Assuming the Celtics manage to win one of the two games remaining in Boston (a pretty good assumption), someone from the Celtics will be named MVP of the Finals. After Game Four I argued that someone should be Ray Allen. But when we look at the performances of the players in Game Five, the R. Allen candidacy appears to be faltering.
Table One: Analyzing Game Five of the 2008 NBA Finals
Table One reveals that the top player in Game Five for the Celtics was Paul Pierce. This was also true in Game One and Game Two. In the third and fourth games it was R. Allen, and R. Allen still leads the Celtics in PAWS. Still, the Pierce candidacy is beginning to look to be better supported by the numbers (although R. Allen still leads in PAWS for the series).
The MVP in LA
Now what if the Lakers pull off a sweep in Boston? Certainly the choice for MVP would depend on how the players played in Game Six and Seven. But given what we have seen so far, the top Laker – after a very strong Game Five – is Pau Gasol. Gasol currently lead the Lakers in PAWS in the Finals. Not only that, it should be clear to everyone that the Lakers would not be in the Finals without Gasol. As I noted last March, without Gasol or Andrew Bynum, the Lakers were about a 0.500 team this year. And as I also noted last March, that’s basically what the Lakers have been since Shaq left town. Yes, Kobe is a very good player. But without a productive big man, the Lakers are not a top NBA team.
Of course, one can also argue that without Kobe the Lakers don’t win much either. Certainly I have no problem with that argument. But given that the Lakers need both Gasol and Kobe to be successful, the choice of team MVP – assuming we don’t do co-MVPs – depends on who had the “biggest” impact on team success. When I look at the Finals numbers – via PAWS – the top player on the Lakers so far is Gasol.
Other Game Five Thoughts
If the Lakers do come back and win the championship, I think the Boston collapse (and it would be a collapse) has to be at least partially attributed to the injuries to Kendrick Perkins and Rajon Rondo. Although much attention is focused on the Big Three, the other two members of the starting line-up in Boston were quite productive this year. Replacing these two players with P.J. Brown and Sam Cassell has not really worked. Although Brown and Cassell were very productive players in the past, in the 2008 playoffs each player has been well below average.
In addition to playing unproductive players, Doc Rivers also appears to have transferred frontcourt minutes to James Posey. In Game Five, Garnett, Brown, and Leon Powe played 63 minutes. This means the remaining 33 minutes at power forward had to be played by Posey. Yes, Posey can also be an above average player. But I don’t think the data suggests that Posey is better at power forward than Powe.
Let me close with one final (and somewhat silly) thought on the Lakers. You might have missed this, but there was a Chris Mihm sighting at the game. In three minute he missed a shot, committed a turnover, and committed two personal fouls. This works out to a PAWS48 of -58.99.
Is this the worse PAWS48 in history? It wasn’t even the worst in the game. Ronny Turiaf played one minute and committed a turnover and a personal foul. That means his PAWS 48 was -82.99.
Now many people think that if they played they could do better. I disagree. For a player to commit a turnover he first has to have the ball. And I am pretty sure that if most people took the floor with a professional NBA team (especially in the Finals), no one would think about throwing the non-NBA person the ball.
Of course, without the ball you could still commit personal fouls. Trevor Ariza played one minute and committed a foul. Since he didn’t do anything else, his PAWS48 was -31.44.
So what does all this tell us? A non-NBA player could post a better PAWS48 than either Mihm or Turiaf. And if the person didn’t touch anyone, he or she could probably top Ariza as well.
– DJ
The WoW Journal Comments Policy
Our research on the NBA was summarized HERE.
The Technical Notes at wagesofwins.com provides substantially more information on the published research behind Wins Produced and Win Score
Wins Produced, Win Score, and PAWSmin are also discussed in the following posts:
Simple Models of Player Performance
What Wins Produced Says and What It Does Not Say
Introducing PAWSmin — and a Defense of Box Score Statistics
Finally, A Guide to Evaluating Models contains useful hints on how to interpret and evaluate statistical models.
porteno
June 16, 2008
Clearly this invalidates the whole model.
Logic
June 16, 2008
Wow.
I feel like your data has led you far off the path of reality.
How far off must you go before you realize that your interpretation of the data is wrong?
Kevin Garnett just as good as Tim Duncan? He may be as “productive” by your measures, but that seems to have no correlation to being a better player.
Pau Gasol the MVP of the Lakers? I can’t tell if you are serious or not. Are you so arrogant to believe that your data trumps the analysis of former players, coaches, and scouts? I know they are not always in agreement, but this borders on ridiculous.
It flies in the face of sense (and I don’t mean common sense, I mean basketball sense).
Now, I can’t back my claims up with “hard evidence”, but then again I don’t believe basketball statisitics constitute “hard evidence”. This is not baseball. Basketball is more like soccer, and so any credible judgment must be given with a heavy dose of experience and qualitative analysis.
ilikeflowers
June 16, 2008
Thanks for the anecdotes, Logic. They are very original and thought provoking.
mrparker
June 16, 2008
Logic,
Your train of thought makes no sense. You cannot base your opinion simply on what the experts think. Afterall, the sun does not revolve around the earth and the world is not flat.
Logic
June 16, 2008
mrparker, you are absolutely right. The “experts” are not always right, and in fact can be terribly wrong, as in your example of the flatness of the world. I guess I am just exasperated with the measures of the wages of wins, mainly because I don’t think current basketball statistics provide a very good description of what happens in a game.
So it seems I’m just exasperated by any statistical “system” used to measure players.
I understand the attempt to interpret statistics, but when the statistics themselves are flawed (point, rebounds, steal) and don’t necessarily represent even half of the entire game (again, as opposed to baseball, where the stats can describe most of the game pretty well), you can come to completely off-base and/or meaningless conclusions about the value of a player within a team.
For example, try taking the stats of soccer (goals, steals, even plus/minus) and see if you can figure out who the best player is using those measures.
Basketball stats are somewhere in between the stats of baseball and soccer, so I expect the conclusions that you can draw to be somewhere in between those two as well. Unless, of course, you think you do have a “wages of wins” of soccer, in which case I would love to see it and what it tells us.
J
June 17, 2008
Logic makes a good point. For instance, the scoring of “assists” is so flawed as to render the statistic near meaningless. Chris Paul gets an assist when he passes the ball to David West, who can then dribble once or twice & shoot a turnaround J. David Friedman has documented this during the playoffs.
Further, are you aware of or do you use any statistic to measure the impact of a play like the following (which occurred in Game 4 or 5 of the Finals): Bryant drives along the left baseline, 2 or 3 Celtics defenders rush to help. Bryant swings a crisp cross-baseline pass to an open player on the right baseline. Celtics defenders scramble to rotate. That player tosses the ball along the perimeter, and Celtics defenders scramble to rotate. That player (Fisher, I think), then passes it again Walton, cue more rotating, and Walton dishes it to Vujacic, who drains a three-point shot.
Walton is credited with the assist, but clearly that play belongs to Kobe: his penetration scrambled the defense, and frankly only a drive by him (or a player of his quality) would have so fully occupied the attention of the remainder of the Celtics’ defenders to create the opportunity for the Lakers to pass the ball across the perimeter as they did.
That example is just another illustration of Logic’s point re: soccer. Soccer, like basketball, consists of highly fluid and interdependent team play, components of which are not easily measured. Thierry Henry, a former forward for Arsenal, would constantly make distracting diagonal runs from the center to the left, dragging a central defender along with him. This would open space centrally for midfielder A to run into, and if midfielder B passed the ball to the MF A, who scores a goal, then Henry doesn’t show up on any stat sheet (and, indeed, MF B might not either, as soccer leagues are not terribly consistent about keeping “assists” stats), but clearly his run created the goal by carving open the space.
Likewise, Kobe might pass the ball to Gasol at the freethrow line and make a dash to the bucket down the right. KG & Perkins see that and both see that & slightly position themselves to prepare for meeting Kobe at the basket. As a result, Odom is able to escape Perkins’ attention, and when Gasol whips him a pass, Odom is able to catch & dunk on the left side. Does Kobe get any statistical credit for that? Almost certainly no, but the respect that KG and Perkins have for him undeniably created that play.
This is why your blog is neat, but ultimately not an outstanding analytical tool for the NBA. Statistics can certainly tell *a lot* about the game, but so, so, so many things remain entirely uncaptured by statistics.
PS Logic — ProZone and one other system have been trying to develop stat systems to measure the impact of players in soccer, and many coaches (including Arsene Wenger at Arsenal) do use them. A key component is tracking the distance covered, passes attempted, completed, etc, etc. Still I doubt any system will ever fully measure performance in soccer.
ilikeflowers
June 17, 2008
Wow, a lot of verbage for two hypotheses.
[1] The assist stat is crap.
[2] Player x makes his teammates better.
J
June 17, 2008
The verbiage is necessary, ilikeflowers, b/c a simple assertion that “player x makes his teammates better” can be laughed off or blithely dismissed.
Offering a few, very common scenarios that occur multiple times in a game but do not appear in any way in the box score, however, suggests the fundamental weakness of basing one’s analytical approach entirely on the box score.
Tball
June 17, 2008
J
I don’t think you can take an anecdotal situation that is uncaptured by statistics and throw out the statistical tool.
First, in your examples, Kobe didn’t find the open man. If two or three players rush to defend Kobe, then he should be able to find the open man and, if he cannot, give credit to the man that does. If Walton doesn’t find the open man, maybe the play breaks down there and Kobe has not had a positive influence on the play at all.
Baseball, the most convenient of sports for statistics, has similar holes, like offensive players not getting credited for moving a player from first to second or second to third on a ground ball (unless it is a sacrifice). But the fact is, it is such a small impact on the game, it lacks statistical value. Similarly, in your first anecdote, half the time Kobe will find an open man and 70% of the time that player will score (hypothetical data). As a result Kobe picks up about 5 assists/game in those situations. There is a finite number of players that average more than 5 assists per game in the NBA and most of those players probably initiate their fair share of passes that become additional passes that become assists. The point is, it averages out.
That said, I will do you one better. Rondo is playing PG for the Celtics when Pierce has the ball. Kobe goes to double him, leaving Rondo open. Kobe knows Rondo won’t shoot it, Pierce knows Rondo won’t shoot it, so Pierce tries to force something else that causes a turnover. Rondo’s unwillingness to shoot diminishes Pierce’s win score.
At the end of the day, though, the win score system appears fairly adept at predicting winners and, particularly, at predicting winners after mid-season trades. It pegged the Celtics as a Finals team after the KG trade and the Lakers as the other Finals team after the Gasol trade. It pegged Posey as one of the best free agents available last summer when the C’s signed him for half of the exception.
Don’t look for anecdotal activity the tool does not explicitly account for to judge the tool, judge the tool by its ability to predict future performance.
Anon
June 17, 2008
Anecdotaly, Pau Gasol was incredible in Game 5. He battled for four quarters and showed that he is, actually seven feet tall. His rebounding was huge. The one thing that he has done consistently throughout the series is attack Kevin Garnett. Gasol has attacked him in the post and either out muscled Garnett or gotten him in the air for the up and under. Who saw that coming? And who saw Kobe Bryant disappearing for seven out of the eight quarters in Games 4 and 5?
J
June 17, 2008
Anon, if you thought Kobe disappeared for 7 of 8 quarters in Games 4 & 5, you weren’t watching (or were heavily inebriated). Kobe was crucial to the huge lead in Game 4 — his dribble penetration repeatedly threw the Celtics’ defense into disarray and enabled his teammates to score, although sometimes the bucket came after a pass or a dribble, such that Kobe did not receive an “official” assist. But if you watch that game again, it is clear that Kobe pulled the strings for that incredible Game 4 first half performance and deserves a great deal of credit for making offensive plays that allowed his teammates to be open and do good things with the ball.
TBall — your Rondo example is great. Rondo has *killed* the Celtics in Games 4 and 5 — has that shown up in the stats used here? If not, I would suggest that the stats are tremendously flawed, even if they manage to display some impressive general predictive power over the course of a season.
Also, you significantly underestimate the amount of times that scenarios such as I described occur. Kobe passes & cuts nearly every possession. And I think you have forgotten the incredible play that I described, which I believe was in Game 4. When Kobe drove, he did find an open man, but that player elected not to shot, thinking that another player might be “more open.” The Lakers wound up with all 5 players touching the ball in quick succession and a made 3-pt shot by Sasha V. But none of that happens without Kobe’s dribble penetration. That basket simply does not happen without Kobe’s drive and the respect that the Celtics have for Kobe’s driving ability. The first player to whom Kobe passed *was* open and could easily have decided to take & drain a shot, but instead he passed up the shot and the Lakers whipped the ball around the perimeter. Kobe still made an important offensive play even if the first pass-recipient did not make the shot.
Likewise, consider a couple plays from Game 5. On the high pick-and-roll, Kobe rolls off Gasol’s screen, gets trapped by KG, and Kobe finds Odom who has flashed to the free throw line. People rush to Odom, who then feeds Gasol who has rolled to the hoop for an easy dunk. Pts to Gasol, assist to Odom, but nothing to Kobe. But, clearly, none of that happens unless KG decides that he needs to run out & trap Kobe; nor does it happen unless Kobe recognizes that KG has unbalanced the Celtics’ defense and passes to the flashing Odom.
Logic
June 17, 2008
Tball – What is the tool’s record in predicting future performance? I’m not sure this has been documented anywhere.
Owen
June 17, 2008
So many Kobe threads, so little time.
J – You should check out 20 Second Timeout, that’ s a blog you should absolutely love.
ilikeflowers
June 17, 2008
Anon vs J, AKA the beauty of anecdotes.
J, Player x makes his teammates better succinctly describes the situations mentioned. If it is as important a factor as you posit, then it should have a measurable effect, even with current statistics. Offensive efficiency +/- would tell you something, especially if the team doesn’t have an assigned secondary dribble-penetrator (dp) for whom they run set plays. Looking at the list of high-W48 guards, I see that most of them can be thought of as scoring dp’s. Dribble penetration may even just be a part of the position and its importance already adequately captured.
How good of a dp do you have to be in order for the team to obtain most from the benefit of that strategy? i.e. is it most of it’s value already captured in the position adjustment?
Is having two moderately talented dp’s better than having one super talented one?
How do dp’s affect turnovers?
How do dp’s affect rebounds?
Overall, how important exactly is the presence of a dp to a team’s success?
How does the dp strategy compare to other offensive strategy’s?
And on and on.
Vince Gagliano
June 17, 2008
By the way, even though Allen and Gasol are their team’s leaders in PAWS for the series, there is more to this story than meets the eye.
Through the five Finals games, Paul Pierce has led the Celtics in PAWS three times, and Gasol twice.
However, Allen is the Celtics’ PAWS leader in the other two games, while Derek Fisher, Sasha Vujacic, and Lamar Odom lead in the other three.
So if Pierce has a strong Game 6 in a Celtics win, one can make the argument that he deserves Finals MVP even if he is second fiddle to Allen, by virtue of being the top performer in more games. Same goes if he does well in Game 7.
Conversely, if the Lakers just happen to sweep the Celtics in Boston, Gasol would certainly make a case for MVP, although it’ll likely be a co-MVP if Kobe gets his way. However it goes, Bryant should not be the sole winner when it’s all said and done.
Another strike against Allen: His strongest performance was in Game 3, a Laker win.
Vince Gagliano
June 17, 2008
One of his strongest performances, sorry.
Dadgummit, I should really think these things over.
J
June 17, 2008
The notion that Gasol has been a more effective player than Kobe for the Lakers in this series is absurd. Try watching a game, and observing aspects of it that do not necessarily appear in the box score (which, you know, is also influenced by quirks like differing standards for awarding assists and hamstrung by an inability to measure every aspect of the game).
Oh, and I missed this from the main post: “I think the Boston collapse (and it would be a collapse) has to be at least partially attributed to the injuries to Kendrick Perkins and Rajon Rondo.”
This is *hilarious*!!!
The absence of Rondo is hugely *helpful* to the Celtics. Even a pro-WoW commenter above noted that. Switching Kobe to have him nominally guard Rondo while roaming on the defensive end was the crucial strategic move that has allowed LA to jump out to its huge early leads. Rondo barely played in Game 5 after helping dig the Celtics in the hole; with him out, the Celtics rallied and made it a close game.
If the author of this blog seriously thinks that Rondo’s presence on the court has *helped* the Celtics, he is not watching the games at all or very closely — and his model is absolutely terrible.
Also, it is non-sense for the author to state that “the other two members of the starting line-up in Boston (Rondo & Perkins) were quite productive this year” in this context of assessing the impact Rondo has had in the Finals. Rondo may well have been effective in the 82-game regular season, in which he could have faced the Lakers only twice. But anybody with any familiarity with basketball could tell you that Rondo has been a giant liability to the Celtics since the series switched to LA and Jackson switched Kobe to nominally guarding Rondo.
PS I do read 20-Second Timeout, and comment there fairly frequently.
Owen
June 17, 2008
“The absence of Rondo is hugely *helpful* to the Celtics.”
Really?
You really think the Celtics are better off as team not playing an injured Rondo than they would be with him fully healthy?
Sort of a bizarre opinion considering how well he played for them all year and what he did in Game 2. I think you must have misunderstood point DJ was making.
Anony
June 17, 2008
You don’t seem to understand the point that the reason Rondo is so ineffective is because he’s injured. He has been a terrible shooter all season long and he’s still a good and productive player because he’s so quick, and he’s a good ballhandler, passer, defender, and rebounder. The injury prevents him from making all the little plays he makes that help the team, and since he was never a good shooter, that’s why the injury hurts them.
And you should try to analyze Gasol’s production compared to Kobe without being massively biased in favor of Kobe. Gasol has been amazing in this series. In game 5, he was the primary person responsible for getting Garnett in foul trouble by attacking him on offense (but the stats don’t measure that, must be a fatal flaw!), and he scored very efficiently, made no mistakes, and got some big rebounds. He’s been playing solid, low-mistake basketball all series. Kobe, on the other hand, while he has been good for stretches, has disappeared for quarters at a time, and has shot a horrible percentage. I’m not saying that Gasol is a better player than Kobe Bryant, but in this series you could definitely make an (anecdotal) argument that Gasol has helped the Lakers more than Kobe has. Good things seem to happen every time he touches the ball.
Vince Gagliano
June 17, 2008
Then again, Rondo has a 3.57 assist-to-turnover ratio in the playoffs. Sam Cassell scores a 2.08 (25 assists, 12 TO) and Tony Allen a 2 (2 assists, 1 TO).
He’s far from the best shooter on the Celtics, but he helps direct the offense and gets the ball to Boston’s Big Three. Hence, he is an important “role player” for Boston.
Perkins is a little bit thicker than Gasol, and can push him around. However, Garnett works better in the paint and has more strength. Theoretically, it shouldn’t matter too much for the C’s, but Perkins is still an important defensive weapon.
J
June 17, 2008
Vince, an assist is not an obvious, accurate statistic like, say, a strikeout or a home run. Scorekeepers do not consistently award assists on any clear criteria. David Friedman has demonstrated this by analyzing the scoring of Chris Paul’s games in the NO-SA series. Scorekeepers gave Paul “assists” repeatedly on plays in which David West took more than enough action to render the “assist” inappropriate.
For that matter, a turnover is not a black-and-white statistic either. Consider a point guard who feeds the ball to a post player who is in a poor position, off-balance and with a help defender very nearby. That post player reaches to catch the less-than-perfect pass, is swarmed by a help defender, and winds up having the ball swatted out of his hands by the help defender. The post player receives the TO stat, but perhaps a good PG would not have created that TO by choosing to pass to another player.
Certainly one can debate the frequency with which scenarios like the above occur, but I would suggest that if you watch carefully, they do occur not infrequently. In any event, they occur enough to have a big impact in a low sample size, like one’s numbers during 5 games of the NBA finals.
Also, to whoever asserted Rondo is “injured” — Doc Rivers disagrees with you.
***
Rondo hurt his left ankle in Game 3, but hasn’t had to miss any games. Apparently the injury wasn’t the problem.
“He’s fine,” Rivers said. “He’s a second-year basketball player and I never lose light of that. There’s a lot of pressure on him, I’m not going to put more on him. I’m just going to keep coaching him and trusting him, and he got us here. He’s gotten us here. But clearly he’s not playing as well as he would like to, but we believe in him.”
***
http://sports.yahoo.com/nba/news;_ylt=Aje.RBcaGnMDavvgDJx7hNSLvLYF?slug=ap-nbafinals-notebook&prov=ap&type=lgns
Anony — I guess we’ll just have to disagree. Gasol has had his moments, but he is not the difference-maker for the Lakers. When he is setting good strong screens and rolling aggressively to the hoop, he has done very well in tandem with Kobe, who has done very well at drawing multiple defenders and accurately passing out of those situations to the Lakers’ (and Gasol’s) advantage. But Gasol hasn’t done this consistently, and his finishing around the basket has been rather inconsistent and weak.
Also, another huge part of the Lakers’ better performances has been the disruptive effect created by Kobe “guarding” Rondo and roaming defensively. Kobe’s defense has killed the Celtics, led to easy transition scoring, and helped the Lakers jump out to giant early leads in the last two games.
porteno
June 17, 2008
I’m tempted to argue here, but honestly I ask myself: would that be the best use of your time?
Clearly we have some commenters who have pre-judged the use of statistics out of ignorance. That is sad, but no matter how perspicaceous my comment, I am unlikely to change such an unwilling intellect.
I appreciated the anecdoal soccer arguments, but even the anecdotal arguments were quite flawed. (Go Gunners!) This comment, however, amused me:
*If not, I would suggest that the stats are tremendously flawed, even if they manage to display some impressive general predictive power over the course of a season. *
Predictive power is very strong, that ignoring it seems like something only a fool would do.
Logic
June 17, 2008
“On a skill level, they are equal — no ifs, ands or buts,” says Will Perdue, a former Jordan teammate and an ESPN analyst. “Kobe has worked on his game to be better than Michael. But he always has done something to screw up his public image.”
Who has more credibility? A former teammate or a system of assorted statistics?
This would be like if someone was telling Columbus that no, the world IS flat, AFTER Columbus reached what he believed to be India.
Go tell Will Perdue that Kobe is not as productive as MJ … and Steve Kerr, and B.J. Armstrong (all on record that Kobe’s skills are on par with MJ … and not just dunking skills).
Logic
June 17, 2008
“On a skill level, they are equal — no ifs, ands or buts,” says Will Perdue, a former Jordan teammate and an ESPN analyst. “Kobe has worked on his game to be better than Michael. But he always has done something to screw up his public image.”
Who has more credibility? A former teammate or a system of assorted statistics?
This would be like if someone was telling Columbus that no, the world IS flat, AFTER Columbus reached what he believed to be India.
Go tell Will Perdue that Kobe is not as productive as MJ … and Steve Kerr, and B.J. Armstrong (all on record that Kobe’s skills are on par with MJ … and not just dunking skills).
ilikeflowers
June 17, 2008
‘Who has more credibility? A former teammate or a system of assorted statistics? ‘
I would give more credibility to whatever approach produced the best predictions.
“This would be like if someone was telling Columbus that no, the world IS flat, AFTER Columbus reached what he believed to be India.”
No, it would be more like going from the Earth is round to the Earth is spherical with a bulge at the equator.
J
June 17, 2008
no porteno, not out of ignorance, but out of suspicion for the accuracy of the stats, and their ability to capture the full panoply of what occurs in an NBA game.
Also, would * love* to hear your views on how my soccer anecdotes are “quite flawed.” Glad to hear you’re a Gunners fan though. For an extra tidbit, that goal I described (with Henry dragging a central defender wide left so a central player could score) was the 2004 game against Chelsea that Arsenal won 2-1 at Stamford Bridge after conceding an early goal. I believe it’s the last game that Chelsea lost at the Bridge.
Finally, what I meant in the comment about predictive power is that a statistical model might do pretty well in predicting a team’s performance over 82 games, but I question its explanatory power or predictive power in the NBA Finals, a 7-game series. The stats from the 82 games were compiled against massively varying levels of competition, whereas the 7-game series involves a very high quality opponent. The 7-game series also allows coaches to hone in on opposing teams to a degree of detail they cannot replicate in the regular season, and this may result in strategies that significantly affect the effectiveness of a player, lineup or offensive scheme (see, e.g., Rondo, Rajon).
Jed
June 17, 2008
Is it just me or does Logic’s analogy support the exact opposite point that he wants to make?
To most any observer, it is “common sense” the Earth is flat. After all, we don’t see any curve when we look around us. But rigorous mathematical analysis and careful physical observation reveals that the Earth is in fact curved. Prof. Berri’s attempt to mine the data on NBA players is similarly about critically examining what common sense tells us to be obvious.
Plus, since Columbus never reached India in the first place, I wouldn’t put much stock in his assertion that he proved the Earth round to begin with!
J
June 17, 2008
PS Porteno, I should say that I love MoneyBall and statistics in baseball. Firejoemorgan is a fantastic blog.
I just feel very strongly that basketball, like soccer, is not adequately captured in any set of statistics (and to a degree that far exceeds any similar such gap in baseball). Sure, stats can offer some very strong, interesting insights, but ultimately, those insights must be paired with actually watching the game and understanding plays and effects that do not show up in the box score, of which I have given many hypothetical examples in this thread.
NickP
June 17, 2008
I think all would agree that Michael Jordan, Isiah Thomas, and Kevin McHale are three of the best players of the last 25 years, to some degree. Jordan is obviously an all-time great, while Isiah and McHale were both above average, very productive players.
I also think all would agree that those 3 have proven to be horrible judges of the talent it takes to field a winning basketball team. Really, their opinion of what is or is not a great basketball player should be just about meaningless. Being good enough to play professional basketball doesn’t mean you *know* who CAN or DOES play good basketball. So extending that point to Will Perdue…why should we care at all what he thinks about Jordan vs. Kobe? What has he done to prove he should be considered an *expert* on the topic? He’s paid to give his opinion, so he has an audience…but that doesn’t mean his opinion has any value.
ilikeflowers
June 17, 2008
Anecdotes For Fun ™
The Lakers are going to destroy Detroit in the 04 finals because they have Shaq AND Kobe!
Wade must be better than Kobe, because he beat a team from the mighty West with nobody besides and old Shaq! (Here, the anecdote stumbles upon a likely truth)
The Lakers are going to beat the Celtics because almost everyone at ESPN thinks that they will! After all they have the best player in the universe! (Maybe they will, we’ll see)
J
June 17, 2008
To join the pile-on, I obviously rate Kobe very highly, but my opinion does not rely on the views of Will Perdue in any way and I agree that his views ought to receive no greater credence simply because he managed to play in the NBA.
Mark
June 17, 2008
While I’m a Laker fan, I agree that Kobe is overrated in general. However, I think the statistics do lie in the case of the finals. Watching the games it is clear Boston focuses unusually huge amounts of defensive pressure on Kobe. When Kobe gives up the ball the remaining team is typically playing 4 on 3.5.
Now you can make the argument Kobe isn’t earning this attention. One could imagine you could replace Kobe with Coby (Karl) and get the same effect as long as Boston promised to overplay Coby the same way. Still, Kobe’s presence seems to be helping the Lakers beyond the boxscore.
We all know the boxscore is good but not perfect. I don’t agree with those that simply dismiss stats out of hand. But in this case, I think we can clearly see what the boxscore is missing.
Logic
June 17, 2008
“Anecdotal” is no worse than “statistical” when the statistics in question are basketball statistics.
So, I’ll bite – what is the predictive value of WOW? Did it predict that the Atlanta Hawks would all of its home games against the Boston Celtics in the first round? Did it predict that the San Antonio Spurs would beat the Suns in five? I know it chose the Celtics as the favorite, but didn’t the “experts” come to that conclusion when the playoffs started, too? Does Mr. Berri even make predictions for every game? I’m not deliberately being antagonistic, I really want to know. I can do some googling too, but maybe you guys can tell me what the track record of WOW is in a “non-anecdotal” sense.
Logic
June 17, 2008
* Correction – did it predict that the Atlanta Hawks would WIN all if its home games against Boston?
ilikeflowers
June 17, 2008
It predicts what it’s supposed to, wins for the season and likely winners in each round given the lineups for the season. Efficiency differential and home court predicts the winner (unless the lineup changes for some reason), I leave it to you to look up the actual winning percentage, it’s somewhere on this site and it’s not specific to W48. Who knows, if we account for the increased minutes of the Spurs big two then maybe it predicts the Spurs winning as well, or makes it close enough to only be a mild upset.
Anecdotaly, Kobe looks like a mostly good sometimes elite player in this series (which is how I’ve always thought of him anecdotaly) and that’s what the numbers say too. And of course W48 only thinks Kobe is one of the top two SG’s in the league (Wade will be the best if he recovers fully from injury) which means that it’s obviously missing something.
Logic
June 17, 2008
Let me say what my viewpoint is NOT, before people inevitably distort my viewpoint.
It is NOT: “How can a statistical model yield decisions that are better than those reached from a lifetime of experience and personal observation? ”
It IS: How can a statistical model be created based upon incomplete data? The statistics in basketball used today (even plus/minus) describe about 40% (my wild guess, I don’t think it’s more than 70% on the extreme end) of the factors that produce wins.
And I agree with some here, predictive value does give credibility to the analysis. But even that is looking at the team statistics. It gets even murkier when trying to separate an individual player’s contribution from the team as a whole.
Owen
June 17, 2008
Logic – What do you think about this bit from one of DJ’s previous posts?
“Since 1982 the weaker team – in terms of efficiency differential – has only won five times. In other words, across the last 25 years, efficiency differential has called the champion correctly 80% of the time.
One should note that three of these times, the weaker team (in terms of efficiency differential) – due to a better won-loss record – had home court advantage in the playoffs. So these upsets were not surprising.
In sum, over the past 25 years, truly surprising upsets – in terms of efficiency differential – have been rare.”
Owen
June 17, 2008
J – Just wondering, is Rondo still killing the Celts? :-)
J
June 17, 2008
Yes, actually. He was terrible in the first quarter, missing his first several shots. Right now, in the 4th, he shot 8-20. Had the Lakers made more of their shots early on, the game might have been very different. As it was, Space Cadet Radmanovic and Sasha V made some very poor plays early for the Lakers (leaving Posey & House open behind the 3-pt line on several occasions).
Sure, Rondo wound up having a very good defensive game, but he remains a major offensive liability.
Also, I hope those who thought Gasol was playing the best on the Lakers actually watched this game. KG pancaking him early on that fastbreak basically sums up my feelings about Gasol’s contribution to this series.
Owen
June 17, 2008
J – Six steals, eight assists, seven rebounds, just one turnover, and 21 points with a ts% of 44%. The latter is not great, but overall he was absolutely outstanding, posting a WS of 12 in 31 minutes.
It’s not just about scoring…
J
June 17, 2008
I recognize that — that’s why I referred to his missing his first several shots. Those allowed the Lakers to get out in transition but LA failed to capitalize on Boston/Rondo’s early shooting woes. At 10-8 or 10-10, I recall the announcers noting that Boston was shooting some ridiculously low percentage from the field (but had a few FTs). Had the Lakers been making their open looks on the other end, the start of this game would have resembled the last couple.
Also, I acknowledged that Rondo “wound up” playing well. He is very much a “confidence” player, and once the Celtics jumped out to a big lead (despite his early poor shooting, and, yes, because of his nice defensive steals), he dropped a couple of his open looks (but still wound up with a pretty poor shooting percentage).
He is still a terrible offensive point guard. And this Boston team, for as outstanding of a regular season as it had, really scraped its way through the playoffs and looked somewhat lucky to beat its final three opponents. PJ Brown saved them against the Cavs, and both the Pistons and Lakers really played them very close, even if they lost 4-2. The Lakers could very easily have stolen Game 2 (even with the massive FT disparity) and then wasted their fantastic first-half performance in Game 4. I guarantee that the Celtics do not repeat.
ilikeflowers
June 17, 2008
No reason to think that the Celtics can’t do it again next year. Rondo, Powe, and Davis should improve – and they’re already very productive. Hopefully Cassell will be jettisoned in favor of House. Barring injury and with the continued improvement of their excellent 1st and 2nd year players they might be better than the Lakers even with Bynum. And by ‘better’ I mean have a better efficiency differential (just like this year).
Watching Posey this series was deja vu from the Miami-Dallas series.
mrparker
June 18, 2008
When will the Kobe nation stop making excuses for him? I work in a bar and somewhere around the 3rd quarter the same patrons who had been predicting Lakers in 5 started pronouncing the deficiencies of Kobe’s teammates as the reason that Kobe was not going to get his 4th ring.
This seems to be what they are saying. Kobe is the greatest player in the world except in the case when his teammates are not as good as him in which case he will not be held accountable for his awful numbers.
He’s alot like GWBush. Everyone hates him except for the people who love him who think he is the best president ever.
David
June 18, 2008
Logic – as for the predictive value of WOW, one impressive anecdote is in regards to the Iverson trade. At the time of the trade, the 76er’s were 5-18. WOW correctly predicted that their remaining record would be 30-29:
https://dberri.wordpress.com/2006/12/19/first-impressions-of-the-iverson-trade/
I believe that Iverson is more talented than Andre Miller. But if you take any NBA team and replace their PG with Miller, and then take that same team and replace their PG with Iverson, I know that the first team will win more games. That’s pretty much an established fact now. It takes cold-hearted statistical modeling to discover facts like this. Humans have trouble with this because they confuse talent with winning.
porteno
June 18, 2008
J –
At times I thought you were a troll, but let me attempt to persuade you with a pithy comment: you argue that current statistics don’t track anything in basketball.
I used to think that too. The problem is that the data does not agree with that assertion. Dave’s model explains something like 97% of the game, so clearly the statistics ARE tracking something worth paying attention to. As hard as it is to believe, that’s true.
Furthermore, the model doesn’t just explain the past. If it explained the past but had zero predictive power, then it would just be fitting the line to the data. But, Berri’s model has predictive power!
Those two things make it unwise to blithe dismissing the model because it doesn’t comport to your visual perceptions.
Jeff
June 18, 2008
This just in: Kobe’s sweat cures cancer. I’ll bet your stupid mathematical model couldn’t predict that, David.
I wonder how many Kobe-philes are still holding out for the discovery of wmds in Iraq.
ilikeflowers
June 18, 2008
Since Paul Pierce is now obviously the best player on the planet having bested LeBron and Mamba, his sweat must edge out Kobe’s.
Tball
June 18, 2008
Since Pierce was clearly Kobe’s superior on both ends of the floor, does Pierce get to proclaim himself “King Cobra” – the only snake more deadly than the black mamba?
It takes a black mamba 20 minutes to kill a man, whereas a king cobra can kill a man in 15 minutes. These stats were provided by Wikipedia and do not account for situations where the black mamba passes a dagger to Gasol, who passes it Luke Walton, who fumbles the dagger and kills himself.
Logic
June 19, 2008
Tball- why just Pierce? Ray Allen was better than Pierce, KG, and Kobe according to PAWS.