Brian Chan is a Southern California transplant and life-long Clippers fan living in the Bay Area. The things keeping him content are his camera, the game of basketball and the beauty of Chris Kaman. Today, he digs inside the data for ClipperBlog to establish a benchmark for the 2010-11 season. It’s a tall task because the Clippers have overhauled their roster from a season ago. With all those moving parts to account for, here are Chan’s conclusions:
Clipper Win Projections Using WARP
With the start of the season upon us, there are many questions in the minds of Clippers fans. Will Blake Griffin live up to the hype? Will Eric Gordon develop into an elite 2-guard? Can the Clippers stay healthy? All of these questions boil down to one overarching question: how many games will the Clippers win this season?
There have been many projections for the NBA Season, most of which seem to paint a fairly grim picture for the Clippers. I thought I would try my hand at a simple projection that would give reasonable results and could show the drivers of a team’s wins or the lack thereof. For this projection, I used Kevin Pelton’s insightful Wins Above Replacement Player (WARP) System, to try to see if a team’s cumulative WARP is representative of its eventual total wins.
WARP attempts to evaluate a player based on how many wins the player would produce on a team with four average players relative to the wins produced by a team consisting of four average players and one replacement level player, where replacement level is defined as the level of production of a readily available replacement (not starters). The WARP evaluation is based on various statistics that rate a player’s offensive and defensive production and efficiency, and it allows the rating of players on a per minute basis. The WARP Formula breaks down as:
WARP = (Win% – RL) * (Min/48)
WARP: Wins Above Replacement Player
Win% = Winning Percentage of Team (Player X + 4 Average Players)
RL = Winning Percentage of Replacement Level Team (Estimated: 10 wins/82 games)
Min. = Minutes
There are many detailed formulas that go into WARP; however, by using WARP and Minute numbers from Basketball Prospectus, we can identify a “Win%”, which can in turn be used to recalculate WARP based on projected minutes in 2011.
|Player||Adjusted WARP||Player||Adjusted WARP|
|Al Thornton||-0.6||Kareem Rush||-0.1|
|Baron Davis||8.2||Marcus Camby||8.8|
|Bobby Brown||-0.3||Mardy Collins||-1.2|
|Brian Skinner||-0.3||Rasual Butler||-3.5|
|Chris Kaman||2.6||Ricky Davis||-1.0|
|Craig Smith||2.5||Sebastian Telfair||-0.2|
|DeAndre Jordan||1.3||Steve Blake||-0.2|
|Drew Gooden||4.5||Steve Novak||-1.1|
|Eric Gordon||0.6||Travis Outlaw||0.0|
This win projection assumes that teams are about as good as the sum of their parts. After finding the 2010 WARP for each player who played on a team and adjusting for minutes for players who were moved mid-season, we have a roster broken down by WARP. By adding the WARP for the roster, you have a cumulative WARP that represents contributions by all players during the season. Comparing cumulative WARP to team wins produces interesting results:
|Team||2010 Cumulative WARP||2010 Wins||Correl|
|Oklahoma City Thunder||38.9||50|
|Portland Trail Blazers||36.2||50|
|San Antonio Spurs||36||50|
|Golden State Warriors||28.3||26|
|New Orleans Hornets||24.4||37|
As seen in the table, there is a strong correlation (0.92) between cumulative WARP and season wins. There are a few inconsistencies, such as the Lakers with a 37.2 WARP, but in general, the trend appears to be that a higher team cumulative WARP results in a higher number of team wins.
Applying a simple regression will allow us to approximately predict wins in the 2011 season based on projected 2011 WARP.
Using the previous season’s WARP Win % and basic estimations of players’ minutes based on their spot on the depth chart, our projected 2011 WARP looks like the following:
|Clippers 2010-2011 Roster|
|Player||2010 Win%||Projected Min||Projected WARP|
This method allows for simple projected WARP for players with historical figures; however, it doesn’t account for the rookies who will be seeing playing time. Instead of attempting to estimate their WARPs, one can look at the possible scenarios.
|No.||Condition||Expected WARP||Expected Wins|
|4||BG+ (12.7) EJ+ (6)||29.18029649||40.07258555|
|5||BG+ (12.7) EJ+ (6) Rookies+ (3)||32.18029649||43.6846448|
Scenario 1: In this situation, there is no player progression and the rookies all play at a replacement player level. The expected WARP would be 11.12 and the expected win total becomes 18.33. This drop in wins reflects the loss of Marcus Camby, the Clipper with the highest WARP in 2010 despite not being on the team for the entire season.
Scenario 2: In Scenario 2, we assume Blake Griffin at least replaces Marcus Camby’s contributions and has a WARP of 8.8. This results in 28.93 expected wins.
Scenario 3: In Scenario 3, Blake Griffin plays as well as Camby did over the entire 2010 season and has a WARP of 12.7. This results in 33.62 expected wins.
Scenario 4: Blake Griffin plays as well as Marcus Camby and Eric Gordon increases his WARP to 6, which is about average for a good 2-guard. This results in 40.07 expected wins.
Scenario 5: Finally, on top of Scenario 4, the rest of the rookies play at a high level and produce a WARP of 3. This results in 43.68 expected wins.
There are many potential scenarios, but these were ones that seemed reasonable enough to ponder on.
Although this projection may not be bulletproof, it’s an interesting way to look at the season’s potential. This projection’s proximity to other projections is a good sign, and it allows us to see which positions the Clippers can improve at in terms of productivity and efficiency. The addition of Ryan Gomes presents an improvement at small forward, as he will cannibalize minutes from Rasual Butler, who had one of the worse WARPs (-3.5) in the Western Conference. Randy Foye provides improvement and stability at the backup guard positions in terms of WARP, and the great potential of the Clippers’ rookies will be fun to follow.
Ultimately, this and other statistical projections should not lessen your excitement for the season. There are factors, such as player development, change in style of play, and chemistry, that are too difficult to capture into a neat number. With the new additions to the roster and the coaching changes made, I know I, as a fan, can’t wait for tipoff.