Monta Ellis is obviously known for his dynamic scoring ability, but it's his ability to create scoring opportunities for others that stands out as significant for the Golden State Warriors this season. Photo by Ezra Shaw/Getty Images.
For those not familiar with Pro Basketball Prospectus - or the work of co-authors Bradford Doolittle and Kevin Pelton at BasketballProspectus.com - it is like an encyclopedia of NBA statistics and team analysis for the season that's worth a read for any die-hard basketball fan.
Due to the NBA lockout, Pro Basketball Prospectus 2011-12 was released later than their past two editions and has come with a .PDF supplement as well as a few supplementary articles on BasketballProspectus.com to keep you as updated as possible during the whirlwind start to this shortened season.
So perhaps you can consider this Q&A with Pelton as the "Golden State Warriors supplement", which includes thoughts on what Nate Robinson might contribute to the team, what might help rookie Klay Thompson develop into a productive NBA player, the team's most surprising statistic through 10 games, and rumors of little-known cloning experiments in Los Angeles during the 2005-06 season.
GSoM: In the book, you predicted that the Warriors would win 28 games this season. So far they're off to a 3-7 start. I saw you note on Twitter a while back that "Golden State got a contending schedule...with a non-contending team". You followed that with a late-December article with schedule-adjusted projections for the season that have the Warriors finishing dead last in the West. Can you say more about what makes the Warriors' schedule so difficult?
Kevin Pelton: The way I evaluated the difficulty of schedule was to take our generic SCHOENE projections and run them through the actual schedule, simulated 10,000 times by a program written by my colleague Bradford Doolittle, and see how it affected win projections. That exercise really confirmed the notion that this year's schedule pitted the top teams against each other as much as possible in terms of the interconference matchups for TV purposes. So most of the teams that dropped were the best teams in the league, and the biggest draws. Then there was Golden State. Just two projected lottery teams saw their win total go down using the actual schedule, and the other was Memphis. Add in the ESPN opener against the Clippers and it becomes apparent that the league viewed the Warriors as a marquee team for scheduling purposes.
GSoM: For those who are unfamiliar with SCHOENE, can you summarize in plain terms what exactly it is and why it gave me false hope by predicting the Warriors to win 40+ games a couple of years ago?
KP: SCHOENE is our projection system, named for former Sonic Russ Schoene. There are two components--individual projections and team projections. For players, we take their performance over the last three seasons and use it to generate a lost of the most similar predecessors at the same age. We then use those players' actual development to adjust that baseline and come up with a projection for each player. The team projections take these stats as well as some important team factors and our best guess at playing time to come up with a complete projected stat line for each team. Those stats are translated into a win total.
Now, as for why the Warriors in particular have been overpredicted, I think there are a couple of reasons. One factor that wasn't a part of SCHOENE until this year is passing, so there was no way for the system to see that all of Golden State's scorers wouldn't work well together. Last year, the addition of David Lee really threw things off. Lee's good rebounding meant we projected that the Warriors would go from the bottom of the league in defensive rebound to above average. But Lee has never had that kind of impact on team rebounding because he doesn't do a good job of boxing out.
GSoM: So given some of the over-/under-estimating that SCHOENE has done in the past, is there any reason to believe that the Warriors might exceed expectations this year?
KP: Since the system was overhauled this summer, I don't think this year's projection really overstated things for the Warriors. (If it did, I feel bad for fans.) Stephen Curry's ankle really throws off everything because his performance is such a big part of the games Golden State was projected to win, but if he stays healthy it's not out of the question that the Warriors are a part of the playoff race.
GSoM: I found it funny that your player projections had Andris Biedrins listed as similar to Kwame Brown, which clearly means that the Warriors got an advance copy of your book before making this signing. It's only a one-year contract - and now Brown is expected to be out for most of the season - but is there precedent for a NBA team becoming successful with two Kwame Browns...or two Andris Biedrinses... or one of each... getting heavy minutes in their center rotation?
KP: Aside from some little-known cloning experiments conducted by the 2005-06 Los Angeles Lakers, no. The Warriors get enough offense from the other four positions that they might have been OK if Biedrins and Brown could just provide defensive rebounding and the occasional made free throw--you know, like the 2006-2008 vintage Biedrins. Andris' health aside, I think the early returns have been decently positive.
GSoM: Of course, the Warriors made some other additions to their roster since your book was published and all of them - Brown, Brandon Rush, Nate Robinson and Ish Smith (as well as Klay Thompson) - have had ups and downs this season. Which player might you expect to increase their production as a member of the Warriors compared to what they've done in the past?
KP: Of those guys, I would say Rush has the best opportunity. The Warriors' free-flowing offense can create more open looks for Rush, especially when he plays as part of smallball lineups against slower defenders. Rush has shown the ability to make spot-up jumpers in this league, so he can be effective if he plays to his strengths.
GSoM: Robinson might be a guy who people hope will play better than he did last year, at the very least. What can he bring the Warriors?
KP: I think Nate has gotten a bad rap in the league because of his emotions and his celebrations. At his core, he isn't a selfish player, but he's obviously best cast as a volume scorer. When he's getting into the paint and knocking down threes, his ability to use a lot of possessions makes him valuable. He hasn't done that as well the last couple of seasons, but was very effective as a bench scorer in New York in 2008-09. With Steph Curry out, and with Dorell Wright struggling, Golden State needs additional firepower from the backcourt, so I think this can be a good marriage of player and team.
GSoM: The Warriors also have a number of younger players with somewhat unknown potential that you have already covered in your book (e.g. Thompson, Charles Jenkins, Jeremy Tyler, and Ekpe Udoh). Hoopdata had some interesting tweets recently about the need to redefine how we think about the concept of "talent" and developing players that supposedly possess it. Going back to one of the subjects you identified as standing out at the 2012 MIT Sloan Conference, how might statistics help a relatively young team like the Warriors that has a number of players with question marks about their future potential? What might they not account for? (Feel free to use an example of a player from the Warriors roster developing into an All-Star.)
KP: I think good statistics can be a good reality check in terms of how athletic potential translates into success on the floor. A player like Nick Young put up big scoring numbers last season, and even shot a decent percentage from the field, but his lack of other contributions indicated he still wasn't helping the Wizards. A coaching staff that can use statistics with players when appropriate can reinforce winning habits.
I think another important lesson from Washington (the team that inspired Hoopdata's rant) is the importance of putting players in situations where they can succeed. To some extent, you need to let them experiment to figure out just what they can do on the court, certainly. But when you're able to put them in the appropriate box, it builds confidence. Let's take Klay Thompson: eventually, I think he may be able to create off the dribble. For now, though, spot-up shooting is the clear strength of his game. The more Mark Jackson can do to take pressure off Thompson from having to do too much and allow him to just shoot, the better I think Thompson's rookie season - and his chances of eventually developing into a quality player (let's not use the A-S word just yet) - will be.
GSoM: So through 10 games, what Warriors stat(s) jump out at you as particularly surprising, given a challenging start that has included significant wins against the Miami Heat, Chicago Bulls, and New York Knicks?
KP: I would say Monta's assist rate. Obviously he's had the ball in his hands a ton, but last year when he played point guard he averaged 7.9 assists per 48 minutes, and he's substantially ahead of that so far this season overall. [Note: as of today, Ellis is averaging 9.4 assists per 48 minutes (19th in the league), according to NBA.com.]
GSoM: Well then...you have forced me to ask: if you had to choose between Steph Curry or Monta Ellis to build around, who would it be based on the numbers?
KP: Absent the ankle, it would definitely be Curry. I think you can build a winning team around Ellis--like the We Believe team--but it would require just the right mix of talent around him. To build a competitive defense with Ellis, you'd have to put him alongside a big guard and to make that work on offense that player would have to both be able to distribute and shoot the three. Those kind of players are rare and valuable. Curry is much easier to translate to any situation.