Round 1: Memphis Grizzlies at San Antonio Spurs - Manu: Even more important than you think
After about a month of bickering whether or not games were even worth playing, the excitement of “money time” has finally arrived. As anxious as we all were to get to the playoffs, Manu Ginobili’s injury (if it’s serious) should make the Spurs anxious about getting out of the first round.
Manu’s importance
When I attended the Sloan Sports Conference in early March, I was involved in a discussion with the legendary basketball stat-guru Dean Oliver. When the conversation shifted to Manu Ginobili, Oliver stated that he felt Manu was the most underrated basketball player in NBA history. (Oliver later clarified that he hasn’t researched the subject, so don’t hold him to it completely.
Although Oliver hasn’t researched the subject, his assertion was not unfounded. Based on Basketball-Reference.com’s Win Share’s per 48 minutes, Ginobili has ranked 5th in the NBA in 2004/05, 2nd in 2006/07, 6th in 2007/08 and 6th in 2009/10. Had Ginobili played enough minutes in 2005/06 and 2008/09, he almost certainly would have finished in the top 10 both of those seasons as well. (Manu ranked 12th this season.)
My version of statistical plus minus looks even more favorably on Ginobili (click here for explanations of my offensive and defensive ratings). At this point, ratings are calculated from 2005/06 to 2009/10. Manu’s per possession rankings: 1st, 1st, 4th, 4th and 2nd (among players with at least 500 minutes played). My version of statistical plus minus favors Manu more than other statistical ratings largely because of his ability to draw charges, his high steal rate, his ability to create unassisted 3 pointers, unassisted layups, and his ability to set up his teammates for layups and dunks. Charges are not readily available this season, so I still need to figure out how I will determine my ratings for this season.
I must disclose that although Ginobili ranked first in the NBA in 2005/06 and 2006/07, it was not by a wide enough margin for me to suggest he was the league’s best player; but he was much closer than most suspect.
Manu’s value isn’t based just on some meaningless stats, either. The Spurs regularly perform substantially better when he is on the court. This season, the Spurs have outscored opponents by 11.41 points per 100 possession when Ginobili is in the game. Only Kevin Garnett has experienced greater success. Incredibly, the Spurs have been outscored by 2.14 points per 100 possessions when Manu is out. Yet somehow, this season was only the second time that Ginobili played in the all-star game.
Suffice to say, if Manu Ginbili can’t play at close to 100% during this series, the Spurs could be in serious trouble.
The Spurs projected playoff rotation
The following represents a unscientific estimate of the minutes per game I expect for the Spurs players during the playoffs.
Player EstMPG Adj +/- 4 Yr Ridge APM
Manu Ginobili 35 12 6.2
Tim Duncan 35 6.1 5.3
Tony Parker 35 -3.1 2.6
Richard Jefferson 28 0.6 1.4
George Hill 22 3.6 0.9
Antonio McDyess 19 -2.3 2
Matt Bonner 19 1.5 4
Gary Neal 18 -3 -0.7
DeJuan Blair 14 -6.4 -0.7
Tiago Splitter 10 -4 -1.1
Others 7 0.9 0
In addition to the expected MPG, the above table also includes adjusted plus minus for the current season. Since Memphis’ statistics are not available on basketballvalue.com, I calculated the results on my own. I made a few adjustments such as including separate estimates for Battier in Memphis and Houston and grouping a few players at the end of the rosters for both teams. Adjusted plus minus used ordinary least squares regression to account for all players on the court and estimate player value.
The numbers represented by “4 Yr Ridge APM” are taken from stats-for-the-nba.appspot.com. Ridge regression addresses a few problems associated with the adjusted plus minus using ordinary least squares regression, such as those from basketballvalue.com. (Regression is the technique of fitting data, scoring margin in this case, based on predictors, such as players. Ordinary least squares is the simplest form of regression, with no transformations or adjustments.)
Ridge APM constrains player estimates in order to reduce the tendency of outlandish estimates. Additionally, ordinary least squares often has difficulty allocating success among the 5 players in common lineups (more on that later). This becomes less of a problem as more years included in the regression. These benefits do come at a cost. The Ridge-based player estimates are biased towards zero, especially for players with few minutes. In certain instances, it is also important to consider the bias towards prior seasons. Despite the costs of making these adjustments, adjusted plus minus using ridge regression is a far more accurate predictor than ordinary least squares.
A couple notes:
Obviously, the figures above assume that Ginobili doesn’t any time. I find it hard to believe he will miss game 1 (based on nothing), but I have no idea if his performance will be limited.
The Spurs depth was a strong suit during the regular season, but as stars play more minutes, this advantage will diminish.
Opponents of Memphis attempted three pointers at the 4th highest rate in the NBA. I suspect it will be more valuable to play Matt Bonner, etc. more minutes in this series. Parker, Ginobili, Hill and Duncan’s presence and ability to get to create at the rim should also help create room for these three point shooters.
One interesting matchup could be Manu Ginobili against Tony Allen. For all of Ginobili’s value, he takes a ton of risks. He throws ambitious passes, makes awkward moves, takes hotly contested shots and attacks the rim recklessly. Of course, these plays are usually calculated risks (at least instinctively) and typically work in Manu and the Spurs’ favor, but Tony Allen might be the most opportunistic defender in the NBA. His steal rate of 4.5% tops the league. Chris Paul is next at 3.6% and Rondo is the only other player above 2.8%. I wouldn’t be surprised if Manu struggles against Allen initially, even if healthy, and adjusts throughout the series.
The Grizzlies projected playoff rotation
Player EstMPG Adj +/- 4 Yr Ridge APM
Zach Randolph 37 4 2.4
Mike Conley 37 7.5 3.1
O.J. Mayo 33 -2.2 -0.5
Marc Gasol 31 -3.2 2.2
Tony Allen 30 -0.3 2
Sam Young 24 -1.9 -1.1
Shane Battier 24 1.5 3.6
Darrell Arthur 18 -0.1 -1
Greivis Vasquez 4 -7.2 -1.4
Rudy Gay 0 -1.1 -0.8
Others 4 0.4 0
The regular season series between the Spurs and Grizzlies was knotted at 2-2. However the Spurs were without Parker for one loss and Duncan for the other.
For more analysis of the Grizzlies, check out my advanced scouting reports prior for the first, second and fourth matchups against the Spurs.
Projecting the series
John Hollinger and Accuscore both set the Spurs odds of advancing to the second round at 82%. Basketball-reference provides an estimate of 76%. Using performance after the trade deadline, basketball-reference adjusts these odds to 48%. The post-deadline figures may be more appropriate for evaluating the Grizzlies, who have continued to play well despite losing Rudy Gay for the season. However, as Neil Paine points out, the Spurs rested players for much of this stretch.
For my projection(s), I decided to apply adjusted plus minus estimates. Although the Ridge based estimates are more accurate from player to player, I suspect that the single season ordinary least squares estimates are more appropriate in this situation. The problem of allocating player value between the 5 players on the court is effectively eliminated when evaluating the lineup as a whole. The wild player estimates are not a significant factor on the team level since the minutes and observations are much greater than the player level. Additionally, the bias’ toward zero and past performance isn’t included in this estimate.
Based on my expected MPG for the Grizzlies and Spurs my team ratings are as follows:
Spurs 8.8
Grizzlies 4.2
The Spurs SRS rating is 5.86. This increase is largely attributable to the increase in playing time expected from the Spurs top players. The Grizzlies SRS rating is 2.55. Memphis appears to be a better playoff team since Tony Allen and Shane Battier missed significant time during the regular season.
Based on my team ratings and a 3.5 point home team advantage, I estimate the Spurs have an 83% chance of winning this series. Of course, like most of the other projections, this does not factor in the possibility of Manu Ginobili being severely limited. Although I have no reason to suspect that Ginobili will miss the whole series, I have developed estimates for this scenario.
My first projection for this series sans Manu assumes that his value is 12.0 (based on Ordinary Least Squares APM). As good as Ginobili is, this is an aggressive estimate in my opinion. Since Ginobili’s absence significantly alters expected lineup combinations, it seems reasonable to adjust his estimated impact as appropriate. The second projection without Ginobili sets his value at 6.2 (which might be conservative). Finally, the last column assumes that Manu will play at a diminished rate roughly in line with my personal expectation.
The following table estimates the odds of each scenario based on 10,000 simulations.
Winner Games With Manu No Manu (value=12.0) No Manu (value=6.2) My Pick
Spurs Series 83% 24% 54% 74%
Spurs 4 17% 1% 6% 10%
Spurs 5 30% 5% 15% 24%
Spurs 6 18% 5% 12% 18%
Spurs 7 18% 12% 20% 22%
Grizzlies 7 6% 14% 12% 7%
Grizzlies 6 7% 27% 19% 11%
Grizzlies 5 2% 20% 9% 4%
Grizzlies 4 1% 16% 5% 3%
Related posts:
- Game 1: Memphis Grizzlies 101, San Antonio Spurs 98
- Familiarity breeds contempt, so get to know the Memphis Grizzlies
- Pregaming: Spurs vs. Grizzlies Game 1
- The 4-Down Podcast, Episode 24: Looking at the Spurs-Grizzlies series with Matt Moore
- Off the Beaten Path, Spurs need Manu Ginobili
- On Manu’s injury and playoff seeding
- Advanced Scouting: San Antonio Spurs at Los Angeles Lakers - Who’s playing tonight?
