Elijah Johnson, Jeff Withey unlikely to assume same offensive roles as Tyshawn Taylor, Thomas Robinson


With eight new scholarship freshmen on the roster, it's hard to predict exactly how the Kansas men's basketball team's offensive roles will establish themselves for the 2012-13 season.

If history is any indication, though, KU fans shouldn't expect seniors Elijah Johnson and Jeff Withey to do the same heavy lifting offensively that departed players Tyshawn Taylor and Thomas Robinson did in 2011-12.

Elijah Johnson (15) along with from left, Jeff Withey, Tyshawn Taylor, and Travis Releford, laugh during a huddle in the second-half of the Jayhawks 60-57 win over North Carolina State in St. Louis Friday, March 16, 2012...

Elijah Johnson (15) along with from left, Jeff Withey, Tyshawn Taylor, and Travis Releford, laugh during a huddle in the second-half of the Jayhawks 60-57 win over North Carolina State in St. Louis Friday, March 16, 2012... by Mike Yoder

The reason for this is a 2007 study from Ken Pomeroy that attempted to look at players' offensive roles from one year to the next.

For this study, he used possession percentage, which is the percentage of possessions a player ends by making or missing a shot or creating a turnover with a few adjustments made for offensive rebounds and assists (20 percent is average).

After looking at possession percentages of players one year to the next, Pomeroy came to the following conclusion in his study:

"Players do jump from being decoys to go-to guys in one season, and some even regress the other way. Those are the exceptions. By and large, a player's role on his team in one season is a good indicator of his role the following season."

One of the examples he used from the time was Duke's Josh McRoberts. After playing his freshman year with J.J. Redick and Shelden Williams, most Blue Devils fans expected him to carry the load offensively during his sophomore season.

There was only one problem: McRoberts played passively his freshman year (17.1 percent possession percentage), and it's hard for a player to dramatically shift his role from one year to the next.

Though McRoberts' possession percentage went up to 21.9 percent his sophomore year, Pomeroy said the forward still received criticism for not taking over games.

Let's take a look at last year. Did Taylor and Robinson have the offensive profiles to suggest they could become go-to guys?

Kansas forward Thomas Robinson, top, hugs teammate Tyshawn Taylor after Taylor's feed to Robinson for a dunk against South Florida during the second half on Saturday, Dec. 3, 2011 at Allen Fieldhouse.

Kansas forward Thomas Robinson, top, hugs teammate Tyshawn Taylor after Taylor's feed to Robinson for a dunk against South Florida during the second half on Saturday, Dec. 3, 2011 at Allen Fieldhouse. by Nick Krug

Here are both of their possession percentages from the past two seasons, according to

2010-11 — Taylor 21.2 percent, Robinson 26.7 percent
2011-12 — Taylor 27.7 percent, Robinson 29.7 percent

Taylor's jump in offensive involvement was significant — according to Pomeroy's study, only about 1 in 20 players will experience a usage increase this large from one season to the next. Still, Taylor was an above-average offensive contributor his junior year, so a jump to 26.7 percent wasn't completely crazy.

Robinson, meanwhile, was a high-usage guy even when he wasn't the focal point of the offense his sophomore year. It shouldn't have come as any surprise that he could handle a go-to guy role for KU last season.

So what about Johnson and Withey, the two guys who are being expected to produce the most offensively for KU next season?

Here's a look at their possession percentages from a year ago:

Johnson — 17.5 percent
Withey — 18.0 percent

At times last year, both players drew criticism for being too passive offensively. For Johnson, this was mostly focused on his lack of aggressiveness with penetration, as he had just 46 free-throw attempts while hoisting up 64 more three-pointers than two-pointers.

Withey also rarely looked for his own shot, with many of his attempts coming off open looks created by assists (before the Final Four games, a whopping 79.7 percent of his "close twos" were assisted last year).

So what does this all mean?

Well, if Pomeroy's study holds true today (he told me that it should with the amount of data he used), KU fans shouldn't expect Johnson and Withey to immediately step in and become the offensive contributors that Taylor and Robinson were a year ago.

Kansas forward Thomas Robinson, center, jokingly pushes center Jeff Withey away from his golf cart as the players prepare to be shuttled to interviews on Saturday, March 24, 2012 at the Edward Jones Dome in St. Louis. At left is KU guard Elijah Johnson.

Kansas forward Thomas Robinson, center, jokingly pushes center Jeff Withey away from his golf cart as the players prepare to be shuttled to interviews on Saturday, March 24, 2012 at the Edward Jones Dome in St. Louis. At left is KU guard Elijah Johnson. by Nick Krug

Though Pomeroy told me it's not impossible for players to make possession percentage leaps from the teens into the high-20s, more than likely, both players will end up in the 20-24 percent range.

That would leave a lot of possessions unclaimed for KU.

So who might pick those up?

Kevin Young is a possibility (19.3 percent), though he needs to improve his defense and reduce his fouls to pick up increased minutes.

Travis Releford, meanwhile, seems unlikely to take on a huge role, as he posted the second-lowest possession percentage of KU's regulars last season (13.9 percent).

It appears, then, that there is an opportunity for freshmen Ben McLemore and Perry Ellis (and potentially Anrio Adams and Andrew White) to make a big offensive impact for KU in their first years.

In all likelihood, KU's offense will be more balanced in 2012-13, with the Jayhawks needing a few good freshmen to immediately step into scoring roles.


clevelandjayhawker 10 years, 1 month ago

thank you for posting an article that is worth reading and gets me excited for basketball (other than HORSE of course)

Dave McClain 10 years, 1 month ago

So, Taylor and Robinson added up to 56.4% possession percentage last season. I wonder if Pomeroy has statistics for situations where two moved on that added to more than 50% and how the remaining two key players faired the next year. I don't disagree with the balance we should probably have this next season, but I also have to believe that the big man will be under good coaching pressure to step up his possession percentage.

Ken Tomberlin 10 years, 1 month ago

If Withey and Johnson add up to a 45% possession percentage, that would seem to be healthier more balanced offense than two players amounting to 56.4%. We have some good talent and looking forward to an offense with multiple threats.

Jesse Newell 10 years, 1 month ago

konkey — You make some good points.

And you are absolutely right at the end: usage rate doesn't help us to predict how well a team/player is going to play.

Usage rate might, however, offer some clues to the balance of next year's team. KU hasn't had many years where it's had two players dominate offensive possessions like Taylor and Robinson did last year. Based on last year's stats, it doesn't look like Johnson and Withey will be those types of players (though they still will probably lead KU in the stat and will be appropriately labeled as KU's go-to guys).

Could KU's offense be better with more balance? Absolutely. The point here isn't to say that KU's offense will be better or suffer without two extreme possession-users, it's to say that KU's offense most likely will have a different dynamic next year, mostly because the numbers would suggest that Johnson and Withey probably won't make monumental possession leaps in one year.

I guess if I were going to compare on balance (not talent), I would think KU next year will more resemble 2009-10 (with Collins, Aldrich, Henry and the Morris twins all using their share of possessions) than last year (Taylor and Robinson taking most of the offensive load).

As I mentioned on Twitter earlier today, it's pretty amazing how Self has had offensive success both with balanced teams (2008, 2010) and also unbalanced ones (2009, 2011, 2012).

Steve Kubler 10 years, 1 month ago

Although Withey showed flashes of offense in various games I have wondered if he, and Elijah, have the moxy to be the main guys in the upcoming season as Taylor and Robinson was last. My other concern is at guard and who is going to step up and fill that role beside Elijah.

I would like to see a big scoring threat in that position, but one that can handle bringing the ball up against pressure. Someone has to take some defensive pressure off of Withey, as I do not see him being able to score reliably while double teamed. Johnson stepped up during the year but will need a reliable ball handler to side him or pressure defense could become a problem. Then there will be who the 6th and 7th men on the floor will be.

It appears that playing time is wide open leaving me with a great deal of anticipation for the trip to Europe. I hope we get some solid video of those games so we can start getting a feel for what this season will bring. LJW, are you listening here?

All in all looks like HCBS will earn Cindy's money this year!

master16 10 years, 1 month ago

I definitely think we will be more balanced next season, which will affect the offensive productivity of EJ and JW in the first place. I fully expect guys like BMac and Ellis to put up some points and Rele will prob be a solid contributor too. Last I checked, Releford was known for being an offensive threat in highschool, so far we've only see very minimal flashes of his scoring ability. I hope it comes out in his senior year. Withey will no doubt be a solid defensive presence, perhaps the best in the nation, but I'm not expecting him to even come close to averaging TRob's offensive numbers.

I think Johnson does have the potential to score a lot of points simply because of his toughness and passion, but I don't see that happening unless the other guards all have bad seasons. EJ either has to work on his 3 point % or get back to attacking the rim. If he plays as tough as he did in the post season, we'll be in pretty good shape as long as the other guards come out to play too.

David Lara 10 years, 1 month ago

Interesting that they don't list his most recent accolades. A great way to wrap up the article and prove how silly the close-minded perspectives are would be to show the specific success he has achieved on the international stage.

RJ King 10 years, 1 month ago

Agreed. "...Withey also rarely looked for his own shot, . . . a whopping 80 percent of his "close twos" were assisted."

Feed the post. Feed the post. Feed the post. And yes to the need for defensive help for JW!

dtdjayhawk 10 years, 1 month ago

Don't forget about Traylor who has been practicing against T-Rob and Withey for a year and he still had Manning.Plus, all the players that will be stopping through during the summer to practice with them.

ironhead80 10 years, 1 month ago

REALLY!!!!!!!!!! Nothing else to write this time of year I guess?

RJ King 10 years, 1 month ago

The last part of your comment must have been cut off. I was unable to see where you suggested at least five other, timely, more significant, and more interesting topics for the week of June 25th.

Jesse Johnson 10 years, 1 month ago

I'm confused about your post. For me as a real KU BBall fan, these types of articles are an oasis in the middle of a desert (the off-season). I get so excited when I see a headline like this that I'm sure my face visibly lights up.

shottybotch 10 years, 1 month ago

Jesse, please ignore him. I love these articles any time of year but especially during the sports draught we call baseball season

Sam Constance 10 years, 1 month ago

This is a great article, and I loved reading it.

That being said, I don't know how much we can glean from this information, basically because of how big a departure last year's balance was from typical Jayhawk teams.

I can't remember the last time a KU team was as top heavy as last year's with TT and TR.

When I think of two-man teams at Kansas in recent history, a couple obvious examples come to mind first:

C&C Basketball Factory (Cole and Collins)

The Iowa Connection (Collison and Hinrich)

Jesse--do you happen to have readily available usage percentage for the aforementioned duos from 2009-10 (Aldrich/Collins) and 2002-03 (Collison/Hinrich)? I would be curious to see how it compares to last year's workhorse seasons from Taylor and Robinson.

I personally think the biggest gap in the usage on KU's team will be taken by Ben McLemore, with a half-season of D-I practice under his belt and the fact that he's likely going to step into one of the positions on the floor that typically takes a large scoring role (SG).

Jesse Newell 10 years, 1 month ago

StatSheet is the only site I've seen that goes back that far. And I assume you mean 2008-09 Aldrich/Collins, which was pre-Xavier Henry.

2008-09 Collins (27.9 percent); Aldrich (21.4 percent) ... Aldrich almost never turned it over that year, which lowered his usage percentage. Then again, he didn't shoot it THAT much (22.7 percent shot percentage).

2002-03 Collison (25.1 percent); Hinrich (21.9 percent) ... Hinrich had microscopic turnover rate as well ... plus, Keith Langford (21.0 percent) and pre-injury Wayne Simien (21.9 percent) were using up their share of possessions, too.

Here are the two highest duos I could find under Self behind Taylor/Robinson last year, according to StatSheet ...

2010-11 Marcus Morris (26.0 percent), Markieff Morris (25.4 percent)

2004-05 Simien (26.6 percent), Langford (24.5 percent).

RJ King 10 years, 1 month ago

Nice to know that the stats pretty much back up the "eye test."

Sam Constance 10 years, 1 month ago

Thanks! Much appreciated, and yes, assumed correctly that I meant 08-09 Aldrich and Collins.

Curtis Stutz 10 years, 1 month ago

Even though they're different players I love the comparison to the 2002-2003 team. You've got several starters returning from a team that made the Final Four, some newcomers that should contribute a lot, and some role players looking to step into the limelight a bit more. I don't think people are giving Withey credit for the tough matchup he poses defenses. He has some shooting touch and he's a big boy. He's not going to show the offensive game Collison did, few do, but he can definitely get it going if the guards feed him and he can get the occasional putback. EJ's case is a bit extreme with TT playing so many minutes and dominating offensive possessions. You have to agree with the suggestion from Kenpom's study that he has some odds to overcome to become a high usage guy, but the tools are all there. TRob wouldn't have had as high of usage his sophomore year had Marcus Morris constantly been in the game with him. The majority of EJ's minutes came with TT on the floor. In the end I think everyone feels more comfortable with balance. If 2 guys are dominating the ball it's probably not putting enough pressure on the rest of their team on defense and in the case of Collins/Aldrich can lead to becoming too predictable.

Jayhawk1958 10 years, 1 month ago

No new revelation here. We know what we lost. Just going to be like every year a different team, look, and chemistry. I'm sure Bill will be tinkering with it a lot. I still think we will have more talent next year because for example we won't have a converted walk-on be the first man off the bench. We will be longer and more athletic with a lot of different options which we can exploit with.

REHawk 10 years, 1 month ago

I envision BenMac with B Rush scoring numbers, 13+ av. If Perry Ellis can chip in 10+, we should be okay. EJ, Travis or Jeff will occasionally explode with 20 pt. outings.
Going into last season, many of us felt that the success of the team hinged upon the physical health, stamina and development of Withey, as well as the steadfast focus of TT. This year we have to hope that EJ recovers near 100% from surgery, and that high energy players KY and Tharpe show marked improvement. I look to see pleasant surprises from any of the other newcomers. D. Manning would be salivating at the opportunity to work with so many promising big men. With so many new eligible players, we might again have to be patient prior to February. But this squad bears the potential to mold itself into another championship contender, league and national.

Jim Jackson 10 years, 1 month ago

Do we know anytihng about Anrio's situation?

notjustbread 10 years, 1 month ago

Fun stuff. I for one, love, reading about Jayhawk hoops in late June. Is there an update to H-O-R-S-E picks coming tomorrow? I could be all over that as well.

Tony Bandle 10 years, 1 month ago

Biggest Impact newcomers: Perry and Andrew Biggest Impact semi-newcomers: BenMac and Jamari Biggest Impact veterans: Jeff and Elijah Wild Cards: Milt, 'Rio and ZMan

Jeff, Elijah, Travis and Kevin, whether consciously or unconciously, took a back seat to TRob and TT last year. This year they are in the front seat.

Frankly, I think Travis and Kevin are who they are and most likely won't change much one way or the other this coming season.

It should be interesting.

Alohahawk 10 years, 1 month ago

Since it's that slow (KU BB news wise) time of the year, and this is when every fan has numerous months to speculate on what may or may not happen this coming season, this article helps fill that void and re-opens those re-evaluations of team members roles. Thanks for the on-going articles, LJW. Whether one responds to the subject matter, or the discussion veers away from it, as what happened with the posts over KU's/UK's school name a few days ago, the articles give us fans the chance to voice new ideas or update old arguments. Instead of adding comments which may be missed when posted under an older article.

As for player production, I'd guess due to the sheer number of new recruits and their capabilities, that as suggested above, the scoring should be more balanced this coming season. Points production I predict would be, in descending order: McLemore, Johnson, Ellis or Releford, and then Withey. All with 10 or more ppg. In a previous post (a month ago, or so) I said that Releford would start out big, and once the new comers got comfortable, his point production would drop slightly. IMHO, Johnson will excel at point guard (barring injury/and recovery from his operation), and with the ball starting out in his hands, he will drive the lane much more than last year. If the Jayhawks can hit threes consistently, that should open up the middle for him to drive and kick for an open three.

Finally, something that would significantly effect future recruiting: "If" McLemore has an exceptional year there's a chance he could jump to the league after playing just one season. I'd hate to see him leave so soon, but if he did, it would change the thinking of a number of recruits.

pizzashuttle 10 years, 1 month ago

Did Pomeroy's study look at players moving from their junior to senior year? My guess is there would be more correlation with a higher increase in offensive production narrowed to that category. Seniors typically take more of a leadership role and would look to create more offense. That and they also know they need to impress NBA scouts since it's their last college BB season.

jaybate 10 years, 1 month ago

"Beware Large N Data Sets Used for Localized Predictions in Events Lacking Obvious Fungibility and Randomization"

I suspect KenPom's mountain of data, and the averages it generates, reveal a great deal about average situations, but sharply misleads about non-average situations.

Based on KenPom's data, last year, we would have been very confident that Tyshawn would not leap to the percentage he leaped to.

We would have been very wrong!

The quantitative question then is: would Tyshawn have been a near random anomaly, or would the data set have been improperly parsed an analyzed?

Obviously, it would have been improperly parsed and analyzed.

Here is why.

Tyshawn was obviously on a fast track trend to being a more dominant PG prior to his senior year. He had only become a full time PG for the first time in his career his junior year of D1!!!

Next, Tyshawn had been a big part of the offense with a broken shooting hand as a junior, so it was obvious that he could become much more effective with a good shooting hand his senior year.

Next, Tyshawn entered his senior season without any impact freshmen being added to the perimeter, or to the paint; this nearly ensured that Tyshawn would have to impact more.

So: Tyshawn developed exactly as one should have expected in terms of his role in the offense, given what we knew coming into last season.

What was not foreseeable was that EJ and Travis would both become so injured that they could not be credible impact options for extended periods of the season.

jaybate 10 years, 1 month ago

So: both Tyshawn's predictable development, plus other player's injuries, plus failure to recruit impact players, combined to make him have a big leap in percentage exceeding what KenPom's averages would have lead us to expect.

So: the problem with using KenPom's large N averages is that without parsing that data base to include data solely from players in an analogous developmental cycle playing on teams with injured potential impact players, and on teams that recruited no impact players, the large N data base obscures reality, rather than revealing it in a statistically significant way.

Now, just as misusing the large N data base without proper parsing and analysis can produce invalid expectations that outcomes do not conform with, so too can the same produce invalid expectations that reality conforms with without correctly explaining.

For instance, using KenPom's large N data base without parsing leads Jesse to say that the probability is that EJ and Jeff will not assume as much of a burden of the offense as did Tyshawn and Thomas.

This expectation is very likely to be fulfilled this year to some degree, but for reasons that the large N data base also obscures.

The fact is EJ's situation entering his senior year is distinguishable from the average and from Tyshawn in the following ways:

EJ played with a bad knee most of his junior year. If it heals, as Tyshawn's hand did, he should become sharply more involved, perhaps even by a large increment than Tyshawn did, as it is easier to participate with an injured hand than with an injured knee. Therefore, if other things were equal, we might expect EJ to make a very similar leap in percentage to what Tyshawn made.

But other things are not equal in this case.

EJ is not likely to have to play with injured potential impact players at the 2 and 3 most of his senior year, because there is more depth. Therefore, he is apt to have more active players sharing more of the load on the perimeter.

Further, EJ will have to play with several new freshman impact players that Tyshawn did not have to play with. Therefore we can expect even more sharing of the impact plays with these new players.

jaybate 10 years, 1 month ago

These are the easily recognizable drivers of what will determine EJ's eventual percentage of offensive participation and not KenPom's large N driven averages.

So: once again, to achieve something more than coincidental fulfillment, or refutation, of expectations, the large N data set would have to be parsed to yield a strong circumstantial likeness of player development, health of surrounding perimeter players, and similar levels of infusion of freshman impact players, in order to make reliable, statistically significant estimates about what EJ's and Jeff's eventual participation percentages will be.

I take the time to expand on this, because this is one of the most crucial and frequent errors made when laymen look at large N data sets and try to draw localized inferences from them.

A large N data set is a great thing for making inductive inferences from highly fungible things, in highly randomized error processes.

If you want to predict how many nearly identical and fungible bolts will be stamped out with in a narrow range of variance from the same machine in a period before wear and tear of the machine becomes a factor, then large N data sets are the cats meow for estimating with.

But the more variable, variously structured and variously interconnected, and so in essence, biased, localized circumstances are, the more tendency there is for unparsed, large N data sets to actually obscure the why of outcomes and so enable both erroneous forecasts and merely coincidentally accurate forecasts.

Know the data.

Know the event.

Know the circumstances of the event.

Fear what you don't know about the error factor.

Fear violation of modeling assumptions.

Never leave you thinking cap at home.

mikehawk 10 years, 1 month ago

I find the article very interesting and arrive at similar conclusions without the use or access (or enjoyment) to that level of quantitative analysis. I come at this from a more qualitative standpoint, looking at personalities in various situations and across the season. Tyshawn and Thomas were anything but quiet retiring types, and even with players on the floor in earlier years like Sherron and the twins, they were willing to take shots and assert themselves. EJ, and even more so Travis and Withey, are not the personality types to carry load minute in and minute out. They all three play defense and pick their spots offensively. Your point about Kevin Young is interesting as he has the personality type who likes to assert, take risks, makes mistakes but doesn't let it get him down or pout, and can at times in his first year with veteran players around him come in and be a "game changer." What we (or at least I don't know) what personality types were redshirted and what types enroll this Fall. Tharpe's personality had a fearless quality which put him on the bench for long periods, and yet he comes off the bench to bury a big three late in the tournament. Hopefully, he channels his talent and doesn't become tentative. No one is talking about Tharpe, but he could be a surprise next year, or the one after.

Curtis Stutz 10 years, 1 month ago

I just want to say I've been waiting patiently for the full schedule to come out for next year, so someone please point me to it if it's out!

Also this was a good article. Jaybate made some good points but I think wrote 3-4 times as much as Jesse.

Also for anyone who will be awake at 1am, the OSU@KU game is on ESPN Classic I think, but it ticks me off that during decent hours they only show Duke/UNC/UK games.

jaybate 10 years, 1 month ago

Conflation usually takes fewer words to assert than to unsnag. :-)

Michael Sillman 10 years, 1 month ago

The statistics in the article and the conclusions are compelling but I think that we might still see Johnson and Withey step up to fill most of the scoring vacuum.

KonkeyDong has already pointed out that Johnson is changing positions and his possessions will increase dramatically. With the ball in his hands he will be much more likely to drive unlike last year with Taylor around. It would be unusual for both the point guard and shooting guard to be penetrators. Someone has to hang back.

If Withey can make that hook shot his go-to move, he could become a big time scorer. You don't need to be strong to be effective with the hook. You just need to be longer than everyone else. I think it is an underutilized skill.

Commenting has been disabled for this item.