Little White Statistics
Or, “Stats are like bikinis - nice to look at, but don’t tell you the whole story.” -Brent Barry

Alright, let’s wrap this puppy up. It’s been way too long. This post will summarize everything we’ve done, so if you don’t want to read the pages and pages I’ve written, you won’t miss the conclusions: just the reasoning.

Why were we doing this all again? Oh yeah, because a few months ago, Christopher Reina made the claim that the Lakers are a better team when Kobe isn’t a “High Volume Shooter”. His reasoning was simple: the Lakers were 26-18 when Kobe took 20 or more, but 31-7 when he took 19 or less. Pretty straight-forward, right?

The fellas at Ball Don’t Lie subsequently slammed the analysis for a variety of reasons - mostly, not considering the game situation, and for drawing such an arbitrary cut-off. Their analysis, however, wasn’t backed by any numbers, so it was really a matter of one blog’s opinion against another.

So, what we’ve been trying to do is put better numbers behind both of those blogs’ conclusions. And once we did, it became pretty clear that Ball Don’t Lie is right, while Christopher Reina was wrong (sorry, Chris).

First of all, it’s true that Kobe averages more shots in losses than wins - four more shots per loss than per win. It’s also true, however, that Kobe averages more minutes in losses then wins - by about 4 minutes. With his shooting rate (a bit under 22 shots/game overall), that difference automatically accounts for about 2 shots per game, meaning that - balanced for minutes - Kobe averages closer to 21 shots per game in losses. That pretty much immediately throws out the 20-shot cut-off used by Reina. But, that’s still more shots per game in losses then wins.

Then we considered the fact that blowouts are special instances in which Kobe’s statistics are drastically off. 18 times during the season, the Lakers blew out their opponents by more than 12 points while Kobe played below-average minutes. Only one time were the Lakers blown out in similar fashion with Kobe playing below-average minutes. Averaging without those games, we discover the same thing we discovered above: that Kobe only averages 2 shots less per win than per loss.

But why 2 shots more per loss? Well, we then proceeded to go way too in-depth into individual games to discover when Kobe’s shot attempts increased, and we discovered something interesting. Kobe’s shot attempts and the Lakers’ corresponding winning percentage fell quite plainly into three categories. When Kobe takes less than 17 shots, the Lakers were 9-1. When Kobe took 18 to 28 shots, the Lakers were 28-19. When Kobe took 29 or more, the Lakers were 2-4. Note that these aren’t arbitrary cut-offs like the 20-shot cut-off, but rather they’re cut-offs where the change in the Lakers’ fortunes is most evident. And note that these divisions are not including that special case, blowouts where Kobe plays below-average minutes.

With those categories in mind, there are only two options: either (a) Kobe’s increasing shot attempts are causing the Lakers’ fortunes to change, or (b) some third variable is causing both Kobe’s shot attempts to increase and the Lakers’ fortunes to change. If (b) is true, there should be characteristics that group together all the games in each category.

And, upon examining the games quite closely, such characteristics do emerge. In the games where Kobe takes 17 or less shots, the Lakers are nearly never even challenged. They get an early lead and never relinquish it. Several of these games are blowouts where Kobe still plays his average number of minutes, but the important thing is that the Lakers are really never challenged.

In the games where Kobe takes 18 to 28 shots, the Lakers are almost always challenged. Sometimes they jump out to an early lead but have their opponent come back. Sometimes they fall behind early and comeback. Sometimes the game is close before the Lakers (or, rarely, the opponents) pull away. And sometimes the game is just close throughout. But the point is, all these games are united by one fact: the Lakers are challenged.

And within those games, we discover something very interesting. There are definite particular instances when Kobe’s shot attempts increase. More often than not, it’s in the second half, sparking either a comeback or a run to seal the game. But in a general sense, it’s when the Lakers are most challenged and most at risk for letting the game slip away. In these instances, Kobe becomes more aggressive and shoots more - and often succeeds in shooting the Lakers back into the game.

And in that final category, where Kobe shoots 29 shots or more and the Lakers are 2-4, we observe an extremely clear pattern: Kobe is the only offensive weapon. Gasol played in none of these games, and Bynum played in only one. And, like in the previous categories, these are all games where the Lakers were at risk of losing.

So, we’ve noticed a pattern, an answer to the (b) from a few paragraphs up. The “third variable” is how much the Lakers are challenged. In games where they aren’t challenged, they overwhelmingly win - an obvious conclusion, since they aren’t challenged. In games where they are challenged, they don’t win as often. This isn’t a revolutionary concept: they’re more likely to lose games where their competitor actually competes and plays well. That’s quite simple. And in games where they’re challenged and don’t have two of their top three offensive weapons, they lose even more often. Again, not a revolutionary concept.

This third variable is also what causes Kobe’s shot attempts to increase. As the best player on his team, the primary offensive weapon, and the ultimate competitor, he will tend to see more shots when his team is threatened. In games where they aren’t, there isn’t the need for him to be as aggressive. In games where they are, his aggressiveness and shot attempts increase. And in games where they’re threatened and there are no other options, his shot attempts increase even more.

The problem with the initial study conducted by Reina was misplaced causation: he suggested that Kobe’s high-volume shooting caused the Lakers’ fortunes to drop, but in reality it was the risk of the Lakers’ fortunes dropping that caused Kobe’s shooting volume to increase. Had Kobe restricted himself to the same shot counts in those games where the Lakers were challenged, it is reasonable to assume the Lakers would have lost more because oftentimes, it was Kobe’s shooting that sparked Lakers comebacks. His higher-volume shooting didn’t cause the Lakers’ fortunes to drop - it caused the Lakers’ already-dropping fortunes to not drop as far.

LITTLE WHITE TAKEAWAYS

The Lakers are not better off when Kobe doesn’t shoot as much, despite him averaging fewer shots in wins than losses. Instead, The Lakers depend on Kobe to lift them over their opponents in competitive games. Therefore, Kobe’s shot attempts increase in competitive games compared to non-competitive games. Naturally, the Lakers also are more likely to lose competitive games than non-competitive games.

The proof for this can be observed in a variety of ways. First of all, in most of the games where Kobe’s shot attempts are elevated, the Lakers are challenged at some point; on the other hand, in most of the games where his attempts are not elevated, the Lakers are never challenged.

What’s more, the actual increase in his shot attempts is observed typically in the second half, and typically sparks either a Lakers comeback (if the opponent leads) or a Lakers run to seal a win (if the game was close). Because these increases are typically observed in the second half, it is completely unreasonable to say that the increased attempts caused the Lakers’ misfortunes in the first place.

In short, the Lakers are more likely to lose competitive games, and Kobe shoots more in competitive games to try to prevent that. The competitiveness of the game causes both the Lakers’ lower winning percentage and Kobe’s increased shot count. Kobe’s shot count does not cause the Lakers’ lower winning percentage in those games.

Apologies for the delay - I started graduate school this past month, and I’ve had precious little time to write. But things are started to calm down, so hopefully we’ll be able to wrap up these two studies (Kobe Bryant’s “High-Volume Shooting” and the Box Score Analysis) before the season started in a couple weeks, and then we’ll try to introduce a weekly or bi-weekly feature.

But for now, it’s time to wrap this up. We’ve discussed how Kobe Bryant doesn’t play as much traditionally in Lakers’ blowouts, thus lowering his shot attempts. We’ve discussed how, even in blowouts where Kobe plays just as much, his shot total is lower. And we’ve discussed how, time and time again, Kobe’s shot attempts go up when his team is threatened: either in the event of an opposing team’s comeback, or when the Lakers trail and are running out of time to start a comeback, or in the waning seconds of a close game.

But now let us consider one more type of game: games where Kobe took drastically more shots than his season average. Six times during the season, Kobe took 30 or more shots in a single game - 30 twice and 32, 33, 37 and 44 once each. In these games, the Lakers were 2-4 - winning in Kobe’s 30- and 44-shot games and losing in the others. So, what prompted Kobe to take so many shots in these games?

Analyzing these games in the detail we’ve used previously is a bit silly considering that Kobe’s shot attempts are pretty high throughout the game, so instead let’s look at the overall storyline of each game and try to figure out subjectively what prompted his high shot total.

• March 23rd: Golden State 115, Lakers 111: Kobe goes 13/30 from the field, 3/9 from 3 and 7/7 from the line for 36 points, along with 14 rebounds and 8 assists. In this game, the Lakers play with a very short rotation (8 players play, no Bynum and no Gasol), and only half of those 8 players actually play well (6/19 for Odom though he does bring in 22 rebounds, 3/8 for Radmanovic, 0/5 for Vujacic, 3/10 for Farmar). The Lakers fall behind in the second quarter and trail by as much as 26 before mounting a furious and almost-successful comeback. 18 of Kobe’s 30 FGAs come during this attempted comeback - his first-half shot total (when the Lakers fell behind) matched his season average.
• March 24th: Lakers 123, Golden State 119: Second verse, same as the first. The Lakers play the same team the next day, fall behind around halftime (though not by as much, maximum of 13 points), and comeback. Same short rotation, and still no Gasol or Bynum. And this time, they win in overtime. And, almost identically to the game the day before, Kobe’s shot total while the Lakers fall behind is only slightly above his season average (12 shots for the first half), while his shot total is higher while the Lakers come back and win (18 shots).
• October 30th: Houston 95, Lakers 93: First game of the season, and still follows the same framework of these first two. Lakers fall behind (in the 3rd and early 4th this time, by a maximum of 13) and come back later. At first glance it doesn’t appear that his shot attempts go up significantly during the late comeback (9 shots in the comeback which spans the last 10 minutes, 32 for the game, so only a moderate increase), but if one counts the times Kobe is fouled in the act of shooting as shot attempts (which, in my opinion, they really should), that total shoots up to 15, while his total for the game (when counting the entire game this way) shoots up to 45 - meaning a third of his shots take place during a fifth of the game, when the Lakers come back. (in retrospect, I should’ve calculated ‘true FGA’ for every game we’ve analyzed, but we’ve considered it already when relevant)
• March 16th: Houston 104, Lakers 92: Two against the Warriors, now two against the Rockets. This game was the final win in Houston’s remarkable 22-game winning streak. You might also know that it falls within that same date range as those earlier Warriors games, which means that Gasol and Bynum were on the sidelines. Like the other three games, this game saw the Lakers fall behind (max of 15) and then come back, although in this instance the Rockets did pull away at the end. In this particular case, his shots actually fell evenly across the two halves (17 in the first, 16 in the second). However, 13 of those first-half shots came while the Lakers were still within 4 points - it wasn’t until Kobe’s shots dropped off in the first half that the Rockets pulled away.
• March 28th: Memphis 114, Lakers 111: You probably saw this coming - yet another game in that group of games where the Lakers were without Gasol and Bynum. The rest of the Lakers should’ve sat the came out too: of the 10 players who played, only two shot at or above 45%: Bryant and Mbenga, who was 2/3. The rest of the team combined to shoot a dismal 20 for 64 for 31%. Bryant’s 19/37 shooting elevated the team’s number substantially. Kobe’s shot distribution was fairly even for the game: 12 shots in the 1st, 10 in the 2nd, 9 in the 3rd and 6 in the 4th. The drop-off in the 4th quarter can be attributed mostly to Memphis finally realizing that Kobe was the only one who showed up to play and devoting even more defensive attention to him, as the 4th quarter is littered with shots that would’ve been Kobe assists if his teammates had actually hit the shots.
• January 14th: Lakers 123, Seattle 121: Lakers fans know this game as the game after Bynum went down with his season-ending injury - so, like the previous games, this was one of the Lakers’ games with no Bynum or Gasol. Vujacic and Radmanovic sat out as well, although the rest of the Lakers (excepting Odom’s 3/15 performance and Walton’s 1/6 debacle) played quite well, with Brown, Fisher, Ariza, Farmar, Turiaf and Crittenton - yes, Crittenton - each shooting at or above a 50% clip (although none attempted more than 6 shots). Despite their play, the Lakers still fell behind by 10 early in the 4th quarter. As for Kobe, he shot at a high clip throughout the game, (11 FGA in the 1st, 7 in the 2nd, 12 in the 3rd), but as usual, a disproportionately high number of his shots (12 over 10 minutes) came during the Lakers’ comeback in the final 5 minutes of regulation and overtime. That’s 27% of his shots being contained in 18% of the game.

So, we uncovered a nice little pattern amongst these games. First of all, five of the six games (excepting Houston on October 30th) are played without Gasol or Bynum, while none of the games feature Gasol at all (two before the trade, four during his late-season ankle injury). Five of the six (excepting Memphis) games also saw the Lakers behind by double-digits in the second half, a very rare occurrence for the season, though commonplace amongst this sampling of games. And, five of the six games (excepting Memphis) also saw Kobe’s shot attempts increase during the Lakers’ comebacks.

That’s a pretty clear correlation: Kobe’s six highest FGA games come when Gasol and Bynum don’t play, and when the Lakers are threatened.

LITTLE WHITE TAKEAWAYS


For the six games where Kobe takes drastically more shots than his season average (30 shots or more while he averages 21.44 in non-blowouts), there are three very clear criteria that each take place in at least five of the six games: first, the Lakers are playing without Gasol or Bynum; second, the Lakers fall behind by double digits at some point during the second half; and third, the Lakers come back - in every single game - to make it competitive, while Kobe’s shot attempts increase during the comeback.

It’s a very match between these criteria and the games. The lack of the Lakers’ other offensive weapons will obviously lead Kobe to shoot more, and that the Lakers are threatened at some point has been shown in our study to lead to an increase in Kobe’s shot attempts. And, as we’ve seen, when Kobe’s shot attempts go up, the Lakers come back.

Considering how large this analysis has been, I’ll be posting a wrap-up summary post in the next couple days to re-state our conclusions. And it won’t be another month and a half this time, I swear.