Shopping

The Jam: Extras is a collection of Krucial Kuts that fall outside of our usual lines of classification.  You’re listening to side one/track two: Shopping, the Unimpeachable Guide to Danny Ainge and the NBA Draft (2003-2011).

Friend of the Kuts Nigel recently asked for our take on the Celtics’ draft evaluation acumen and strategy under the iron-fisted rule of former Blue Jays infielder Danny Ainge.  Along with his question, Nigel expressed two ideas.  The first: when it comes to the draft, Ainge has placed an out-sized emphasis on defensive specialists at the expense of capable scorers (Kendrick Perkins, Avery Bradley, and Tony Allen).  The second: Ainge has demonstrated an inability to properly evaluate offensive skill (Gerald Green).

To examine the issue, we first wanted to get a sense of how Ainge’s performance has compared to that of his peers.  We compiled a list of every player selected in the NBA draft since 2003, Ainge’s first year on the job.  Then, using Basketball-Reference’s handy Draft page, we pulled up the Win Shares generated by each player selected.  The Win Share is an all-in-one statistical shorthand that estimates “a player’s contribution [to his team] in terms of wins”.  Essentially, it measures the quality of a player’s performance in a way that normalizes for all different types of players.  Check out www.basketball-reference.com for a detailed description of what Win Shares are and how they’re calculated.

Our initial analysis was a simple one: tally up all of the Win Shares, divide by the total number of picks, and whichever team showed the most Win Shares per Pick drafted best.  Before we get to the results, we must first point out a few things:

  1. When compiling the list of drafted players, we defined a team’s “picks” as players who went on to play for the team that chose them before playing for any other NBA team (obviously), and players who were drafted and then traded to the team in question before the season started.  For example, in 2008 the Minnesota Timberwolves drafted O.J. Mayo with the third pick and the Memphis Grizzlies drafted Kevin Love with the fifth.  Shortly thereafter, the two players were traded for each other.  For our purposes, Mayo was picked by the Grizzlies and Love was picked by the Timberwolves.
  2.  As we are trying to get a sense of how Ainge and his contemporaries have performed as talent evaluators, we are not separating Win Shares that were generated for another team out of the tally.  The Celtics drafted Al Jefferson and later traded him to the Timberwolves for Kevin Garnett.  The Win Shares that he generated for the T’Wolves and, later, the Utah Jazzcount towards the Celtics’ total.Apportioning the Win Shares to the teams that benefited from them would not only be painfully time consuming, it would also transform the exercise into an examination of team building.  This would become increasingly convoluted as we factored in Win Shares generated by veteran players received in trade or looked at the effects of player movement on payroll flexibility.  We’re only interested in how teams compare to each other in terms of identifying players who can perform well enough to maintain a place for themselves in the league.
  3. During Danny’s tenure as the Celtics’ President of Basketball Operations, several teams have experienced turnover in their front offices, going through one or two or more changes in the General Manager’s seat.  We’re interested in Danny’s performance against the other teams in the league, not the other executives, so we make no distinction between, say, the picks made for the New York Knicks by Scott Layden, Isiah Thomas and Donnie Walsh.

The following table shows the total number of picks and pick-generated Win Shares for each team since the 2003 draft, sorted by Win Shares per Pick.

Ainge has done well in the draft, finishing eighth with 9.3 WS/Pick, well enough above the league average of 7.4.  Other notables:

  • The Chicago Bulls (14.8 WS/Pick) have been the elite, landing high-quality players such as Luol Deng (46.9 WS), Kirk Hinrich (42.3) Ben Gordon (32.6), Derrick Rose (28.2) and Joakim Noah (27.3).  They’ve benefited from six top-ten picks in the last nine years, tied for the second most in that range in that span.
  • Of course, high picks don’t mean much if your management doesn’t know how to use them.  The Minnesota Timberwolves (3.1 WS/Pick) have had a league-leading seven top-tens since 2003, yet finish dead-last in this measurement.
  • The Cleveland Cavaliers rank high on this list thanks entirely to the selection of LeBron James, whose 126.4 Win Shares are by far the most generated by any player drafted in the past nine years.  This number makes up 60% of the Win Shares that the Cavs have generated through the draft.  Just for JKs, let’s say that Cleveland drafted Carmelo Anthony (third selection overall, 58.7 WS) instead of LeBron. Their new average of 8.4 WS/Pick would drop them from fourth in the league to eleventh.  For heartier laughs, let’s imagine that they drafted Darko Milicic (second selection, 7.3 WS).  At 5.3 WS/Pick, they would rank 22nd.
  • Some may be surprised to see the San Antonio Spurs, who have a reputation as one of the very best organizations in the league, second to last on this list, amongst perennial cellar-dwellers the Clippers, Nets and Timberwolves.  This is because nine of the Spurs’ 19 picks have never played a minute of NBA basketball.  Their no-show percentage of 47.4% leads the league by a fairly wide margin, ahead of second-place Dallas and Denver at 41.7% each.

As we pointed out, the Bulls have had six top-ten picks since 2003.  The Celtics are one of only four teams[1]  to have had zero such picks in that time[2], and they are one of seven teams[3] to have had two or fewer top-fifteens.  Despite drafting from a point of disadvantage year after year, Ainge has been able to consistently identify and select players who have, at the very least, established themselves as rotation regulars.

The following graph describes how the Celtics have performed in the draft against the league average and against the other nine teams in the top ten in WS/Pick, in five-spot increments.

We see that the Celtics, represented by the green bar, have generated zero WS/Pick from draft spots one through five because they haven’t made any selections in those spots.  On average, the other 29 teams in the league have used turned those spots into 21.0 WS/Pick, and the other nine teams on our top ten list have turned them into 40.5 WS/Pick.

Both the league and the top ten operate along the same trend line: they make their biggest scores within the first five picks, drop off by half or more at the six-to-ten range, and then experience gradually diminishing returns as the draft goes on.

The Celtics, on the other hand, have thrived in the middle and back end of the first round and the middle of the second round.  The following table lists the players that the Celtics have taken in these spots.  The table is sorted by Win Shares generated, and includes Win Shares per 48 minutes.

The Celtics had their best success in the draft from 2003-2007, a stretch that saw the team go from average to terrible in the standings.  With the onset of the Garnett era, the Celtics won a title, lost out on a second one in seven games, and made the playoffs four seasons and counting.  They also stopped getting draft picks in remotely promising spots.  From 2003-2007, the Celtics had seven first-round picks and seven second-round picks.  On average, these were the 20th and 45th overall.  Since then, they’ve had three first-rounders and five second-rounders, which averaged out as the 25th and 54th picks overall.

This table shows the players chosen in spots where the Celtics have not fared as well.

It seems fitting that Gerald Green is at the top of this list[4], a player whose selection was based mostly on the promise of his uncanny leaping ability, which translated into some truly spectacular displays of in-game dunking.  Unfortunately, it didn’t translate into much else.  At the end of the 2007 season, Green was dealt to the Timberwolves as part of the package that brought back Kevin Garnett, and was out of the league two years later.

The selection of Green is often used as a symbol of Ainge’s failings as an executive.  Ainge had hoped to select Danny Granger (42.8 WS and counting), but he was chosen by Larry Bird and the Indiana Pacers with the pick immediately preceding the Celtics’.  The following players were still on the board between the selections of Green and Ryan Gomes, who the C’s would take with their next pick (50th overall): David Lee (44.7 WS), Louis Williams (22.7), Monta Ellis (22.3), Jarrett Jack (22.0), Hakim Warrick (20.2), Brandon Bass (18.7), Nate Robinson (18.1), and 13 others who have had more impactful careers than Gerald’s.  Of course, when this is brought up, no one mentions that Lee, Williams and Ellis were chosen 30th, 45th and 40th.  No teams foresaw that they’d turn out to be as good as they have.

Ainge dreamed big on Green’s athleticism and came up with the short end.  It happens.  We don’t have graphs or tables to back this up, but we’re pretty sure that every GM has a Gerald Green, Stromile Swift, Hasheem Thabeet, or Jonathan Bender that they’d like to forget about.  A player taken high in the draft based on the promise of his raw, physical gifts and little else.  A player that didn’t pan out.

The draft is a gamble, with the odds stacking higher and higher against its participants as each selection is made.  Our final word on this: between 2003 and 2005, 15 players, including Gerald Green, were selected in the 16-20 range.  Eight of those players have been worth 1.0 or more Win Shares per year[5] since their selection.  Seven were worth less.  The space that separates Joey Graham (2005, 16th pick, 1.1 WS/Year) from Danny Granger (17th, 7.1) from Gerald Green (18th, 0.2) from Hakim Warrick (19th, 2.9) from Julius Hodge (20th, 0.0) is the same as with every draft choice, year after year: one pick.

To conclude, we return to Nigel’s initial query: does Danny Ainge overvalue defense when assessing draft prospects?  To answer the question, we looked at the Offensive and Defensive Win Shares for each of Ainge’s draft picks (once again, see http://www.basketball-reference.com for definitions).  We set a minimum games played limit at 50, which is still probably too small a sample size, but allows us to include Avery Bradley.  Next, we simply subtracted the DWS from the OWS.  A negative number tells us that the player has provided more value on the defensive end; a positive number tells us that he provided more value on the offensive end.

This is an admittedly crude analysis, and certainly has its flaws: it suggests that Gerald Green contributed more with his defense (probably not), that Rajon Rondo does significantly more with his defense (maybe), and that Ryan Gomes is an offense-first player (doubtful).  Flaws or no, with Kendrick Perkins rating as the most defense-heavy and Al Jefferson rating as the most offense-heavy, we think it works well enough to tell us where Danny’s mind is at when he’s making his draft board.

The following table shows that 14 of Ainge’s picks have been “defensive” players and four have been “offensive,” though there are five to seven who may be too close to a zero difference to call:

It does appear that Ainge prefers to draft with defense at the front of his mind, and acquire his offensive weapons through trade and free agency.  To say whether he over-, under- or properly values defense in comparison to offense would mean that we’d have to do this for all of the other teams’ picks, which would take a very, very long time.

When putting this together, we turned up all sorts of list fodder and fun facts about the last nine years’ drafts.  Check back with us later this week for all of the dirt, plus a very special interview with a guy who went to grade school with Stephen Curry and to high school with Jeremy Lin.  That’s a scoop you won’t get anywhere else.


[1] Spurs, Rockets and Suns.

[2] In 2006 they had the seventh pick (Randy Foye) and in 2007 they had the fifth (Jeff Green), but traded both away.

[3] Spurs, Rockets, Mavericks, Lakers, Nuggets and Heat.

[4] What does it say about us that the player whose season we’ve decided to document in mind-numbing detail and our favorite of the lost children from the post-Antoine/pre-Garnett dark days are the only two who have “contributed” negative Win Shares?

[5] The formula here is WS/(2012-”Year Drafted”).  This counts the years that a player like Gerald Green did not play in the league.  Part of success in the draft comes from choosing players who last for several years, so it makes sense to us ding selections by factoring in the seasons that a player doesn’t play.

4 thoughts on “Shopping

  1. This is a very well-done and very well-thought-out argument, and my kudos to the proprietors of this blog for it.

    I would also like to add that one could make the case that Danny Ainge’s predisposition for defensive players over this time could represent both team philosophy and team need.

    Clearly they had enough offense with the Big 3, so it would make sense to be focusing on defensive role players during that era. And prior to that era, one could make the case that players like Gary Payton, Antoine Walker, Al Jefferson, and the like were more than enough scoring.

    The emphasis that Doc Rivers places on playing both ways may have contributed to Ainge’s decision to focus on defense.

  2. Pingback: The Unimpeachable Guide to Danny Ainge and the NBA Draft (2003-2011) « Krucial Kuts

  3. Pingback: The Unimpeachable Guide to Danny Ainge and the NBA Draft (2003-2011) - Hardwood Houdini - A Boston Celtics Fan Site - News, Blogs, Opinion and More

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s