July 11, 2007 4:22 PM Subscribe

How many points per game or per x minutes does the average NBA player score?

I want to know this so that I can create a basic over-under guestimate of how an incoming player will fare in the NBA, and say, for conversation's sake, "he will probably do better/worse than the average." I am guessing that there are probably quite a few ways to calculate this, but I am lazy and know nothing about what would make a valid statistic or what method would best achieve my purpose. I have searched the web and failed at finding anything so far.

I want to know this so that I can create a basic over-under guestimate of how an incoming player will fare in the NBA, and say, for conversation's sake, "he will probably do better/worse than the average." I am guessing that there are probably quite a few ways to calculate this, but I am lazy and know nothing about what would make a valid statistic or what method would best achieve my purpose. I have searched the web and failed at finding anything so far.

Take the total number of points scored by both teams in a game. Divide by 10 players. This is the average number of points scored by a player per 48 minutes played. Do this over a bunch of games.

posted by ldenneau at 4:39 PM on July 11, 2007

posted by ldenneau at 4:39 PM on July 11, 2007

I'm sure you could ask at the APBRmetrics forum, and get an answer in minutes. You also might wanna poke around at 82games.com, which is a pretty in-depth b-ball stats site.

posted by danblaker at 4:45 PM on July 11, 2007

posted by danblaker at 4:45 PM on July 11, 2007

Unless you're surrounded by people who don't follow basketball at all, saying that is almost meaningless, because as a statistic, it's not indicative of anything. Players are known for points, rebounds, assists, steals, blocks, being able to take the double team, 3-pointers, dunks, driving to the basket, good defense, 6th man/role players, and the list is endless...

And ultimately, one statistic averaging all players of all positions and all amounts of time played gives you a messy and almost useless number.

What I would do, however, is google the name of the player(s) and get a sense of what they do well and what they don't. Just a couple of bullet points will suffice, and that search would likely take you less time than any statistical analysis, plus make you seem slightly more knowledgable than you really are.

As an example, John Stockton of the Utah Jazz (one of the greatest players of all time (arguably)) only average 5.6 points per game his rookie season. But he was never known for being a scoring machine, and would never put up huge numbers consistently. And yet, he was among the league's best in steals and assists. Stockton to Malone was a classic line that existed through the 90s. So if you start a conversation by saying, "Well, Stockton should only be an average player because of his 'x points per game'", people are going to look at you like you're crazy.

posted by SeizeTheDay at 4:50 PM on July 11, 2007

I want to emphasize Idenneau's answer, 'cause it might get buried in minutiae. Average number of points per player per minute is trivially easy to find; take average number of points per game, divide by minutes in a game, divide by number of players on the court.

posted by Justinian at 4:55 PM on July 11, 2007

posted by Justinian at 4:55 PM on July 11, 2007

I'd like to point out that LDenneau's answer is a mean, and when it comes to this kind of thing the mean is nearly meaningless. (Ahem)

Player scoring in the NBA almost certainly falls on something much more like a power-law curve. So the majority of players would score less than the mean.

As always, if you want a single summary number, the median would tell you far more -- but that's not going to be anything like as easy to derive.

posted by Steven C. Den Beste at 5:25 PM on July 11, 2007

Player scoring in the NBA almost certainly falls on something much more like a power-law curve. So the majority of players would score less than the mean.

As always, if you want a single summary number, the median would tell you far more -- but that's not going to be anything like as easy to derive.

posted by Steven C. Den Beste at 5:25 PM on July 11, 2007

I want to emphasize SeizeTheDay's answer is the more useful one. As much as you want to think understanding a sport is as easy as that, it isn't. And the number of points a player scores does not in fact say very much in basketball.

posted by dame at 6:21 PM on July 11, 2007

posted by dame at 6:21 PM on July 11, 2007

I guess none of you play fantasy basketball . . .

Adding up all of the 2006-07 regular season averages for all 450* players who averaged more than 0.0 points per game** and then dividing by the number of players yields**8.3** as the mean.

As has been pointed out, though, this is not necessarily the most statistically significant number when it comes to what you're talking about, considering that a majority of players averaged less than 8.3 points per game.

The median (meaning half averaged more and half averaged fewer points per game), which is far more significant for an over/under scenario, was**6.6** points.

However, there is a lot more than just these number to measure success for a rookie in the NBA. Just going by either the median or the mean scoring alone won't really give you an indication of whether a player is doing well in the NBA -- a whole lot depends on factors like number of minutes played per game, what position he plays, whether he's starting or coming off the bench, what other contributions to the team he's making (statistical or otherwise), etc. This is to say nothing of the value considering where the player was drafted. A lottery pick had better be averaging a lot more than 7 points per game to be considered a success if he's a scorer and is playing starter's minutes. But an undrafted free agent coming off the bench for 10 minutes a game would likely be considered a success for averaging 5 points in those minutes, especially if he's contributing in other areas.

* This is, of course, more than the total number of active roster spots on all the NBA teams, but this data has every player who played during the season, which includes players who are signed to fill in for injured/dropped players.

(Semi-interesting sidenote: the lowest per-game average for a player who played all 82 games was Eric Snow's 4.2 ppg.)

** There were five players who played in at least one game but failed to average more than 0 points, but since the total number of games among them was only 21, I left them out of the calculations.

posted by camcgee at 6:48 PM on July 11, 2007

Adding up all of the 2006-07 regular season averages for all 450* players who averaged more than 0.0 points per game** and then dividing by the number of players yields

As has been pointed out, though, this is not necessarily the most statistically significant number when it comes to what you're talking about, considering that a majority of players averaged less than 8.3 points per game.

The median (meaning half averaged more and half averaged fewer points per game), which is far more significant for an over/under scenario, was

However, there is a lot more than just these number to measure success for a rookie in the NBA. Just going by either the median or the mean scoring alone won't really give you an indication of whether a player is doing well in the NBA -- a whole lot depends on factors like number of minutes played per game, what position he plays, whether he's starting or coming off the bench, what other contributions to the team he's making (statistical or otherwise), etc. This is to say nothing of the value considering where the player was drafted. A lottery pick had better be averaging a lot more than 7 points per game to be considered a success if he's a scorer and is playing starter's minutes. But an undrafted free agent coming off the bench for 10 minutes a game would likely be considered a success for averaging 5 points in those minutes, especially if he's contributing in other areas.

* This is, of course, more than the total number of active roster spots on all the NBA teams, but this data has every player who played during the season, which includes players who are signed to fill in for injured/dropped players.

(Semi-interesting sidenote: the lowest per-game average for a player who played all 82 games was Eric Snow's 4.2 ppg.)

** There were five players who played in at least one game but failed to average more than 0 points, but since the total number of games among them was only 21, I left them out of the calculations.

posted by camcgee at 6:48 PM on July 11, 2007

for benchmarking purposes, kobe bryant lead the league in scoring this past season, and he scored an average of 31.6 points per game while playing 40.8 minutes, .774 points per minue.

posted by andifsohow at 2:57 AM on July 12, 2007

posted by andifsohow at 2:57 AM on July 12, 2007

I did an undergraduate senior thesis on statistical analysis in basketball.

The best all-round metric for measuring offensive success is John Hollinger's PER. This metric is weighted such that an average NBA player gets 15; stars score above 20 and players in danger of slipping out of the league score less than 10. It considers every important offensive category, from true shooting percentage to low turnover ratio to shot creation ability to passing.

This metric doesn't consider defense, so the ability of someone like Bruce Bowen, who routinely scores below 10 despite starting for the reigning NBA champion Spurs, is not properly evaluated. The statistics just aren't there for adequate analysis of defense; it's the next big project for people like Hollinger.

posted by Kwine at 8:11 AM on July 12, 2007 [1 favorite]

The best all-round metric for measuring offensive success is John Hollinger's PER. This metric is weighted such that an average NBA player gets 15; stars score above 20 and players in danger of slipping out of the league score less than 10. It considers every important offensive category, from true shooting percentage to low turnover ratio to shot creation ability to passing.

This metric doesn't consider defense, so the ability of someone like Bruce Bowen, who routinely scores below 10 despite starting for the reigning NBA champion Spurs, is not properly evaluated. The statistics just aren't there for adequate analysis of defense; it's the next big project for people like Hollinger.

posted by Kwine at 8:11 AM on July 12, 2007 [1 favorite]

This thread is closed to new comments.

posted by rancidchickn at 4:30 PM on July 11, 2007