Charlotte Bobcats – Offseason Review

Since the beginning of their existence, the Bobcats have come to define mediocrity on and off the court. On the court, various collections of NBA castaways and second fiddles have combined to produce four losing seasons in the (L)Eastern Conference. The executive one-two punch of Robert Johnson and Michael Jordan seem much more interested in gracing the covers of business magazines and passing off blame to others than building a successful team. But there is some hope for this franchise in 2008-2009. The Bobcats resigned Emeka Okafor to a long-term contract, thus ensuring that one of the few decent players the franchise has cultivated would remain on the team. However, the bigger news was the signing of Larry Brown as head coach.

Brown is one of the most recognizable and respected of current NBA head coaches. While his penchant for changing teams and exerting a tyrant’s hand in the front office has lead some to question his value, those worries should not be a problem for Charlotte. Brown’s ego is likely still bruised from his ignominious exit from the New York Knicks, and he is likely to want to see his new Bobcats project through to a more successful conclusion. Furthermore, Brown is now 68, and has been voicing concerns about his family and his health for some time. Most importantly, he has a strong personal relationship with Bobcats GM Michael Jordan, a relationship much stronger than he has had with team executives in past stops. But even after pushing those questions aside, the question still remains, what can Brown do for the Bobcats?

The Brown Effect

Coaches, unlike players, are judged by one primary stat: win-loss record. In the absence of any other compelling stat, it will be interesting to see if anything more can be gleaned from this most basic statistic. First, it might be interesting to look at the records of Brown’s teams the year before he started as head coach and his first year. (Because Brown took over the Clippers mid-season, I am calling the portion of the season before the takeover “Record before LB,” and estimating the difference between Brown’s 35 games and how the Clippers would have performed in those 35 games if they maintained a consistent w-l record).

Team Record before LB Record Season 1 Difference

Cougars(ABA) 35-49 57-27 +22

Nuggets(ABA-NBA) 37-47 65-19 +28

Nets(NBA) 24-58 44-38 +20

Spurs(NBA) 31-51 21-61 -10

Clippers(NBA) 22-25 23-12 +7 (est.)

Pacers(NBA) 41-41 47-35 +6

76ers(NBA) 22-60 31-51 +9

Pistons(NBA) 50-32 54-28 +4

Knicks(NBA) 33-49 23-59 -10

Bobcats(NBA) 32-50 ?

Overall, an average of Brown’s first year win improvement is apparent (3.7 wins for just his NBA stints, and significantly higher if you factor in his ABA jobs). But such an analysis seems superficial. It doesn’t take into account that Brown mostly took over bad teams. Bad teams usually 1) have high draft picks and 2) and some imperative to improve (as evinced by firing the previous head coach and replacing him with someone else, often Larry Brown). So, on average, a team that finished with 20 wins is more likely to increase their wins by a greater factor than a team that finished with 50 wins.

Such analysis also does not properly analyze individual coaching efforts. While the Pistons only won four more games under Brown than they did under Rick Carlisle the previous year, Brown’s first year with the Pistons is arguably his best coaching job, insofar as he won the NBA title that year by defeating the heavily-favored Lakers. If you were to add playoff victories in an attempt to rectify the situation, then other coaching efforts, like Brown’s turnaround of the Clippers (who broke a significant barrier by making it into the playoffs only to lose in the first round) would be undervalued.

Properly Measuring Brown’s Effect

A more sensitive analysis might examine the difference between the records of Brown’s teams and their projected win-loss record based on points scored and points allowed (commonly called Pythagorean W-L, based upon the formula used to derive the projection). This intuitively makes sense, because good coaches would make the most difference in close games that feature teams of similar talent level, where strategy and analysis could make the difference.

So perhaps the proper analysis is to analyze the difference between the actual record and Pythagorean projection of Brown’s predecessor, the difference for Brown upon takeover (when the team is most similar to what the previous coach was working with) and Brown’s average difference (to establish a quantity with a larger sample size).

Team Pythag before LB Pythag Season 1 Average Pythag

Cougars(ABA) +1 +3 -0.5

Nuggets(ABA-NBA) -4 +6 +2.8 (+1.3 in NBA)

Nets(NBA) -2 +1 +0.5

Spurs(NBA) +2 -2 +2

Clippers(NBA) -1 -1 -1

Pacers(NBA) -5 -4 -1.8

76ers(NBA) -2 -3 -0.2

Pistons(NBA) -2 -5 -2

Knicks(NBA) -1 -1 -1

Bobcats(NBA) +3 ?

While initially promising, problems arise in interpreting anything meaningful from this data. Brown’s W-L records compared to Pythagorean projections show no clear evidence of anything, and probably require much more sensitive analysis (a problem I will return to in a later blog). My cursory analysis here shows little in the way of patterns or other useful indicators. My hunch is that Pythagorean projection work best in baseball, where runs are scored more independently of the game situation (score and time). In basketball, score and time dictate much more, as coaches put in worse players when ahead or behind by a significant amount. Baseball also has more games than basketball, creating a larger sample size.

Furthermore, Pythagorean projection has an error of 3.5 wins. Basically, since the calculation is a projection, it cannot be relied upon to be completely accurate, and the amount to which it could be regularly believed to be inaccurate is the error. As you can see, most of Brown’s individual seasons and nearly all of his cumulative totals fall within the error figure or so close as to render the results meaningless.

Note: After fiddling with how the approach Brown’s split 1991-1992 season with the Spurs and Clippers, I decided to throw the season out, since every way of approaching the problem still lead to results that fell inside the error.

I hesitate to make any strong statements about this disappointing analysis. It weakly stands for proposition that a coach’s impact on his team is a difficult, and perhaps impossible, skill to measure. (It is an interesting question whether or not the near non-existence of useful coaching stats biases statistical commentators on the NBA, such that they neglect or downplay the impact of coaches.)

What to make of the “unmeasurables”

The Bobcats eschewed free agency is year, looking more toward internal (read: cheaper) solutions in improving upon last season.

1) The team will be returning intriguing performers from injury. Neither Sean May nor Adam Morrison have had stellar NBA careers to-date, but their relative youth and initial draft position require some pause before completely writing off their careers. How they might be deployed depends upon how Brown deploys his roster:

Traditional: Beyond Okafor (and perennially frustrating Nazi Mohammed), the Bobcats have little productivity potential in the post. May’s basketball IQ, ability to score around the basket and collect offensive rebounds could be useful in traditional offensive sets, considering Okafor’s limited offensive game. However, traditional lineups required Morrison to defend stronger, quicker players and be a reliable shooter from the outside. In his rookie season of 2006-2007, Morrison was neither. While Matt Carroll was not particularly a good defender or a great shooter last year, he has shown a consistent ability to do both better than Morrison.

Up-Tempo: The Bobcats’ most common (and relatively productive) lineup last year played Okafor, Gerald Wallace, Jason Richardson, Ray Felton and Jeff McInnis. All players excluding McInnis were playing a position “larger” than their traditional position. In such a scenario, Morrison might be a more effective combo-forward, a player who role on offense is less defined and more opportunistic, similar to his role at Gonzaga. While Morrison may never be a good defensive player, accentuating his offensive game would maximize the gain the Bobcats could get from him. However, May would likely flounder in an up-tempo lineup. He would be an extremely undersized center, and placing him at a forward position negates gains in quickness.

2) Much will be made of how DJ Augustin and Larry Brown click during training camp and the coming season. Brown is notoriously hard on point guards, and if previous experience is at all telling, is unlikely to give Augustin the reigns to the team and let him play the sort of up-tempo style that the rookie seems most suited for.

It appears that the Bobcats best course of action would be to stick with one of its most productive lineups from last year, plugging in Augustin for McInnis, and making use of Morrison and Carroll as first reserves. This would marginalize May, but unlike Morrison, he is not the victim of a fluke knee injury, but several injuries that may be chronic, and the result of his lack of conditioning.

Boston Celtics - Offseason Review

Opinions differ as to who deserves credit for masterminding the Celtics’ turnaround last year. Cynics say that Kevin McHale’s ineptitude and Kevin Garnett’s will to win were the main ignition behind the return to glory, and thus, that factors mostly out of Coach Doc Rivers and General Manager Danny Ainge’s hands conspired to create the 2007-2008 Celtics. Perhaps this is true, or perhaps the public read one too many of Bill Simmons’ columns. While I don’t have much to say about Rivers (except that as a coach of a veteran group, he largely stayed out of the way, and that might not be a bad thing), Ainge’s role in the Celtics’ reconstruction merits further review.

Ainge Would Be At Home In Vegas (Except that Mormons Don’t Gamble)

Specifically, Ainge’s draft strategy highlights an interesting core philosophy: high-risk, high-reward. Below is a chart of all the picks made or traded for on draft day by Ainge in his tenure as GM, with the pick number following:

2003: Troy Bell (16), Brandon Hunter (56)

2004: Al Jefferson (15), Delonte West (24), Tony Allen (25), Justin Reed (41)

2005: Gerald Green (18), Ryan Gomes (50), Orien Greene (53)

2006: Rajon Rondo (21), Leon Powe (49)

2007: Gabe Pruitt (32), Glen Davis (35)

2008: JR Giddens (30), Bill Walker (47), Semih Erden (60)

After a rocky start in 2003, Ainge seemed to find the correct formula, and/or gained enough control over the draft “war room” to apply said formula. Over the last 5 years, Ainge has mostly taken combo guards who would never be confused with “pure point guards” (West, Allen, Rondo, Pruitt, Greene), sleek athletes who lack polish, shooting and consistency in their games (Reed, Green, Giddens, Walker), and raw, unchiseled big men (Jefferson, Davis, Erden). Leon Powe stands out as posing a different kind of risk. Coming out of California, Powe had already undergone multiple reconstructive surgeries to his knees. Only Ryan Gomes really stands out, as he represents the polar opposite approach to drafting.

The Beating Drum of Productivity

Recently, some members of the basketball statistical community have embarked upon a not-so-subtle attempt to critique the drafting strategies of NBA teams. Notably, Wages of Wins has applied its central statistic, Wins Produced, to college statistics, in order to discover and promulgate the most productive college players. (See an example here: http://dberri.wordpress.com/2008/05/27/the-2008-nba-draft-preview) The logic that productive college players will become productive pros has considerable merit, and when you consider where how well Ryan Gomes played at Villanova, how low he was drafted, and his production in the NBA, it appears that there may exist something of a market inefficiency.

And its not just the stats community, folks. No less a TV talking head than Dick Vitale stands as the vanguard for players like JJ Redick and Adam Morrison, whose college productivity were matched only by a equally mediocre set of measurables and workout performances. One of the few interesting elements of ESPN’s otherwise nauseating coverage of the NBA Draft has been picking up on Dicky V’s unstated premise that college success is a huge indicator of pro success. As he clamored for teams to pick All-Americans before “one-and-dones,” Vitale evinced a spirit not dissimilar to that of the stats community (!).

So What?

In the face of this critique, several facts about the NBA should be mentioned in rebuttal:

1) Having a mass of above-average, productive talent is the surest way a team will avoid a deep playoff run. In recent years, only the 2003-2004 Detroit Pistons can reasonably say they eschewed the “all-star(s) surrounded by role players” model on the way to a title. Maybe this is just evidence of a copycat league, but it also supports the notion that, in a league with a very finite amount of difference makers, there is a huge premium on finding the next one. Given that NBA rosters require only 13 spots to be filled, teams are not going to waste their time developing young players with marginal upside.

2) Shane Battier and Carlos Boozer stand out as clear examples of what Wages of Wins is talking about: high college productivity correlating with high professional productivity, despite being drafted below players of much lower productivity at any level. However, a quick look through their stats shows that even Battier and Boozer were not close to playing their most productive basketball yet. The NBA is dissimilar to the NFL or MLB, where rookies routinely step in as difference-makers and league leaders. Perhaps because of officiating, perhaps because of the difference in the length of season, rarely do rookies even merit All-Star consideration. In this way, rookies are less valuable in basketball then in other major American professional sports.

3) Most importantly, the rookie salary rules are essentially set up to induce teams to take projects in the first round, and productive types in the second round. Why is that? First round picks have guaranteed contracts for the first two years of their career, which can be extended at the discretion of the team for an additional two years. The salary of all four possible years is dictated by a set scale, which most would agree is below the market rate for players with the abilities of most first round picks. Second round picks, however, are essentially little more than restricted free agents; teams drafting players in the second round are under no obligation to sign the drafted player, and when they do, may negotiate terms not subject to any scale.

What does this mean? As I see it, NBA teams have a strong incentive to control project players for four years by drafting them in the first round, much less than they do with players who may enter the pros at a higher productivity level, but with a lower ceiling. Such players are available in free agency every year, and teams the advantage of having a larger sample size for these veterans than they do for rookies. Additionally, the flexibility of the system is such that a team can cut bait if the player does not progress after two or three years. If the player does progress, they will have the first shot at signing the player to a second contract (which typically runs through the player’s prime), and even a second shot (by allowed the player to hit restricted free agency, and merely matching another team’s offer).

The Value of Young, Cheap, Controlled Talent

If monitored correctly, such a system allows a rebuilding team to build a core through the draft, take stock of how the players together, and keep that team together by signing the best of those players to long-term contracts through their prime. This was Boston’s approach. However, before the 2007-2008 season, the approach was not achieving results, and lead many to wonder if Ainge had invested his reputation in young players who might never gel, or who would gel long after he had gotten the axe.

That’s where the secondary value of his draft strategy comes into play: if Team A has 1) young, talented players, who are 2) locked into short-term contracts at reasonable rates, that 3) Team B will have a right of first refusal in retaining upon completion of the rookie contract, then Team A has some very nice trade commodities. Essentially, Ainge cashed in these commodities to obtain the veteran talent necessary to re-build the Celtics as a title contender. West and the high draft pick that became Jeff Green were the main pieces in the Ray Allen trade, while Jefferson was the main cog in the Kevin Garnett trade. This veteran talent meshed with Paul Pierce and the remaining young talent (Rondo, Allen, Perkins, Powe, Davis) to create the Celtics’ title team.