't have some voodoo defense when the opposition takes foul shots. Last year, teams shot 73.2% against them from the line, 280th in the D-1. This year they are 3rd, and (And!) the best a team did at free throw "defense" last year was 65.6%. There's obvious regression there.

Now, three-pointers you can actually defend against. However, it's pretty well-accepted you don't have that much influence on an opponent's 3P%. If you haven't heard the spiel, I'll give it a go. The theory (to match the cold, hard data) is teams like to take threes when they're open. You can limit the amount of open threes (ie 3PA%), but inevitably good looks will happen, and the opposition will make threes around a third of the time. Now, this isn't to say you have no control. There is a correlation with 2P%. The theory, if points are really hard to get inside, then the team will settle for harder three-pointers. So, it isn't rare for a Boeheim-Syracuse or a Bennett-Virginia to have a good three-point defense.

However, 24.7% is crazy low; it's unsustainable. The lowest in the last ten years is 27.5% and that was ten years ago when people were worse at shooting. UVa had the 9th best three-point defense last year, and teams still shot 31.0% against them. This year is an outlier. They're going to regress.

And yet, if they give up 31.0% in the next (last?) 15 games, they would be just a hair under 27.5% on the season. They would have the best three-point defense in recent history. Tony Bennett could hang his hat on that. Although, he probably won't notice. He is probably just concentrating on getting out of the first round. (Sorry, I had to.)

Now, when I say up-transfer you probably think of Damion Lee, a player going from a mid-major to a power conference. But, here I am using up-transfer to mean Duncan Robinson, a player going from D-II (or D-III in his case) to D-1.

Think of all the up-transfers you know in recent history. There's Duncan Robinson. He was pretty good. There's Derrick White. He was also pretty good. Then, there's... Max Strus? Also, pretty good, but not a household name. So, the list is short, but very good.

So, the question is are all up-transfers good or do we just not hear of the ones who don't pan out? I don't really know, but hopefully at the end of this year, we will have an answer.

Here is a list of up-transfers for this year...

Trey Dreschel - Grand Canyon
Jarrett Givens - Austin Peay
Shea Feehan - Evansville
Zach Hankins - Xavier
Keenan Gumbs - Liberty
Jarell Spellman - Sacred Heart
Skyler Hogan - SEMO
Yanni Wetzell - Vanderbilt
Freddie Gillespie - Baylor

That is a staggering nine players. Though, if you count (which I don't for this exercise) there are more on North Alabama and Cal Baptist.

We may look back at the end of the season and see some great players, but as it stands now Hankins is getting 20 mpg and Wetzell is just getting 10 mpg for the two high-majors.

Up-Transfers, Always Great?

11/28/18

RPI No Longer Reigns

2/13/18

A lot has been made about the resumes given to the tournament committee having more than just RPI. Specifically, it will include rankings from results based metrics KPI and SOR and rankings from predictive metrics Kenpom, Sagarin, and BPI. (For the record, my Tournament Projection is a results based metric and my Dominance Rankings is a predictive metric.) But is all the hubbub grounded?

Fortunately for us, we don't have to wait until Selection Sunday to see if there is an effect as college basketball is following college football's business plan by teasing rankings.

So, we have a sample to test from. In past years, my Tournament Projection was more accurate using RPI rankings as an input compared to using Kenpom rankings as an input.

Pretty simply, I compared my current Tournament Projection to the committee's release using both inputs. It turns out, the Kenpom input produced more accurate results. The committee has taken a step away from RPI! (Though, it has also taken a step towards using predictive measurements as descriptive statistics.)

Now, the most accurate Tournament Projection used a blend of the two, so that is what I will now use going forward (using the full six did not add any accuracy). It only really missed on Duke, Oklahoma, and Arizona.

This is a pretty important happening. The committee is human, and those three teams have arguably the three most marketable players - Marvin Bagley, Trae Young, and DeAndre Ayton. I don't mess with biases in my projections - I have more of a perfect world projection.

Why is Virginia So Good?

2/1/18

This was not expected to be Tony Bennett's best Virginia team, and yet it is. It's not like they were expected to be bad (27th in both preseason AP and Coaches Polls), but last year's version was the worst in four years, and they had two seniors-to-be transfer out from that team in Marial Shayok and Darius Thompson. They just don't seem to have as much talent as they had in their dominant three year stretch from 2014-16 when they had Joe Harris, Akil Mitchell, Justin Anderson, Anthony Gill, Malcolm Brogdon, and London Perrantes.

Years from now, we will have the benefit of hindsight, and maybe this will seem like a really talented team. But for now, we can attribute Virginia's elite play to their defense.

Now, every year they have an elite defense. Bennett is renown for his pack-line defense. The pack-line defense is a man-to-man defense with a slight wrinkle. Instead of tight cover off the ball, the defenders sag off their men towards the middle. This allows them to already be in help position, but still allows them to quickly close out their man if he is passed to.

But this year, they are even better. They are allowing only 0.822 points per possession (adjusted for competition). The next closest team is at 0.863 ppp. The best Virginia has ever done is 0.855 ppp. And, the best any team has done since 2002 is 0.842 ppp. Add in the fact teams in 2002 averaged 1.03 ppp compared to 1.06 ppp today, and Virginia has cemented themselves as the best defense in recent memory. They're killing it. So, why is that?

Their defense is a little different this year. They aren't packing it in quite as much. Because of this, they are giving up more offensive rebounds. Bennett's team has been top 50 in defensive rebounding for ten straight years, and they are outside the top 100 this year. However, they are more than making up for it by getting steals.

Now this isn't enough in it's own. Plenty of teams create turnovers. But, there's one more special thing about Virginia. They create turnovers without committing fouls. This means they are not swiping, but picking passing lanes. Syracuse usually does pretty good at this with their 2-3 zone, but good, long defenses like this year's Cincinnati can do this too.

Since 2002, there have been only three defenses in the top 20 in turnovers created and fouls committed - 2004 East Tennessee State, 2014 Ohio State, and 2018 Virginia. That was a solid ETSU team, but the OSU team led by Aaron Craft is a better comparison. Most great college basketball defenses are anchored by a shot-blocking big, but neither of these was. They had ball-hounding guards with lengthy, athletic wings/forwards.

What this allows is three types of steals - tipped passes, forced bad passes, and double teams - all in the half-court defense. In comparison, West Virginia gets their steals via three main situations - full-court pressure, in-bounds plays, and opportune moments.

Virginia's best defensive outing this year came against ranked Clemson (sans Donte Grantham, unfortunately). Virginia had 14 steals (they forced 19 turnovers), but only committed 10 fouls. Their 14 steals were comprised of six tipped passes, two picked passes, two forced bad passes, and four double teams.

Now, all of those are good ways to get steals. Ty Jerome reads the game well and had both picked passes. Forcing difficult entry passes down low is a sign of good defense. Plus, having big men who can move their feet and take correct hedging angles allows for steals while doubling. But my favorite is stealing a pass while face-guarding your man. They did this a few times, let's look at a couple of them.

Kyle Guy gets the steal here. He sees a Clemson player pick up his dribble, but doesn't double. Instead, he stays with his man. And not only that, but stays facing his man, so he is ready to defend if the pass does get through.

This next one is a good example of the pack-line defense. Notice how crowded that lane is. However, Devon Hall doesn't stay and double when the Clemson player picks up his dribble. Instead, he goes to defend his man while putting his hands in the air, hoping for a steal, but not needing one.

And that's the key. When they play good defense and have a chance for a trap, they usually don't take it. They know trapping a ball-handler can result in a steal, but it also leaves a man open. They don't like that risk. By playing defense this way, there is no risk, but there is still a chance for a steal.

Is Projecting Freshmen a Crapshoot?

11/6/17

The answer is, obviously, no. #BetteridgesLaw

Anyways, there is a (steadily decreasing) minority who think freshmen haven't earned anything and, ipso facto, them and their teams shouldn't be ranked highly. That is at odds with my system which projects Marvin Bagley III and Duke as the best player and team, respectively.

More interestingly, there is a majority who think freshmen are inherently more inconsistent, unpredictable, unreliable, fickle, volatile, and so on. Anecdotally, it makes sense; freshmen have never played college basketball before.

What do the stats say? My freshman algorithm has an R^2 value of just over 0.5 with an average error of 11.5 in terms of player rating. My returning player algorithm has an R^2 value of over 0.6 with an average error of 9 in terms of player rating.

Freshmen are harder to predict, but they don't lag too far behind returning players. Plus, top 100 freshmen are easier to predict (9.5 average error).

Top 25 Players Preseason

11/5/17

The top ranked freshman is consistently very good. Josh Jackson, Ben Simmons, Jahlil Okafor, Andrew Wiggins, Nerlens Noel, Anthony Davis, are the top ranked freshmen in each year since 2010. Expect a lot from Marvin Bagley.

Jevon Carter is a surprise at number two, but steals are worth a lot, and he had the sixth most last year. Additionally, he is a quality offensive player for a good, high major team.

Gary Clark is another surprise. He hasn't had a possession rate over 20% in any of his three years, but he can do everything on the court. He scores efficiently, rebounds on both ends, takes care of the ball, and doesn't foul while getting both blocks and steals.

The rest of the top 10 is much of the same - top freshmen or players who can play both sides of the ball.

It will be interesting to see who is the better of the M Bridges...