Rookie-scale Option Decisions


Oct 14, 2014; New Orleans, LA, USA; New Orleans Pelicans forward Anthony Davis (23) dunks the ball against the Houston Rockets during the first quarter at the Smoothie King Center. Mandatory Credit: Derick E. Hingle-USA TODAY Sports

Few topics draw a stronger emotional response than pessimism towards young players. NBA teams selecting near the top of the draft are investing their sole reward for suffering through a miserable season, making the resulting players the focal source of hope for dejected fans. This hope tends to soldier forth, even when the on-court production falls well below expectations. September and October is the time when every fan is convinced their underwhelming lottery picks are just one offseason of “losing 15 pounds”/”training with Kobe”/”working out with a former navy seal”/”having their [vision : deviated septum : tonsils : ingrown hair] corrected”/”working with a shooting-coach”… from leading them to the promised land. It often seems that front-office personnel feel a similar emotional investment in their young players.

This response is not very surprising. Teams select players who they think will become valuable contributors. Fans may or may not initially share this evaluation, but most eventually fall in line. This sets everyone’s expectations to “positive.” Positivity is probably a wise approach for psychological well-being, but it creates a collective atmosphere of confirmation bias. People automatically focus on events that support their view of the world, but deftly ignore those that do not. Add in the community element of similarly incentivized observers, and it takes overwhelming evidence to reach the point where folks know it just is not going to work out.

Sustained support of poor-performing youngsters is bolstered by the popular opinion that you cannot judge a prospect early in his career. There are certainly exceptions, but usually you know a ton about a player just one season into his career. The standards are much lower for younger players, but generally good young players become great players and bad young players eventually drop out of the league. There is nothing wrong with a little hope, especially in the service of supporting young men in their careers. However, the necessary roster spots and cap-space come at the expense of fringe prospects looking for a chance, and useful veterans hoping to continue contributing in the league. These opportunity costs should become a major consideration as we approach the October 31 deadline for picking up rookie-scale options.

Each year, teams need to evaluate young players coming off of their 1st, 2nd, or 3rd year in the league and decide whether they want to make them an offer for their 3rd, 4th, and 5th year. This puts teams in the awkward position of deciding in 2014 whether a still-developing player will be worth tying up cap and roster space in 2016. Rookie-scale contracts are appropriately considered a great bargain in the NBA; however, it seems teams are often a bit too quick to pick-up player options due to conservative thinking and overestimation of players’ odds of figuring things out and becoming stars. In this article, I briefly discuss a model I made to project the future production of young NBA players, and apply this model to former lottery picks at each stage of the rookie-contract process.

Rookie-scale projection model:

Everyone knows that young players improve as they mature and gain experience. The relationship between age/experience and production has been captured in many studies showing expected improvement between seasons. This approach is understandably unpersuasive to those considering the tradeoffs of abandoning a young prospect. Sure, the guy hasn’t been very good, but you do not want him to suddenly realize his true potential then hit free-agency the following summer. The concept of linear improvement does not fit the intuitive model that fans are really working from. The gamble is all about paying in cap and roster flexibility while hoping the prospect explodes into a highly-productive and underpaid player.

In order to capture this approach, I built multinomial regression models to generate the odds a given player’s career will ultimately peak at one of five different levels:

“Bust”:  Never worth more than 2.5 wins in a season.

“Bench warmer”:  Peak value between 2.5 and 5 wins.

“Starter”:  Peak value between 5 and 7.5 wins.

“Stud”:  Peak value between 7.5 and 10 wins.

“Star”:  Ultimately worth more than 10 wins in a season.

[“Wins” are determined using a combination of Win Shares and RAPM-wins]

The models are trained on 20 years of historical data and use age, experience, NBA production, and prior expectations to make predictions.

Apr 14, 2014; Chicago, IL, USA; Orlando Magic guard Victor Oladipo (5) dribbles the ball against Chicago Bulls guard Kirk Hinrich (12) during the first quarter at the United Center. Mandatory Credit: Mike DiNovo-USA TODAY Sports

1-Year Players:

The best model for projecting players who have completed only one NBA season uses: age, height, rookie season production (the same “Wins” formula noted above), draft slot, and “EWP” my draft projection model based on college and international production (you can browse projections here).

The table below shows the weight of each factor in the model.  The values in each column give the “importance” of each factor for explaining the likelihood a player will peak at the level indicated by each row.  Positive values indicate “more likely” and negative values indicate “less likely”.  I recommend not focusing on the raw size of the values, but rather their size relative to the other factors:


[“Draft” is coded such that players drafted early have a higher score]

Unsurprisingly, actual NBA production is the most important factor, more than twice the combined impact of the two measures of prior expectations. That said, expectations still matter. I have previously made the claim that a player’s draft slot becomes irrelevant the second he steps on the court. This is not true. Interestingly, draft slot does nothing to help discriminate between bench players, starters, and studs after controlling for other factors, but it does help highlight future stars. My EWP measure also continues to have value in projecting players after a single season, and in a more regular fashion across levels of production than draft slot. Players are also penalized for age and height. The height effect may be surprising given the popular idea that big-men are slower to develop; however, players are typically as good as they will ever be at “big man skills” like rebounding and blocking in their rookie season.

Here is how the 2013 lottery picks look according to this model:

Anthony Bennett32%49%16%3%1%5,803,560
Victor Oladipo0%35%38%18%10%5,192,520
Otto Porter22%38%30%7%3%4,662,960
Cody Zeller0%22%48%22%9%4,204,200
Alex Len44%34%19%3%1%3,807,120
Ben McLemore11%44%32%11%3%3,156,600
Kentavious Caldwell-Pope0%29%42%22%7%2,891,760
Trey Burke5%40%34%16%5%2,658,240
C.J. McCollum49%36%12%2%0%2,525,160
Michael Carter-Williams5%32%42%18%3%2,399,040
Steven Adams0%27%46%22%5%2,279,040
Kelly Olynyk18%41%33%7%1%2,165,160
Shabazz Muhammad63%31%5%1%0%2,056,920

Clear opt-ins: Last season’s rookie crop was one of the worst on record. No player looks likely to ever reach “Star” or even “Stud” quality (though my personal favorite, Nerlens Noel, is yet to log a minute of regular season action). That said, rookie contracts are great deals even for non-star contributors. The quartet of Adams, Carter-Williams, Caldwell-Pope, and Burke all warrant two more years of commitment thanks to decent production and cheap contracts. Each is very likely to be worth more than his contract in 2016. Zeller and Oladipo are a bit more costly, but also have the best odds of growing into “Stars” (though admittedly not very good odds).

Marginal keepers: Ben McLemore had a very unimpressive rookie campaign, and this was anticipated by his college numbers. However, McLemore still scores nearly a coin-flip chance of ultimately developing into at least an NBA starter. Otto Porter is in a similar situation, but as the #3 pick his price-tag is less friendly. There are situations where cutting Porter loose would be the right play, but especially given the context of his injury and limited opportunity, I would keep him under contract. Alex Len and C.J. McCollum both looked like decent but flawed prospects coming out of college and had very rough rookie seasons. I am giving both a pass since injury issues clearly muddied their performance. Kelly Olynyk had a decent season, but after accounting for his age he did not perform like a future high-impact player.

Let them walk: Joe Alexander, Patrick O’Bryant, Kendall Marshall, and Yaroslav Korolev are the only lottery picks I can think of whose teams decided not to pick up their third-year options. Despite the historical ineptitude of this class, it does not look like any names will be added to the list. I think this is the wrong decision in a couple of cases.

Anthony Bennett had one of the worst seasons ever for a #1 pick. This is a big deal since top picks start making significant money early in their careers. Bennett is slated to draw a starter’s paycheck in 2016, but the numbers have him at only a 20% chance of ever being worth that much. More likely, Bennett will be a marginal rotation player on a contract that cannot be moved without a pick in tow. The Timberwolves’ other second year player, Shabazz Muhammad, is only due $2-million in 2016. My concern with picking up his option has less to do with the salary hit, and more to do with the roster space. Shabazz’s odds of ultimately developing into at least a starter are only 6% based on college production, draft-slot, age, and rookie production. In contrast, second-rounders typically have better than a 10% chance of doing the same. Continuing to invest in players who do not work out is a decision not to invest in D-Leaguers, under-the-radar prospects, and international players. These tradeoffs are nicely demonstrated by the two second-round picks Minnesota sold for roughly the price of Muhammad’s 2016 salary, and the current concern over how to make room for rookie Glenn Robinson Jr.

Oct 9, 2014; Auburn Hills, MI, USA; Detroit Pistons center Andre Drummond (0) looks on during the second quarter against the Milwaukee Bucks at The Palace of Auburn Hills. Mandatory Credit: Raj Mehta-USA TODAY Sports

2-Year Players:

The best model for projecting players with two seasons under their belts is similar to the model above, but looks at production in both years one and two:

%AgeYear 2Year 1DraftEWPHeight

Here actual NBA production begins to clearly outpace other factors. The combined importance of the 1st and 2nd year is about 4-times more important that any expectations for the player entering the league.

Here is how the 2012 lottery picks look according to this model:

Anthony Davis0%0%4%14%82%7,070,730
Michael Kidd-Gilchrist0%18%39%28%16%6,331,403
Bradley Beal0%11%34%30%25%5,694,674
Dion Waiters0%63%28%8%1%5,138,430
Jonas Valanciunas0%13%52%27%8%4,660,482
Thomas Robinson30%55%14%2%0%4,660,482
Damian Lillard0%0%23%51%27%4,236,286
Harrison Barnes0%38%42%17%2%3,873,398
Terrence Ross0%76%21%3%0%3,553,917
Andre Drummond0%0%21%25%54%3,272,090
Austin Rivers52%44%3%0%0%3,110,796
Meyers Leonard68%29%4%0%0%3,075,879
Jeremy Lamb0%65%29%6%0%3,034,356
Kendall Marshall44%47%8%1%0%waived
John Henson0%58%35%7%0%2,943,221

Clear opt-ins: Anthony Davis looked like the best NCAA prospect since Shaq. He has since backed that up with two seasons of dominant production (though his RAPM has been surprisingly mediocre). Davis’ option is the easiest decision a team will ever make. The 4% chance that he fails to become more than just a starter is basically the odds he suffers a career-ending injury. Andre Drummond and Damian Lillard are also obvious choices. Bradley Beal has a 55% chance of developing into a “Stud” or “Star”, which would be a major steal for only $5-million. Michael Kidd-Gilchrist has been a disappointment, and his contract is starting to get pricey, but he has shown enough potential to remain an obvious keeper. Jonas Valanciunas is very likely to at least earn his contract, and still has the potential to become a great player.

Marginal keepers: Harrison Barnes’ projections give him good odds of growing into a starting role, and show some additional upside on top of that. John Henson and Jeremy Lamb are both proven basketball players. The odds are not good that either develops into anything special, but they deserve roster spots and are not slated to make too much money. Terrence Ross is a bit more questionable due to his higher salary commitment, but he seems a safe bet as well. Dion Waiters is the final player in this middling class of production. Waiters’ price-tag is a bit higher, putting him in the mid-level exception range for 2016. Waiters’ “upside” is not completely dead after two mediocre seasons, but it is on life-support. The best strategy with his option really depends on how Cleveland wants to manage contracts as they transition into annual contender mode. Overpaying limited players can be a major problem for sustained success, but Bird Rights can trump cap concerns if ownership is not worried about the luxury tax.

Let them walk: Austin Rivers, Thomas Robinson, and Meyers Leonard. These players have done little to demonstrate their ability to compete in the NBA. It is highly unlikely that this changes in 2015. The cap and roster flexibility offered by clearing them in 2016 seems an easy choice. These cases are likely less controversial than the third-year options I recommended waiving. All three may ultimately have their options picked up, but there is at least serious consideration in New Orleans and Portland about how to handle these players.     

Aug 22, 2014; New York, NY, USA; United States guard Kyrie Irving (10) controls the ball in front of Puerto Rico forward Alexander Franklin (6) during the first quarter of a game at Madison Square Garden. Mandatory Credit: Brad Penner-USA TODAY Sports

3-Year Players:

By the time players have finished their third season, the 5th year qualifying offer is usually an obvious choice. These players still have a few years of expected improvement, but the rate is starting to slow down, and the stars have demonstrated their ability while the busts have begun washing out.

%AgeYear 3Year 1+2EWPHeight

Draft-slot no longer carries any information at this stage, but EWP still lingers as a weak predictor. Most of the weight is on actual NBA performance in the current and previous seasons. The combined importance of the first three years in the NBA carries about 14-times more information than any expectations you had for a player entering the league.

Here is how the 2012 rookie class looks according to this model:

Kyrie Irving0%0%41%39%20%Maxed
Derrick Williams0%67%29%4%0%8,262,482
Enes Kanter0%71%26%4%0%7,471,412
Tristan Thompson0%16%58%23%2%6,777,589
Ricky Rubio0%0%32%56%12%6,723,730
Jan Vesely48%49%3%0%0%waived
Bismack Biyombo0%56%36%7%0%5,194,227
Brandon Knight0%62%32%6%0%4,790,680
Kemba Walker0%0%61%36%3%4,433,683
Jimmer Fredette94%6%0%0%0%waived
Klay Thompson0%0%69%29%2%4,210,880
Alec Burks0%70%26%4%0%4,175,274
Markieff Morris0%0%91%9%0%Extended
Marcus Morris0%79%19%2%0%Extended

Clear opt-ins: Kyrie Irving already signed a max contract. You can debate whether he deserves that, but he is a clear bargain at his qualifying offer. Ricky, Klay, Tristan, and Kemba are also easy decisions. These are all starters on good contracts with the added benefit of some remaining untapped potential. The real decision on these players comes next season when other teams have the opportunity to raise the pot. Neither of the Morrii looks like a future star, but both were a near lock to be underpaid on their qualifying offers. Phoenix made things easy by extending both to four-year contracts.

Marginal keepers: None.

Let them walk: Most of the debate here stems more from interpretation of players’ production in the NBA rather than expected future improvement. The obvious cases, Jimmer and Vesely, have already been dealt with. Derrick Williams and Enes Kanter should both be relatively easy decisions as well. Their top-3 pick status and corresponding contracts are disproportionate to their on-court impact. Both players deserve to be in the NBA, but Sacramento and Utah are unlikely to regret letting them hit free-agency. Knight, Biyombo, and Burks are all slightly tougher decisions. None of the three are bad players and their qualifying offers are not egregious. However, after three years of mediocre production it is highly unlikely any of them is going to find another gear in 2015, and their expected production should be replaceable at a fraction of the price.