You Are Here > Basketball-Reference.com > BBR Blog > NBA and College Basketball Analysis

SITE NEWS: We are moving all of our site and company news into a single blog for Sports-Reference.com. We'll tag all Basketball-Reference content, so you can quickly and easily find the content you want.

Also, our existing Basketball-Reference blog rss feed will be redirected to the new site's feed.

Basketball-Reference.com // Sports Reference

For more from Neil, check out his new work at BasketballProspectus.com.

Best Performances in a Finals Loss (1991-2010)

Posted by Neil Paine on June 14, 2010

Last night, Kobe Bryant boldly attempted to take over Game 5 of the Finals, pouring in 23 consecutive Laker points in the 3rd quarter on a collection of the toughest shots you'll ever see. However, L.A. couldn't get stops during that span, and nobody but Bryant was scoring, so the Celtics were actually able to extend their lead even as Kobe's outburst was taking place. Bryant finished the game with 38 points, but half of them came in that 8-minute stretch during the 3rd quarter, and he couldn't will L.A. to a late-game charge even as Boston seemed on the verge of a collapse in the final minutes.

The frustration was apparent in Bryant's expressions and body language throughout the 4th quarter, as Kobe was unable to do any damage from the floor in the final 8 minutes of the game. But despite his failure to drive a stake into the Celtics' hearts in crunch time, Bryant's performance was still one of the most valiant in recent Finals history by a member of the losing team. According to Statistical +/-, here are the best individual performances in a Finals loss (minimum 30 minutes played):

Date Player Tm Opp Loc MP FG FGA 3P 3PA FT FTA OR TR AS ST BK TO PF PTS SPM TMates
6/16/1993 Charles Barkley PHO CHI A 46.0 10 19 0 2 12 15 3 12 10 3 1 1 1 32 18.60 -20.57
6/16/1999 Kurt Thomas NYK SAS A 31.0 5 11 0 0 3 4 7 16 0 4 0 0 4 13 18.25 -21.61
6/7/2009 Rashard Lewis ORL LAL A 45.3 12 21 6 12 4 4 5 11 7 1 0 2 2 34 15.15 -14.40
6/7/1996 Shawn Kemp SEA CHI A 40.0 8 18 0 1 13 16 8 13 2 1 4 3 3 29 14.35 -12.98
6/2/1991 Michael Jordan CHI LAL H 40.0 14 24 1 1 7 9 2 8 12 3 0 4 5 36 14.01 -17.69
6/9/2005 Chauncey Billups DET SAS A 43.0 9 16 1 4 6 6 0 4 6 4 1 1 1 25 13.05 -26.34
6/8/2003 Kerry Kittles NJN SAS H 34.0 8 16 3 5 2 3 1 4 1 3 2 0 2 21 12.45 -18.21
6/15/2003 Kerry Kittles NJN SAS A 34.0 5 12 2 6 4 4 0 4 2 3 0 0 2 16 12.23 -16.89
6/23/1999 Charlie Ward NYK SAS H 41.0 4 9 2 5 1 2 2 4 8 4 0 1 3 11 11.44 -21.32
6/19/1994 Derek Harper NYK HOU A 42.0 2 10 1 5 5 6 1 4 10 4 0 1 3 10 10.91 -8.27
6/13/2004 Shaquille O'Neal LAL DET A 47.0 16 21 0 0 4 11 3 20 2 0 1 2 4 36 10.71 -16.00
6/6/2010 Pau Gasol LAL BOS H 42.0 7 10 0 1 11 13 3 8 3 1 6 1 3 25 10.35 -22.59
6/15/2008 Paul Pierce BOS LAL A 48.0 10 22 2 6 16 19 1 6 8 1 0 5 5 38 9.93 -11.48
6/9/2002 Jason Kidd NJN LAL H 43.0 13 23 1 5 3 5 1 5 10 3 1 2 3 30 9.90 -15.71
6/11/1993 Charles Barkley PHO CHI H 46.0 16 26 0 1 10 12 6 13 4 1 1 2 3 42 9.90 -16.30
6/10/2008 Ray Allen BOS LAL A 41.4 8 13 5 7 4 6 3 5 2 1 1 1 3 25 9.84 -11.70
6/19/2000 Dale Davis IND LAL A 35.0 8 10 0 0 4 6 5 14 3 0 3 0 4 20 9.66 -8.82
6/7/2009 Dwight Howard ORL LAL A 47.2 5 10 0 0 7 9 3 16 4 4 4 7 4 17 9.54 -9.94
6/11/1995 Anfernee Hardaway ORL HOU A 41.0 4 10 1 4 10 11 1 4 14 1 2 3 4 19 9.32 -7.44
6/13/2010 Kobe Bryant LAL BOS A 43.9 13 27 4 10 8 9 2 5 4 1 1 4 5 38 9.28 -12.01
6/21/1999 David Robinson SAS NYK A 39.0 6 12 0 0 13 17 1 10 1 0 2 2 3 25 9.09 -13.29
6/15/1994 Hakeem Olajuwon HOU NYK A 43.0 14 20 0 0 4 4 0 8 3 0 5 2 4 32 9.09 -15.05
6/19/1994 John Starks NYK HOU A 46.0 9 18 5 9 4 5 1 2 8 2 0 2 5 27 9.07 -7.42
6/1/1997 John Stockton UTA CHI A 38.0 6 10 2 4 2 3 1 3 12 3 0 7 2 16 8.67 -5.52
6/14/2000 Reggie Miller IND LAL H 50.0 9 19 6 9 11 12 0 5 3 1 1 1 1 35 8.59 -13.82
6/18/1993 Michael Jordan CHI PHO H 44.0 16 29 2 7 7 10 1 7 7 0 2 2 5 41 8.42 -22.67
6/6/2001 Shaquille O'Neal LAL PHI H 52.0 17 28 0 0 10 22 6 20 5 1 0 4 3 44 8.39 -17.94
6/12/1994 Derek Harper NYK HOU H 39.0 9 15 3 7 0 0 2 7 3 4 0 3 4 21 8.34 -15.11
6/11/2003 Tim Duncan SAS NJN A 39.0 10 23 0 0 3 3 8 17 2 1 7 3 4 23 8.27 -4.20
6/8/2008 Kobe Bryant LAL BOS A 40.5 11 23 1 3 7 7 1 4 8 3 0 4 3 30 8.27 -10.11
6/5/2002 Jason Kidd NJN LAL A 43.0 11 26 1 3 0 1 6 10 10 3 0 1 1 23 8.27 -9.25
6/6/2003 David Robinson SAS NJN H 33.0 3 6 0 0 4 6 4 8 1 2 2 1 0 10 8.15 -11.66
6/12/1998 Toni Kukoc CHI UTA H 43.0 11 13 4 6 4 7 1 6 0 1 1 2 2 30 7.92 -13.29
6/25/1999 Latrell Sprewell NYK SAS H 43.0 13 27 1 1 8 10 2 10 2 2 0 3 1 35 7.84 -11.92
6/20/2006 Josh Howard DAL MIA H 30.3 5 16 0 2 4 4 3 12 0 4 1 1 5 14 7.79 -11.68
6/14/1995 Nick Anderson ORL HOU A 31.0 1 5 1 4 1 2 2 7 4 3 0 0 4 4 7.41 -13.75
6/4/2009 Dwight Howard ORL LAL A 35.0 1 6 0 0 10 16 5 15 2 2 2 2 3 12 7.36 -30.65
6/12/1991 Sam Perkins LAL CHI H 37.0 5 12 1 4 11 13 3 9 3 0 2 1 4 22 7.34 -16.66
6/14/1998 Karl Malone UTA CHI H 43.0 11 19 0 0 9 11 5 11 7 1 0 5 2 31 7.26 -11.52
6/13/2003 Jason Kidd NJN SAS H 48.0 10 23 4 10 5 6 3 7 7 2 0 2 2 29 7.14 -22.49

Bryant was highly impressive in the loss, but his teammates had a combined -12.01 SPM:

Player Tm Opp GS MP FG FGA FG3 FG3A FT FTA ORB TRB AST STL BLK TOV PF PTS SPM
Kobe Bryant LAL BOS 1 43.9 13 27 4 10 8 9 2 5 4 1 1 4 5 38 9.28
Pau Gasol LAL BOS 1 38.1 5 12 0 0 2 3 7 12 0 2 0 1 4 12 1.48
Ron Artest LAL BOS 1 34.3 2 9 2 5 1 4 1 2 2 1 0 1 4 7 -3.82
Derek Fisher LAL BOS 1 34.2 2 9 0 1 5 5 2 4 2 1 0 2 3 9 -4.69
Andrew Bynum LAL BOS 1 31.6 3 6 0 0 0 1 1 1 0 0 0 1 3 6 -9.33
Lamar Odom LAL BOS 0 26.3 4 6 0 0 0 2 3 8 2 2 0 3 2 8 2.11
Jordan Farmar LAL BOS 0 13.8 0 4 0 1 1 2 0 1 1 1 0 0 0 1 -4.81
Sasha Vujacic LAL BOS 0 10.4 2 5 1 2 0 0 0 1 0 1 0 0 1 5 4.70
Luke Walton LAL BOS 0 7.2 0 0 0 0 0 0 0 0 1 0 0 1 0 0 -11.40
Shannon Brown LAL BOS 0 0.3 0 0 0 0 0 0 0 0 0 0 0 0 0 0 -7.02

Meanwhile, Boston received a strong outing from Paul Pierce and an amazing game by Kevin Garnett:

Player Tm Opp GS MP FG FGA FG3 FG3A FT FTA ORB TRB AST STL BLK TOV PF PTS SPM
Paul Pierce BOS LAL 1 42.7 12 21 2 4 1 2 0 2 2 1 2 0 4 27 4.58
Ray Allen BOS LAL 1 40.2 5 10 0 4 2 2 0 3 2 1 0 2 5 12 -2.70
Rajon Rondo BOS LAL 1 38.4 9 12 0 0 0 0 1 5 8 1 1 7 1 18 -3.29
Kevin Garnett BOS LAL 1 36.2 6 11 0 0 6 7 1 10 3 5 2 3 4 18 16.49
Kendrick Perkins BOS LAL 1 31.6 2 2 0 0 0 2 4 7 1 0 0 2 4 4 -3.51
Rasheed Wallace BOS LAL 0 14.8 2 4 1 2 0 0 0 4 0 0 1 0 0 5 1.25
Glen Davis BOS LAL 0 13.2 0 1 0 0 0 0 1 3 1 0 0 0 2 0 -4.22
Tony Allen BOS LAL 0 13.2 2 6 0 0 0 0 0 1 0 0 1 1 2 4 -16.13
Nate Robinson BOS LAL 0 9.9 2 4 0 2 0 0 0 0 4 0 0 1 1 4 -2.77

And now it comes down to two games in Los Angeles. As we saw last week, the Celtics now have a 64% probability of winning the series -- assuming the 2 teams are evenly matched -- after winning in the 3rd-most crucial situation possible in a 2-3-2 series (Game 5, tied at 2). The second-most crucial possible situation? Game 6, home team trailing 2-3... which is exactly what we'll be seeing on Tuesday night.

ShareThis

24 Responses to “Best Performances in a Finals Loss (1991-2010)”

  1. Hk Says:

    Neil just wondering, do you prefer Win Shares/48 or SPM in this situation?

    I've seen some conflicting results, and I consider Adjusted plus minus to be very noisy (which SPM is based on).

  2. Walter Says:

    Neil,

    John Hollinger wrote an interesting article today discussing who should win the Finals MVP. His arguement was basically that Kobe may become the second player on a losing team to win the award. Essentially the Celtics have had fairly even contributions from its big 4, each having a good or great game followed by mostly mediocre or a very poor game (see Ray Allen game 3) and it is hard to vote for any of them over Kobe.

    I would be interested to see how good the best player on the losing team has played relative to the MVP winner from the winning team. Is Kobe's performance significantly better relative to the other celtics and enough to warrant the award or have we had other situations where a player on losing team was phenominal but still did not get the award.

  3. Neil Paine Says:

    I like Win Shares for bigger samples than single games. I came up with some crazy results with it last week when I did the bench performance post (http://www.basketball-reference.com/blog/?p=6475)... Teams were getting more than 2 WS total in a game, etc. It's not Win Shares' fault -- it was never designed for single game use. Also, you could typically use something like Hollinger Game Score for single game results, because it's quick and dirty and a poor man's version of PER (although I don't really like to use PER anyway), but it doesn't factor in defense very much or very well, which is a big flaw.

    That why I like to use SPM if I want linear weights for small samples. Now that we've fixed the issues caused by the "versatility index", it's probably the best boxscore-based linear weights formula we have. The weights are "organic" (derived by regressing on real +/- results, as opposed to relying the opinions of the equation-maker), and over a large sample like the one I regressed SPM on, the "noise" of APM isn't really a concern. APM may be wildly off for some players because of its multicollinearity problems, but over the course of thousands of player-seasons, those issues even out and you get a good sense for how much the box score stats contribute to point differential. The biggest problem with SPM is that it relies on box score stats, which means it has a decent handle on offensive contribution but has to make a lot of educated guesses about players' defensive value.

  4. Neil Paine Says:

    (That was in response to #1, btw.)

  5. Jason J Says:

    Neil,

    Is that high SPM for Garnett because of the big steals in relatively short minutes?

  6. Walter Says:

    Neil, do you have a post where you provide te weights for the various box-score items?

  7. Neil Paine Says:

    The weights are here:

    http://www.basketball-reference.com/blog/?page_id=4122

    Garnett had all the things SPM likes -- a ton of steals, 2 blocks, good scoring rate, good efficiency, drew a lot of fouls, grabbed 10 rebounds, etc. He did a lot of "good stuff", which SPM sees as him being active at both ends and making a huge impact. I'd say SPM was right.

  8. Grum Says:

    The problem with SPM over small samples is that the coefficients of many statistics are influenced by what the statistic correlates with. For example, look at personal fouls and blocks. Both have nice positive coefficients, but they are clearly inflated by the fact that they correlate with strong defense. Fouls are particularly obvious, since they are actually harmful.

    But suppose some commits 5 fouls. What does that tell you about his defense in that particular game? Not much, yet he will be rewarded handsomely in his SPM score.

    Or suppose someone gets 0 blocks. What does that tell you about his presence in the paint for that particular game? Well, it makes it less likely that he dominated the paint. But there are plenty of games where a player like Dwight Howard dominates the paint by intimidation alone, without getting any blocks. In this game his SPM is likely to be understated. And the logical converse of this fact is that his SPM is likely to be overstated in games where he has 5 or 6 blocks.

    So IMO there's too much noise in SPM to be useful for a single game. Heck, there's probably too much to be useful for a single series. Blocks, steals, and fouls are the main things to look out for when looking for noise.

  9. Neil Paine Says:

    Personal fouls aren't included in the current regression, they weren't found to be significant. But I agree, SPM can be noisy because it isn't technically measuring how much each category is "worth" (in the sense that, say, 1 made free throw is worth 1 point), but instead measuring how that stat tends to correlate with APM. But I would say there's sample-size danger with single game numbers no matter which stat you look at. Something like Game Score seems pretty straightforward, but the weights are really just John Hollinger's opinions. SPM's weights are actually derived empirically, which is a major asset in my book.

  10. Neil Paine Says:

    I think your comment basically boils down to, "The box score doesn't track everything that determines a player's contribution". A statement with which I would wholeheartedly agree.

  11. Grum Says:

    The advantage of game score is that it pretty much ignores the things that aren't included in the box score. Therefore it usually doesn't overstate or understate them by too much.

    SPM, on the other hand, takes account of these things indirectly. This makes it more accurate over the long run after things average out, but in a single game it's basically just adding gratuitous noise.

  12. Grum Says:

    So what it comes down to is that game score will be chronically biased for certain players (specifically, players who make lots of non-box score contributions, and players who make few non-box score contributions) but will have little noise. And SPM will be just the opposite.

  13. Neil Paine Says:

    Right, I pretty much agree with that assessment. You can measure most of a player's offensive contribution with reasonable accuracy and ignore defense entirely (Game Score), or you can measure both offense and defense, but with less confidence because it's really measuring what the box score stats suggest a player's APM is. I suppose it's all a matter of choice at that point. My choice is almost always the latter, though -- I generally dislike Game Score, because it ignores defense and also because the weights are arbitrary. That's my same complaint with PER, really.

  14. Ricardo Says:

    "John Hollinger wrote an interesting article today discussing who should win the Finals MVP. His arguement was basically that Kobe may become the second player on a losing team to win the award. Essentially the Celtics have had fairly even contributions from its big 4, each having a good or great game followed by mostly mediocre or a very poor game (see Ray Allen game 3) and it is hard to vote for any of them over Kobe."

    It seems like Kobe has had two good games in the series though. To me, a Finals MVP on a losing team really ought to shine every night. I don't think Kobe's played to that standard in this Finals.

  15. koberulz Says:

    Isn't that like saying that if everyone in a race jogs lazily, the guy who comes first doesn't deserve the gold medal?

  16. Jason J Says:

    Neil - While PER is arbitrary, isn't it one of the more accurate predictors among the metrics out there? Also, I'd like to reiterate my love for SPM and ask again that it be made available in the Play Index searches. For the kids, man. Do it for the kids.

    Also IF the Lakers do lose this series (and the whole tenor of this defensive stallwart but offensively inconsistent road team up 3 going back to the opposition's homecourt to try to close it out in 2 games is reminding me an ungodly amount of the Knicks v. Rockets right now, and I'm really scared LA is going to take back to back home games and dash our collective hopes into Hollywood's hellfires), I don't think Kobe deserves the Finals MVP unless he closes out in spectacular fashion. I know we've only got 24 players to choose from, but just because Boston doesn't have one individual absolutely killing it, doesn't mean what has been a not-so-great series by Bryant overall should result in an award that has only once gone to a series loser. Garnett's defense has been tremendous since the series moved to Boston. Let the Ticket have the trophy and ditch his "not clutch" label.

  17. Neil Paine Says:

    I did an unpublished study a few months ago on which metrics (with no team adjustment) predicted the next season's results the best: NBA efficiency, win score (poor man's wins produced), game score (poor man's PER), SPM, alternate SPM, alternate win score, Sports Illustrated's linear weights formula (I believe David Sabino invented that one), Tom Thibodeau's linear weights (formula is in Harvey Pollack's most recent yearbook), Bob Bellotti's points created, and Kev Pelton's old linear-weights WARP formula. I regressed each metric 1,000 minutes to the mean (the exact amount of regression to the mean didn't really matter to the results, I just needed to regress by some amount) and predicted the following season's WPct using a minute-weighted average of the metric. Here are the results:

    Metric R
    spm 0.575
    alt_spm 0.564
    winscore 0.533
    alt_wscore 0.513
    thibodeau 0.388
    ptscreated 0.387
    gamescore 0.379
    pelton 0.365
    nbaeff 0.350
    si 0.299

    If it follows that the "rich man's versions" of these stats are better, but generally follow the same relative correlations as the metrics in the study, then PER is actually a really bad predictor, and SPM is arguably the best.

  18. themojojedi Says:

    Nick Anderson's entry is one to ponder. I guess a few steals goes a long way.

  19. Johnny Says:

    When I think of Nick Anderson during those finals I remember his complete meltdown at the free throw line which pretty much stuck with him the rest of his career.

    However, Penny Hardaway had a very good and underrated finals:

    25.5 ppg, 8.0 apg, 4.8 apg, 1.0 spg, 0.8 bpg, 35/70 from the field, 11/24 3pt, 21/23 FT

  20. Ricardo Says:

    "Isn't that like saying that if everyone in a race jogs lazily, the guy who comes first doesn't deserve the gold medal?"

    Find a better analogy. The point of participating in the NBA Finals is not to win the Finals MVP award - it's to win the damn championship.

    In fact, I'll give you a better analogy: Let's call this the 4x100 Relay Finals. Representing Boston, we have Garnett, Allen, Pierce, and Rondo; for Los Angeles we have Artest, Fischer, Gasol, and Bryant. And let's say that besides having a champion relay team, there is also a Relay Finals MVS award. Don't you think that, if we are to have a sprinter from the losing team win the MVS award, he ought to look like Secretariat against four tortoises out there?

    Kobe has had two good games - nothing on the level of earth-shattering. The Lakers are 1-1 in his two good games. Does this really look like an OBVIOUS MVP here?

  21. storyofgreats Says:

    Great post Neil.Kobe is doing everything he can.He owns the only block in Game 5!How others are so clueless and desireless especially Candyman LO is beyond my comprehension.They could not buy one stop during that ridiculuos strech.
    Cheers!

  22. Hk Says:

    #17, How does Win Shares compare?

    I've seen Rosenbaum's criticisms of Wins Produced (which you can see with win score, it is a great team barometer but not individual barometer). I'm just wondering if being able to predict team results is the best way to measure an individual player's efficiency? I have no doubt that SPM is a great metric (superior to WP), but is it better than Win Shares on an individual level? I would be curious to see your response Neil.

  23. koberulz Says:

    Ricardo: My point is that the best player is the best player, regardless of how poorly that player plays. If nobody plays better than Kobe, he's the best player and thus should win the MVP award, whether the Lakers win or lose.

  24. marc Says:

    What about Rondo?
    And for Kobe, he takes a large chunk of the available shots
    and he shots somewhere in the 40ies, while a lot of the lakers shoot better. Is he really the MVP if he still shoots that much then?
    Scoring dominates awards, thats why he will win.
    But somewhere, somehow didnt he make the series closer than it could have been?