A fantastic read from two of the sites friends, Arsine Khayoyan and David Dickey.
Basketball fans often groan when they open the NBA page to find that a sports writer’s ad hoc statistical gymnastics have the Jazz favored for the championship, or the Bucks with a higher power ranking than the Spurs.
It’s been said that “an unsophisticated forecaster uses statistics as a drunken man uses lamp-posts – for support rather than for illumination.” So as avid followers of the NBA, we decided to implement some of our skills as trained economists to our favorite team, the Los Angeles Lakers.
When looking at the recent statistical formulas concocted to analyze the game, it strikes us that often times one tries to make what they do appear far more complicated than it really is. This renders them ‘special’ and ‘different’ than everyone else, just like mom used to tell them. At the same time, simple averages, margins of victory etc. are hardly significant in any scientific manner.
Thus we decided to split the difference and do a statistical regression model and see which variables are significant on a scientific level. The variables used are pretty simple; they are comprised of the numbers seen in a box score after each game.
Specifically, we looked at…
• Number of field goals attempted
• Field Goal Percentage
• The difference between the amount of 3-point shots attempted by the Lakers and opponent
• The difference between the amount of free throws attempted by the Lakers and the opponent
• The difference between offensive rebounds
• The difference between defensive rebounds
• The difference between assists
• The difference between steals
• The difference between blocks
• The difference between turnovers
• The difference between the number of fouls assessed to each team
All in all, 11 variables were used for all 82 regular season games plus the 5 playoff games against Utah. These regression results give us better understanding than most of the other statistics used out there today. The most significant variable for a Lakers victory was the difference in turnovers. This range was -13 (Opponent had 13 more turnovers than the Lakers) to 10 (Lakers had 10 more turnovers than opponent). The probability for victory in this scenario is 99.99% and 21.56% respectively. When the amount of turnovers is even for both teams, the Lakers still have a 96% chance of winning, ceteris paribus.
The next most important variable was the difference in assists. Obviously, the more assists the team has, the more baskets are being made and the chances of winning are higher. When opponents have 10 assists or more than the Lakers, the probability of winning the game is 60% or less whereas if the Lakers have 10 or more assists than the opponent, the probability of winning is 99% or better. When assists are low and turnovers are high, the probability of winning is under one percent (regardless of how well the other stats may be). Likewise, when assists are high and turnovers low it’s very close to 100%.
Some of the other results were somewhat surprising. According to the model, the number of 3-point shots attempted or the number of offensive rebounds is irrelevant and has no significant effect on the outcome of a game. We would have assumed that as the number of 3-point shots attempted increased, less high percentage shots would be taken and it would have a significant affect on the outcome of the game (especially considering how much emphasis Coach Phil Jackson puts on it). However, apparently the number of 3-point shots doesn’t matter so long as they’re good shots.
On the rebounding side, offensive rebounds proved to be a far more insignificant stat than defensive rebounds. During the first round playoff series against Utah, many Lakers fans (us included) were screaming at the Lakers bigs to box out and bemoaning Boozer’s forearm shivers to Pau Gasol’s lower back. However, at the end of the day, as long as the Lakers have an advantage on the defensive rebounds, what happens with the offensive rebounds is not very relevant. When the Lakers have a double digit advantage on the defensive glass, the probability of victory is 99% whereas when they have a double digit deficit on the defensive glass, it’s only 64% or less.
Finally, free throws, free throws, free throws. Looking at differences of free throws is usually the common excuse most fans use as a reason for a loss. According to the results of the regression model, free throws attempts are only significant at the 90% level, not at the 95% level that the previously mentioned stats were. In laymen terms, getting to the free throw line is important, but not crucial for victory. For example, even when an opponent has 10 more free throw attempts, the Lakers still have an 80% chance of winning the game. Only when the Laker’s opponents shoot 17 additional freethrows, does the probability for winning go under 50%. This may be cause for some alarm should the Lakers meet up with LBJ in the Finals.
In conclusion fellow Lakers fans, be more upset at one-on-one play and turnovers than not getting a free throw from the ref. The stats for the 2008-2009 Lakers show that team ball, reducing the turnovers, and securing defensive rebounds are the main keys to victory.