"My rule was I wouldn't recruit a kid if he had grass in front of his house.
That's not my world. My world was a cracked sidewalk." —Al McGuire

Wednesday, August 17, 2022

How to do NCAA Tournament Expansion Right

SEC Commissioner Greg Sankey recently suggested the NCAA take a "fresh look" at the selection process. His comments indicate two things. First, the big boys would be interested in expanding the tournament field. In addition, he spoke about eliminating the automatic qualifiers, which chafes fans who love the idea that at the beginning of March, every eligible team (outside a few Ivy League outliers) has a clear, if not necessarily realistic, path to winning the National Championship.

Greg Sankey is hinting at NCAA Tourney expansion
 Photo by Andy Lyons | Getty Images

I would join the chorus of individuals that does not want the automatic qualifiers to go away. While Sankey based his argument on First Four to Final Four teams like 2011 VCU, 2018 Syracuse, and 2021 UCLA, the removal of automatic qualifiers would likely eliminate every automatic bid 13-seed or worse, which means the even-more-recent 2022 St. Peter's story never would've happened.

More often than not in the constant "tradition vs money" battle, money wins out, but I'd like to explore an option that would allow for more money to be made (and likely consolidated by the major conferences) while also preserving the tradition that gives every team a path to the title. Here is what I would propose:

Step 1: Add First Four sites

Not only is expansion inevitable, if done right it could actually make the NCAA Tournament better. The first step is expanding to more Tuesday and Wednesday sites. Dayton has been a good pick because it didn't host regularly, is in close proximity to another major city, and has a rabid basketball fanbase that will come for the entertainment, not just for the hometown team.

My first addition would be Omaha. They've only hosted once historically (though recently, in 2018), draw a passionate crowd with regularity, and don't have pro sports options to compete with. Omaha is also located toward the western side of the country which will make it a better geographic fit than cramming everything east of the Mississippi.

We shouldn't need a pandemic to put the Tourney at Assembly Hall

Photo by Matt Begala | Indiana Athletics

Next up, you have to appease the big boys. While Knoxville used to host somewhat regularly (three times in the 1990s) they haven't had NCAA games at Thompson-Boling since 1999. This would provide guaranteed revenue to an SEC school and is again a large enough arena that will draw fans. Finally, we need a Big 10 school, and I would propose Indiana's Assembly Hall. Outside of the 2021 all-Indiana bubble, they haven't hosted since 1981 but have dedicated basketball fans and like Dayton, close proximity to a larger city. So let's go with Creighton, Tennessee, and Indiana joining Dayton as perennial First Four sites. You can even keep the "First Four" moniker because those arenas will be the First Four hosts every year.

Step 2: Every 15/16 has to play their way in

Rather than doing away with the 15s and 16s completely, which would rob fans of stories like St. Peter's, UMBC, and Dunk City that are just as compelling as the First Four to Final Four stories, let's send them all to First Four sites. This would accomplish a number of things. It would retain the "everyone has a chance" storyline, preserve the "easier" path for 1/2 seeds, and improve the 64-team field come Thursday.

By putting the "worst" teams into these games, it would still give everyone the theoretical chance at the NCAA Title. After all, if the first team to go from 16-seed to Sweet 16 is someone like post-automatic bid Syracuse (and you know it would be them), it won't quite have the same magic as we felt when UMBC pulled the Virginia upset. Retaining the automatic bids, even if they are diluted in the course, allows the NCAA Tournament to keep the magical feel that makes it so unique among American playoffs. 

Moments like UMBC over UVA are only possible with autobids intact

Photo by Andrew Shurtleff | Daily Progress

In addition, if we were to completely eliminate automatic bids, the opponent quality seen in opening games by the 1 and 2 seeds would be far more difficult to overcome. While upsets are fun, both the monied interests and the fans want to see the big boys in the second and third weekends. Keeping the weakest teams in the 15/16 seeds spots would help deliver that. After all, fans want the top teams to advance and the harder you make their path the less likely that is to continue.

That said, in 1985, the NCAA went to the 64 team seeded format. From 1985 through 2010, just 4 teams seeded 15 or worse won a game. There was one team added, to the 65-team field in 2001, but the results didn't change much. Then, in 2011, they expanded the field to 68. Since that expansion, 7 teams seeded 15 or worse have won a combined 11 games. By knocking out two of the worst 16-seeds in the First Four, the remaining 16s are stronger. Teams that would've been on the 14 or 15 lines years ago are now 15s or 16s, and their results have improved accordingly. Keeping the 15s and 16s would strengthen that pool and make the 1/2 seed first round games that much more compelling.

Step 3: Add 12 at-large teams

This is where people wince. But while it's easy to say "why add bad NIT caliber teams?" the question should be "are there more VCU/Syracuse/UCLA teams out there?" Might 2018 Penn State or 2019 Texas, both of whom finished top-25 at kenpom after winning the NIT, had a similar NCAA run in them?

There are a few reasons to justify adding these teams. The obvious first one is money. More inventory, particularly if it means adding high-major teams on Tuesday and Wednesday (always the most watched games) means more money. The vast preponderance of these bids will also go to high major teams. If the impetus is from the top six leagues, they account for the majority of where those newly awarded bids would go. Looking at the top three NIT seeds from the past three tournaments, 25/36 (69.4%) teams came from the ACC, Big East, Big 10, Big 12, Pac-12, and SEC. More inventory, more money, and a larger percentage going to the programs that are driving the bus.

More inventory like Notre Dame/Rutgers Double OT please
 Photo by Jeff Dean | AP

However, the real benefit is turning Tuesday and Wednesday into legitimate NCAA Tournament dates rather than appetizers for the diehards. There would be a far bigger audience for a slate that featured the following at the First Four sites, using 2022 as an example:

  • Indiana vs Florida
  • Notre Dame vs St. Louis
  • Rutgers vs VCU
  • Wyoming vs Mississippi State
  • Dayton vs Xavier
  • Texas A&M vs North Texas
  • SMU vs Wake Forest
  • Oklahoma vs BYU

Those are games with big brands that will attract eyes regardless of fan allegiance. More games allows for better chances at attractive rivalry matchups or storylines, whether it's Dayton/Xavier in a true rivalry, Texas A&M/North Texas in an in-state tilt, or Oklahoma/BYU in the future and soon to be past Big 12. Mixing those with the 15/16 seeds games will provide a legitimate slate that will have broader interest. Tuesday and Wednesday won't be token games everyone can write off, they will be the legitimate start of the NCAA Tournament, spanning four networks with big brands playing consistently through both evening sessions.

Step 4: The makeup of the 64

So with this expansion, what will we be left with come the Thursday/Friday sessions? This will be the makeup of the remaining field:

  • 16 Automatic bids opening tourney play
  • 32 At-large bids opening tourney play
  • 8 Automatic 15/16 seed play-in bids
  • 8 At-large play-in bids

Without question, the quality at this point will be better. The 1 and 2 seeds will face the best remaining 15 and 16 seeds, many of which in the past would've been 13 or 14 seeds. In turn, the 11-14 lines will be made up of the better mid-major champs and at-large play-in winners. In general, everyone in this range will be a line or two lower than they would've been in the past, meaning better teams filling out those lines. And the top-10 or so lines will be unchanged, the best teams in the country getting in on merit as they always have.

If they really wanted to tamp down on small conference inclusion, they could include 8 more automatic qualifiers in the "play-in" to the 14 line as well. Or they could have First Four mini-tournaments, where the bottom 16 teams are broken into 4 pods and have to win games on Tuesday and Wednesday to advance to the 16-lines in the round of 64, while the 8 teams above them would play-in to the 15-seed games. This would take the bottom 24 automatic bids and pare them down to 8 participants in the main 64-team bracket. Would it be the same? Obviously not, but it would still give a chance to those teams and provide great storyline opportunity while also improving the overall quality of play once Thursday and Friday roll around.

The End Result

This tournament setup would be the best of both worlds. For the high-major conferences, they get more bids and by eliminating 8 autobids before Thursday, increase their odds of having their own teams advance. By increasing the First Four inventory, there is a tangibly higher value to the opening days of the tournament, which will hinge primarily on the big brands drawing eyeballs (as they always have).

What if the Tournament could be bigger AND better?
 Photo by Tom Pennington | Getty Images

For the small schools, they still have a chance to play in the Big Dance and eight coaches of 15/16 seeds will be guaranteed to go home claiming an NCAA Tournament win. In addition, by thinning the herd at the First Four sites the teams that do make it to the Round of 64 should be in better position to pull off an upset.

Basketball junkies also benefit because they get two days of legitimate content, not just a couple games per night that most people view as throwaways. Casual fans benefit because they get to see more big brands on the opening nights and get to tune in to the national brands to start matters. The NCAA as a whole benefits because more inventory means more advertising and more profit. Whether they could renegotiate as a result (the CBS/Turner deal that runs to 2032 is woefully small compared to the product value) remains to be seen but it would certainly enhance their negotiating ability when the next contract talks come up.

Thursday, July 07, 2022

Who Will Score? Part V: Calculating eFG%

Graphic from fourpointplays.wordpress.com

For basketball stats nerds, Effective Field Goal Percentage is king. Offensive efficiency is measured with four factors. The original explanation from Dean Oliver, author of Basketball on Paper, is included in that link, but essentially eFG% tells you how many points a team or player scores per shot taken. It is by far the most important predictor when it comes to winning basketball. The more efficiently you shoot and the less efficiently your opponent shoots, the more you are likely to win. It is similar to typical field goal percentage except three point shots made get an extra 0.5 points per made shot to reflect that 3 is 50% more than 2.

Our predictions were based on returning player that had adequate stats and comparable players that allowed our projections with reasonable mathematical confidence. In short, it's easier to look at what Tyler Kolek did the past two years and compare that to other up-transfers that with similar paths than to project what Ben Gold will be with no idea how his NBADL experience will translate to the D1 level.

We focused on the six most prolific returning players, which led us to look back at the six players who played the most minutes last year. Here are the top-six in minutes, their makes by category, calculated eFG%, and cumulative eFG%:

2022 2PFGM 2022 3PFGM 2022 FGA eFG%
Justin Lewis 137 58 443 50.6
Tyler Kolek 36 36 225 40.0
Darryl Morsell 106 43 334 51.0
O-Max Prosper 57 19 165 51.8
Kur Kuath 82 0 115 71.3
Oso Ighodaro 73 0 108 67.6
Combined 491 156 1390 52.2

While this doesn't include prolific three-point shooters like Kam Jones and Greg Elliott, that 52.2% eFG% is very close to being in line with the team eFG% of 52.1% in 2021-22. As go the top players on the team in terms of minutes, so goes the team as a whole. That 52.1% eFG% ranked 77th overall and helped lead to an Adjusted Offensive Efficiency Rank of 66. So what does that have to do with next season?

David Joplin & Kam Jones putting in work
 Photo by Jovanny Hernandez | JS Online

Because we are projecting expected eFG% for the top-six returning players, we can line up those players against the players above. For further context, the above players accounted for 73.4% of minutes played last year. Using our 2022-23 projections, the top-six (Tyler Kolek, Kam Jones, O-Max Prosper, Oso Ighodaro, Stevie Mitchell, and David Joplin) would account for 69.1% of minutes. A little bit less, but close enough to be relevant for our purposes.

Using the following formula, we can estimate the number of shots each player will take per game:

60 x % Minutes x % Shots = ?

60 is the number of expected shots per game, and by calculating the percentage of minutes played and the percentage of shots taken when a player is on the court, it gives us an expected number of shots taken per game. Then by multiplying that by 32, which is the number of games Marquette played in 2021-22 and the number of games they would play in 2022-23 assuming all regular season games and one Big East Tournament game, we know the number of shots each player would take. By multiplying that total by the eFG%, we have a number of effective makes. It won't give us actual 2PFGM or 3PFGM, but for instance in 2021-22, Justin Lewis would have an effective make total of 224, taken from the eFG% numerator calculation. Divide that 224 into his 443 field goal attempts and you get 50.6%, matching his eFG%.

The same would apply to our 2022-23 projections, with the added bonus that by calculating the top-six players' effective makes, we can measure those against their expected field goal attempts and by adding it all together, come up with an expected eFG% for those six players, and by extension the entire team. Here are the numbers:

Eff. FGM 2022-23 FGA eFG%
Tyler Kolek 129.4 265.7 48.7
Kam Jones 173 302.5 57.2
O-Max Prosper 123.1 218.2 56.4
Oso Ighodaro 109.7 164.4 66.7
Stevie Mitchell 79.2 148.7 53.3
David Joplin 100.2 207.4 48.3
Combined 714.6 1306.9 54.7

To provide some context in terms of reliability, the minute tally above equates to 94.1% of the 2021-22 minutes and the field goal attempts equate to 94.0% of the 2021-22 attempts. So if 2022-23 team eFG% is in close alignment with the top players in terms of minutes the way the 2021-22 team was, this is a strong indicator of eFG% improvement.

If the players improve at a rate comparable to the average sample (which included both the worst case and best case, so about average development) we can expect a 2.5% increase in eFG% in 2022-23. This makes sense as the two players that took the highest percentage of shots both had eFG% ratings under the team 52.2% average. While Marquette will be losing points, there is reason to believe they will be increasing their offensive efficiency.

To put into context what that hypothetical 2.5% increase would mean, Marquette's 52.1% ranked 77th in eFG% in 2021-22 but if they increased their team eFG% to 54.7% they would've ranked 21st in the category. And while the overall offense finished 66th, in 2021-22 every high-major team that ranked in the top-30 of eFG% was ranked in the top-41 of overall Adjusted Offensive Efficiency. That would mean an minimum expected 25-spot jump for Marquette in Adjusted Offensive Efficiency.

While Marquette doesn't project to have the individual scoring leaders they had last year, they are in line to have more players capable of stepping up on any given night and a more diverse and efficient offense. So the answer to the question of "Who will score?" is "It doesn't matter, as long as it's more efficient."

Tuesday, July 05, 2022

Who Will Score? Part IV: The Dark Horses

Today we continue our Who Will Score series for Marquette in 2022-23. Our first installment looked at the lack of correlation between returning volume scoring and offensive efficiency, particularly for teams that lose between 20-60% of points from volume scorers, like 2021-22 Marquette. In our last installment, we looked at the three most likely candidates to lead the way offensively. Today we turn to a fourth very strong candidate as well as the highest projected "best comparison" player whose potential ppg exceeds even Kam Jones.

Our initial focus today will be on returning players, with some words reserved for newcomers at the end. There will be one more installment after this with some takeaways as far as what to watch for offensively in 2022-23 as well as providing some reasons for optimism.

Before we begin, a reminder of how this will work. We are going to look at the following three categories that lead to scoring:

  • Minutes Played: The more you are on the court, the more you have chances to score.
  • Percent of Shots Taken: This tells us how likely a player is to get scoring opportunities when they are on the floor.
  • Effective Field Goal Percentage: This calculates what a player's shots are worth, with a 50% boost given to three point field goals made since three is 150% of two.

Expected scoring totals are based on 60 shots per game incorporating the three above factors as well as free throw rates and percentages from last year.

Determining the three categories will come by using multiple comparisons from Marquette, Shaka Smart, and kenpom comparison history to establish how similar players' games evolved from season to season in the past., also creating worst and best case scenarios by using the three season developments at the low and high ends of the developmental spectrum. Enough explanation, let's dig in:

Fourth Starter: Oso Ighodaro
Photo by Jeff Hanisch | USA Today Sports

Our last article focused on the three returning players with starting experience, but it seems like a virtual certainty that Oso Ighodaro will also enter 2022-23 as a starter since he is the only returning center on the roster that logged minutes last year. Of returning players, Ighodaro also led the team in offensive efficiency, eFG%, and free throw rate. Here are the projections:

  %Min %Shots eFG% PPG
Sophomore Sample 44.4 13.4 55.1  
Junior Sample 55.8 16.2 54.5  

Oso Ighodaro (So) 45.4 12.4 67.6 5.5
Oso Ighodaro (22-23) 57.1 15 66.7 8.3
Oso Ighodaro (Worst) 29.8 15.8 65.6 4.5
Oso Ighodaro (Best) 74 17.9 74.6 14.1

Oso is probably least likely to be close to his worst case scenario. Barring injury, he is going to log heavy minutes on a team without much depth at the 5. The question is which Oso we get. The one that was #2 nationally in eFG% at 77.3% through the first 23 games or the one that slumped to 45.5% in the final 9 games? Much of this will likely rely on the success of the pick and roll and whether the staff can develop providers other than Tyler Kolek. If they can regain that early season form, don't be surprised if Ighodaro is among the team's leading scorers.

Oso Ighodaro Comparisons: Ousmane Barro, Mike Edwards, Dan Fitzgerald, D.J. Haley, Theo John, Scott Merritt, Akil Mitchell, Jericho Sims, Ray Spalding, Kameron Woods

Worst Case Comps: Mike Edwards, Dan Fitzgerald, D.J. Haley

Best Case Comps: Ousmane Barro, Akil Mitchell, Jericho Sims

X-Factor: David Joplin
Photo by Jeff Hanisch | USA Today Sports

For this Marquette team, David Joplin is the ultimate wildcard. The reason for this is because he was a high usage, low minutes played option as a freshman. If his minutes spike to Justin Lewis levels while maintaining his percent of shots taken, the sky is the limit. That potential is why people like Scrambled Eggs' Phil Bush are on the Jopwagon. But if he takes a step back like sophomore Steve Taylor, he could be an afterthought. Check out the wild projection variances:

  %Min %Shots eFG% PPG
Freshmen Sample 18.6 19.5 49.7  
Sophomore Sample 42.9 19.9 49.7  
David Joplin (Fr) 17.6 26.1 48.3 2.8
David Joplin (22-23) 40.6 26.6 48.3 6.6
David Joplin (Worst) 17.7 24.5 36.7 2.1
David Joplin (Best) 78.1 29.9 65.7 19.2

Minutes will be the primary factor in determining Joplin's scoring, because he's never met a shot he didn't like. If a player who's willing to take shots is earning minutes, players like Troy Daniels show how quickly a star turn can come. But if his defense keeps him off the floor, going the other direction is possible. I think something in the middle is far more likely. Becoming a secondary scoring option off the bench feels most likely and despite the gaudy best case, I expect averaging double-digits will wait until next year.

David Joplin Comparisons: Ousmane Barro, Joe Chapman, Troy Daniels, Drew Friberg, Davante Gardner, Jajuan Johnson, Andrew Robinson, Steve Taylor, Daniel Utomi, Tyler Wood

Worst Case Comps: Joe Chapman, Steve Taylor, Tyler Wood

Best Case Comps: Ousmane Barro, Troy Daniels, Drew Friberg

Sparkplug: Stevie Mitchell
Photo by Collin Nawrocki | Marquette Wire

Did anyone do more to establish themselves late in the season more than Stevie Mitchell? When Marquette was struggling, he was the guy trying to pick them up. Mitchell's minutes dropped early in Big East play. After the calendar turned to 2022, he posted double-digit minutes 10 times and 100+ AORtg games 7 times. 6 of those 10 and 4 of those 7 were in the last 8 games of the year. So how does 2022-23 look?

  %Min %Shots eFG% PPG
Freshmen Sample 27.4 15.2 47.4  
Sophomore Sample 50 17.1 48.5  
Stevie Mitchell (Fr) 26.7 14.1 52.1 2.8
Stevie Mitchell (22-23) 48.7 15.9 53.3 5.9
Stevie Mitchell (Worst) 32.3 13.4 43.1 2.7
Stevie Mitchell (Best) 59.5 17.2 68.3 9.6

Mitchell has the look of a potential starter who will likely be part of four guards competing for three spots. However it's unlikely that he'll have a major scoring impact. Quite simply, guys with modest minutes and usage rates, even when they are efficient from the field, rarely explode into offensive focal points. He will more likely have an impact on the defensive end, though reports out of early camps are that Stevie is one of the most improved players in the offseason.

Stevie Mitchell Comparisons: Brendan Bailey, Ousmane Barro, Joe Chapman, Raheem Dickerson, Marcus Dickinson, Jase Febres, Theo John, Jajuan Johnson, Jeremiah Martin, Steve Taylor, Darius Theus, Kellon Thomas

Worst Case Comps: Raheem Dickerson, Steve Taylor, Darius Theus

Best Case Comps: Jase Febres, Jeremiah Martin, Kellon Thomas

The Rest of the Roster
Photo from @marquette.basketball Instagram

Last year, the top six players in terms of minutes accounted for 54.6 ppg in 73.4% of the minutes. Using the baseline projections, the six players we've highlighted would account for 50.8 points per game in 69.1% of the available minutes. Factoring that out would equate to 73.5 ppg as a team, about on par with the 74.0 ppg scored last year.

Kam Jones was not one of those top-six in minutes, nor was Greg Elliott, which means there is room for players outside those mentioned to break through. Without any track record, any projection would be a sheer guess, so we're going to focus on the rumors out of camp rather than trying to project points.

Zach Wrightsil: The NAIA Player of the Year seems likely to earn minutes, but without floor stretching ability he'll have to do his work inside. Wrightsil will likely have some games where he can exploit mismatches and put up some points, but his impact should be on the defensive and rebounding end more than as an impact scorer. We do see the occasional Ryan Hawkins or Max Strus type of up-transfer, but until he proves otherwise, tempering expectations seems wiser.

Sean Jones: All the reports indicate Sean Jones (pictured above) is the real deal. Despite being small of stature, he's strong, has great leaping ability, and is incredibly quick. The worst case scenario is likely a freshman impact somewhere between what Stevie Mitchell and Kam Jones did last year, though playing point his production may come more from creating for others than scoring himself.

Ben Gold: I've been pleasantly surprised by how strong the positive vibes are around Ben Gold. He is said to be a better than expected shooter and everyone I've talked to believes he will be able to contribute right away. Of all the newcomers, he seems the best positioned to add some bench scoring.

Keeyan Itejere: The simple lack of depth in the middle will probably give Itejere some opportunities. He is clearly an excellent athlete, though I would hedge my bets in terms of how much he offers on the offensive end in his first year on the court.

Emarion Ellis: He remains the high-upside player that might not be quite ready yet to have a major impact.

Chase Ross: True wildcard in that I haven't heard anything one way or the other about what impact he might have.

In our final piece, we are going to look at the most important factor in offensive success and why despite the lack of big-time individual scoring projections, Marquette may be in line for a significant improvement in offensive efficiency next year.

Tuesday, June 28, 2022

Who Will Score? Part III: The Leading Candidates

In our first two parts, we dug into the history books to determine whether or not the need to replace volume scoring had a negative impact on offensive efficiency going forward. While the general conclusion seems to be that other than at the extremes there is very little correlation between scoring lost and offensive efficiency, what fun would it be to throw up our hands and say "well, hopefully someone will score, let's be done with it!"

Now before you dig in too far, please keep the following in mind. This has been expanded to a five-part series, and while some of the parts may have less satisfying takeaways than readers (and, frankly, the author) expected, to fully understand the goal of the team and the overall final projections, you will need to read the series all the way to the end.

Who might carry the scoring lead after Justin Lewis & Darryl Morsell?
 Photo by Charles Fox | Philadephia Inquirer

When projecting who will lead a team in scoring, there are really three main factors:

  • Minutes Played: If you aren't on the court, you aren't going to score. However, because simply being on the court doesn't guarantee you will score, it can't be considered alone.
  • Percent of Shots Taken: Usage rate alone doesn't tell the story because it factors in possessions that end in ways other than with a shot taken. Looking at the percentage of shots a player gets while on the floor indicates how many scoring chances they will have.
  • Effective Field Goal Percentage: Stat nerds (and I imagine most readers here) have long preferred eFG% to other measures of shot efficiency. The easiest explanation is that it is field goal percentage with a 50% boost given to made three point shots because 3 is 50% more than 2.

To approximate next year's performance, we looked at players with similar past profiles to Marquette's returnees and weighed those past seasons against their future results. When you look at the charts, the portion above the break is that player's comparison player average (comps listed at the end of each player's synopsis) and the player's past and projected future numbers are below the break.

After breaking those down, we will be able to project 2022-23 expected minutes played, percent of shots taken, and effective field goal percentage. Those numbers will be balanced against 60 shots per game to produce an expected points per game average, with a slight adjustment for free throw rate and percentage using last year's numbers.

We will also provide expected best and worst case scenarios by using the three top and bottom comps from individual samples. For all players, we took the three biggest single-year comparable jumps and smallest comparable jumps (or biggest declines) and applied the average of those three to each player's 2021-22 season at Marquette. Let's count down the three most likely candidates to lead the team in scoring.

#3 - Tyler Kolek
Photo from espn.com

In his first year at Marquette, Kolek became far more of a creator than he was at George Mason as his percent shots and eFG% both dropped. While that seems like a negative harbinger for his scoring, it's worth noting that the comps we found of players moving up a level often went through similar situations. Looking at Kolek's comparable players, when you counted their pre-transfer seasons twice and their first season at the new program once, then averaged those three, it had a near perfect second-year post-transfer correlation. The respective percentages were accurate to 1.6% of minutes, 0.2% of percent shots, and 0.02% of eFG%. We used that transfer correlation for Kolek only (it didn't fare the same with O-Max's comps). Let's check the numbers:

%Min %Shot eFG% PPG
Pre-Transfer 76.8 25.0 50.3
Transfer Year 1 60.5 19.2 49.3
Transfer Year 2 73.5 23.0 50.0

Tyler Kolek (PT) 75.5 19.5 53.1 10.8
Tyler Kolek (TY1) 72.3 16.7 40.0 6.7
Tyler Kolek (22-23) 74.4 18.6 48.7 9.3
Tyler Kolek (Worst) 76.6 18.2 38.4 7.7
Tyler Kolek (Best) 80.4 22.4 48.7 12.2

The first thing to note is that based on his comparable players, Kolek is highly likely to bounce back in terms of eFG%. Most of the comparable players that saw an eFG dip in their first post-transfer year improved that in their next year. With an expected increase in minutes and usage, Kolek will almost certainly increase his scoring as well. I was a little surprised by the 12.2 ppg best case comparison. That will likely depend on his role in 2022-23. If he continues to primarily play point guard and run the offense, that number is probably the best case scenario, while if he moves off the ball his percent of shots taken and eFG% could both be higher.

It's worth noting we have already seen improvement with Kolek in his first year. He shot just 22% from deep in non-conference games but improved that to 33.3% in Big East play. He also hit 51.4% of his catch-and-shoot threes, so if others are creating for Kolek rather than him creating for himself (he hit a dismal 14.7% on threes off the dribble) he could even exceed his best case scenario. In addition, the early reports out of camp is that Kolek looks like the most improved player on the team and has been scorching the nets after apparently taking the criticisms of his shooting last season personally. Expecting him to be around double-digit scoring is a relatively safe assumption.

Tyler Kolek Comparisons: Jared Bynum, Donald Carey, Torin Dorn, Elijah Harkless, Ithiel Horton, Koby McEwen, Quincy McKnight, Marcquise Reed, Andrew Rowsey, Eric Williams

Worst Case Comparisons: Ithiel Horton, Quincy McKnight, Eric Williams

Best Case Comparisons: Jared Bynum, Elijah Harkless, Koby McEwen 

#2 - Olivier-Maxence Prosper
Photo from Marquette Athletics

Already showing up on some of the 2023 NBA Draft boards, O-Max seems like the safest bet to make a jump in 2022-23. Our comparisons back that up, as even his worst-case options project to improve his percent of shots and eFG%. Quite simply, similar statistical players that played limited minutes in their first collegiate year, then transferred and saw big minute jumps like Prosper did in their second year tended to continue ascending in year three. There was no equivalent comparison like Kolek that incorporated the pre-transfer year, so we strictly did transfer year one to year two improvement percentages. With Justin Lewis gone, it seems likely Prosper will get every chance to increase his scoring.

%Min %Shot eFG% PPG
Pre-Transfer 21.8 16.3 45.2
Transfer Year 1 51.7 17.8 51.8
Transfer Year 2 60.2 20.2 56.4

O-Max Prosper (PT) 22.3 17.3 37.5 2.5
O-Max Prosper (TY1) 51.2 16.8 51.8 6.6
O-Max Prosper (22-23) 59.5 19.1 56.4 9.4
O-Max Prosper (Worst) 47.1 19.5 53.7 7.3
O-Max Prosper (Best) 77.6 18.2 62.3 12.7

In all honesty, the best-case scenario might be the most realistic for Prosper, and it's possible he could even exceed that. 80% of the comparable players improved in percent of shots taken, 80% improved in eFG%, and 90% increased their minutes. Double-digit scoring seems highly likely as he will get all the shots he previously got and many that Justin would've taken last year. Though while the gut feel is that the best case scenario is attainable, it did seem a bit surprising that the second highest scoring projection was still under 10 ppg.

Olivier-Maxence Prosper Comparisons: Nick Babb, Jemarl Baker, Colin Castleton, Luke Fischer, Dan Fitzgerald, Anton Gill, Myke Henry, Tariq Owens, Jonathan Tchamwa Tchatchouwa, Jamil Wilson

Worst Case Comparisons: Jemarl Baker, Colin Castleton, Myke Henry

Best Case Comparisons: Nick Babb, Anton Gill, Tariq Owens

#1 - Kam Jones
Photo from jsonline.com

I imagine it's no surprise to see Kam Jones as the expected leader on this list. His three point shooting ability, high usage, and a long but productive comparable list really lead to high expectations. However, there are some rather unencouraging comparisons that keep us from ruling out a sophomore slump and while his best case scenario is lofty, it isn't the highest best case scenario we could see, but that will come in the next part. Let's break out the numbers:

  %Min %Shot eFG% PPG
Freshmen Sample 43.8 20.7 48.7  
Sophomore Sample 64.1 21.9 49.8  

Kam Jones (Fr) 44.5 22.9 55.9 7.4
Kam Jones (22-23) 65.1 24.2 57.2 11.3
Kam Jones (Worst) 47.8 23.0 51.1 7.1
Kam Jones (Best) 84.6 25.9 68.9 18.9

The average improvements for Kam's comps come up favorable, but considering how far he was ahead of the average eFG%, it seems unlikely he will jump far (almost certainly not to the best case scenario). This admittedly includes some players who saw precipitous eFG% drops, such as Markus Howard and Steve Novak, who mostly fell because they were otherworldly in terms of shooting the ball as freshmen.

I would caution those excited by the big numbers in Jones' best case scenario, a guard reaching 68.9% eFG% is highly unlikely. Not impossible, but his three best case scenario comps all came from players who had freshman year eFG% in the 40s, so they had significantly more room to grow and it was easier for them to have large percentage jumps because of their initial low shooting percentages, unlike Jones for whom it will be harder to improve on 55.9%. I think Jones is the most likely player to lead the team in scoring and reach double-digits but he might also be the least likely to hit his best case scenario.

Kam Jones Player Comps: Brendan Bailey, Justin Blake, Vander Blue, Sandy Cohen, Eric Davis, Jase Febres, Lazar Hayward, Markus Howard, Ryan Kreklow, Justin Lewis, Dameon Mason, Wesley Matthews, Shemiye McLendon, Steve Novak, Kerwin Roach

Worst Case Scenarios: Justin Blake, Eric Davis, Shemiye McLendon

Best Case Scenarios: Brendan Bailey, Jase Febres, Justin Lewis

Rapid Takeaways
At a glance, I felt like these numbers, at least the most likely projections, felt low. Of the six returning players with the highest minutes, these were indeed the three that projected to have the highest points per game output. That said, there is more encouraging news on the way. While scoring won't likely be as consolidated as it was last year (remember, Justin Lewis' 16.8 ppg is the highest ever for a Shaka player) it seems likely that the scoring will be more spread out. For instance, these top scorers all have an expected 2022-23 baseline of over 9.0 ppg, whereas in 2021-22 only two players (Lewis & Morsell) exceeded 8.0 ppg.

Thursday, June 23, 2022

Who Will Score? Part II: Shaka Smart

Shaka Smart will say goodbye to leading scorer Justin Lewis
 Photo by Mike de Sisti | Milwaukee Journal-Sentinel

In our last edition, Cracked Sidewalks looked at Marquette's history when it comes to the loss of volume points per game scoring and whether that has an impact on subsequent offensive efficiency. Because Marquette is losing four of the top seven points per game scorers, the obvious question we are trying to answer is "Who Will Score" in 2022-23.

Today, Cracked Sidewalks has gone back and looked at the entire coaching career of Shaka Smart. Due to relevance, we are also looking at the final seasons of the coaches that preceded him: Anthony Grant at VCU, Rick Barnes at Texas, and Steve Wojciechowski at Marquette. Using the same method as the last article, we are starting with the total number of points scored. We are then looking at the number of significant points lost, which were points scored by players that averaged 5.0 ppg or more. Using those two numbers, we could establish the percentage of significant points lost and match that up with the team's Adjusted Offensive Efficiency Rank as well as the Rank Change for the following year once they lost that percentage of points.

These seasons include Smart's entire career, with the "VCU/TX/MU" designations indicating the years before Coach Smart took over so we can see what impact his hiring had. We will again start with the raw data and then break it down into smaller bits to look at trends.

Year Total Pts Sig Pts Lost % Pts AdjO Rank Rk Change
2021-22 2369 1329 56.1 64 ??
2020-21 MU 1883 1450 77.0 94 30
2020-21 2008 1069 53.2 28 2
2019-20 1989 0 0.0 153 125
2018-19 2628 1375 52.3 29 -124
2017-18 2450 802 32.7 89 60
2016-17 2215 1084 48.9 177 88
2015-16 2355 1226 52.1 49 -128
2014-15 TX 2290 675 29.5 47 -2
2014-15 2609 934 35.8 58 0
2013-14 2624 749 28.5 106 48
2012-13 2770 680 24.5 20 -86
2011-12 2448 484 19.7 96 76
2010-11 2864 1671 58.3 47 -39
2009-10 2737 598 21.8 28 -19
2008-09 VCU 2440 937 38.4 64 36

While Smart's 2021-22 team sees 56.1% of the significant scoring depart, his team last year had more to replace, as did his 2010-11 Final Four team, with nearly opposite results. Last year they replaced 77.0% of the scoring and saw the Adjusted Offensive Efficiency improve by 30 spots while in 2011 he lost 58.3% of the scoring and saw the Adjusted Offensive Efficiency decline by 39 spots. Let's group it out to see if there are any meaningful trends.

Major Point Losses: 60+%

Year Total Pts Sig Pts Lost % Pts AdjO Rank Rk Change
2020-21 MU 1883 1450 77.0 94 30

Compared to Marquette's own history, Smart has quite the short list. The only time he had to replace more than 60% of his scoring was last year when he took over Marquette from Steve Wojciechowski and saw a significant Offensive Efficiency improvement. Bear in mind the only number here that is actually Smart's responsibility is the "Rank Change" improvement. If this says anything, it might be that Smart is better at balancing classes and retaining players so that the only time he's ever been in position to replace so much was taking over a new program.

Moderate Point Losses: 40-59%

Year Total Pts Sig Pts Lost % Pts AdjO Rank Rk Change
2010-11 2864 1671 58.3 47 -39
2021-22 2369 1329 56.1 64 ??
2020-21 2008 1069 53.2 28 2
2018-19 2628 1375 52.3 29 -124
2015-16 2355 1226 52.1 49 -128
2016-17 2215 1084 48.9 177 88

Hard to look at this and not see those two massive declines. The -124 and -128 are the two biggest drops in the 37 total seasons we reviewed between Smart's career and Marquette's history. And yet Smart's second greatest single-season improvement is also in this category. It's worth noting the 2 point improvement was made by Chris Beard as that was Smart's last season at Texas. The data here indicates Smart has at times struggled replacing significant scoring, though the small sample size and the 2017 outlier definitely shows it's not a definitive struggle.

Minor Point Losses: 20-39%

Year Total Pts Sig Pts Lost % Pts AdjO Rank Rk Change
2008-09 VCU 2440 937 38.4 64 36
2014-15 2609 934 35.8 58 0
2017-18 2450 802 32.7 89 60
2014-15 TX 2290 675 29.5 47 -2
2013-14 2624 749 28.5 106 48
2012-13 2770 680 24.5 20 -86
2009-10 2737 598 21.8 28 -19

This brings us back to the essential "nothing matters" conclusion. Three seasons where most of the scoring is retained end up in improvements, three end up in declines (including bizarrely the two that kept the most), and one had no change (though that was in a coaching change to Will Wade).

Minimal Point Losses: 0-19%

Year Total Pts Sig Pts Lost % Pts AdjO Rank Rk Change
2011-12 2448 484 19.7 96 76
2019-20 1989 0 0.0 153 125

Another small sample size, but retaining 80+% of his scoring led to big improvements in the rare events when it's happened. The general takeaway seems to be that while losing scoring won't necessarily hurt you, retaining scoring seems likely to help. Though all one has to do to question that is look at the two entries immediately above these minimal loss results.


In total, there were 37 past seasons reviewed between Marquette's history and Smart's career. When looking at the most extreme outliers, there were four seasons classified as minimal loss (less than 20% of significant scoring lost) and four seasons classified as major loss (60% or greater significant scoring lost). With last year's 77% loss resulting in a 30-spot improvement with the transition from Wojciechowski to Smart being the outlier, the other seven results all resulted in improvement or decline as would be expected. So at the extremes, keeping scoring is good and losing scoring is bad.

The scatter plot below supports that. When you look at the seasons where the Percent of Points Lost (X-Axis) is below 20%, it always yields positive results in the AORtg Rank Change (Y-Axis). When you look at seasons where the Percent of Points Lost is greater than 60%, it usually yields negative results (2021-22 was the 77% outlier).

However in the minor and moderate ranges, highlighted yellow on the plot, the results were all over the place. In 17 seasons of minor loss (20-39%), there were 8 teams that improved, 8 teams that declined, and one team that stayed the same, with the range being anywhere from an 86 spot decline to a 108 spot improvement. In 12 seasons of moderate loss (40-59%), there were 6 teams that improved and 6 teams that declined, with the range being anywhere from a 128 spot decline to an 88 spot improvement.

Trivia Answer: Justin Lewis' 16.8 ppg is the most ever for a player under Shaka Smart
Photo by Aaron Gash | AP Photo

Back to our trivia question from our preview article, Justin Lewis' 16.8 ppg is the most ever scored by a player on a Shaka Smart team. If anything, that should be reassuring to fans as typically, Smart simply doesn't have offenses that heavily revolve around one player. Replacing scoring is always going to be a team effort. Treveon Graham (16.2 ppg in 2014-15, 15.8 ppg in 2013-14, 15.1 ppg in 2012-13), Jamie Skeen (15.7 ppg in 2010-11), and Isaiah Taylor (15.0 ppg in 2015-16) are the only other players to reach the 15 ppg mark. In general, the scoring load for Smart is spread among a number of players rather than centered on one or two individual talents like we've sometimes seen at Marquette in recent years.

If there's any takeaway, it's that when you lose between 20-60% of your scoring, there's virtually no correlation between an offensive efficiency improvement or decline. Statistically speaking, replacing scoring is not something fans should get worked up about.