Sometimes the answer is “I don’t know.”
We analyze the world to further our understanding. Sometimes, though, our observations of the world are insufficient to give us clear answers.
The NBA has just concluded its pre-season and one might wonder if we now have additional observations of the world to tell us something about the upcoming season.
The most games any team played in the pre-season was eight, which is a very small sample. Plus, each team played a collection of players who either will not play in the NBA this season or play very little. Finally, the games didn’t actually count. Okay, these games counted for the people who will not play in the NBA or who earned the right to sit on an NBA bench. But for the NBA regulars, these games were not quite as important as the regular season.
Still, a few teams and players did play well in the pre-season and certainly this has to mean something. Doesn’t it?
For fans of the Toronto Raptors and Golden State Warriors, two teams who only lost once in the pre-season, this collection of meaningless games might be providing a sense of hope for the upcoming season. That hope might be tempered, though, if we consider the top performers in the pre-season.
To identify the top pre-season performers I began with data from Doug Steele’s NBA and MLB Stats Home Page (a very convenient, although not perfectly accurate website). With data in hand, I then calculated Win Score for each player who played 100 minutes in the NBA pre-season. And because what matters is how a player performs per-minute relative to his position, I calculated each player’s Win Score per minute and then adjusted for position played by subtracting off the position averages (noted HERE). With each player evaluated, I then created a list of the top 15 NBA performers.
Dwight Howard tops the list. This is not too surprising. Howard offered a very high Win Score per minute last year, so it is not a shock that he would do well in the pre-season. Baron Davis and Kevin Garnett are listed two and three. Again, these players have a history of playing well in the NBA.
And then there is Paul Millsap. Unless you really are addicted to the NBA, you probably have never heard of Millsap. He was 47th player taken in the draft and Steele lists him as a small forward for the Utah Jazz. I am not sure he is a small forward, and if he is a power forward he would drop to 15th on the list. Still, Millsap has played well in the preseason which means…. okay, not sure what that means. To the extent this is important, Millsap’s performance does mean that I will offer more thoughts on him when I discuss the NBA rookies tomorrow.
Other players on the top 15 list are also surprising. Brian Cook, Kris Humphries, and Matt Barnes each had a negative adjusted Win Score last year, indicating these players were not very productive. Yet here they are among the top 15 performers in the NBA in the pre-season.
Except for the rookies we have two samples to consider in evaluating these players. First we have what they did in the pre-season. We also have what these players did in the regular season last year. Which one should we believe?
Obviously the sample from last year is more important. We can expect Howard, Davis, and Garnett to be good this year. This is not because they have played well in the pre-season, but because they have played well in the past when the games actually counted. Likewise, Cook, Humphries, and Barnes are going to need to be productive when the games count before we believe that they can truly be good NBA players. And given what each did in the past, it will be truly surprising if any of these players rank among the top 15 performers in the NBA regular season.
By the way, three of the top 15 preseason performances were turned in by the Golden State Warriors. Does this mean Don Nelson is going to turn this team around? If you are a Warriors fan the answer is yes. If you understand small samples, the answer is “I don’t know.” And this is sometimes the actual story the sample is telling.
– DJ
Our research on the NBA was summarized HERE.
Wins Produced and Win Score are Discussed in the Following Posts
Simple Models of Player Performance
What Wins Produced Says and What It Does Not Say
Wages of Wins Journal 
Erich
October 29, 2006
Perhaps another interesting angle to take on preseason stats would be pace.
Which teams have seemingly changed their pace from last season? Are the big changes limited to the teams with new coaches?
dberri
October 29, 2006
Erich,
I am not sure what I have is an interesting angle.
Seriously, what you suggest might be worthwhile. One would expect that changes in pace would be associated with new coaches. It would be interesting to see if new coaches can indeed put a stamp on a team that quickly.
Travis
October 29, 2006
Another stat that may be worth noting is whether or not there is an increase in slam-dunks from the pre-season to the regular season. While some players, like those trying to earn a spot on the bench, obviously would try and play their best; other players, like stars, may be tempted to do some show boating. I would think that if your LBJ or D-Wayde , it would be hard to play “smart” basketball in the forth quarter of a preseason game. Maybe, in the interest of the fans, the superstarts would treat a preseason game similar to an all-star game, and go out with sole purpose of entertaining the fans.
Franco
October 29, 2006
Have you considered using your information and trend analysis to develop a guide to fantasy basketball? Good fantasy basketball teams tend to win because they address the holes that you have isolated as the key to your ranking system. Don’t you think that this would be a good $money making$ opportunity? At least, will you send me a chart listing performance for all of last year’s players?
dberri
October 29, 2006
Franco,
A few quick thoughts…
Not sure what I do is really of much help to fantasy basketball players. If you are playing a league with casual fans, then I would advise ignoring scoring in evaluating players. Casual fans, I suspect, will overvalue scorers so that is where you will have the advantage. In leagues with serious players, though, this advantage is probably already exploited. So if you are in a serious league, I have no real advice, other than maybe you should broaden your circle of acquaintances to people who are not devoting their life to fantasy basketball (just a small joke).
About doing something just to make money… if you bothered to go get a Ph.D. and become a college professor then I imagine that money alone is not a motivator. After all if you are just about making money you could have earned an MBA and started making very serious money quite a bit quicker. Not that I am opposed to making money. But I prefer to make money doing something I like. So when people pay me to write about sports economics, I take the money. When people pay me to talk about sports and economics, I also take that money as well. But starting up a business of selling information to fantasy basketball players seems like quite a bit of work that I am not sure I would enjoy. So the money would not be much of a motivator.
Last thought… I do play fantasy basketball. And I did finish tied for first in my league last year. Looking at my team this year, though, I think I over applied my strategy. At least the players assigned via our automatic draft sure don’t look like winners to me.
Ed Devere
October 31, 2006
What is also curious, Jamal Sampson had the best per minute score both preseason (Denver) and last season (Sacramento), if you don’t set a minimum to minutes played (and don’t adjust to position).
dberri
October 31, 2006
Ed,
It is interesting that Sampson, in 506 career minutes, has posted a career Win Score per minute of 0.319. That is well above average. Per 48 minutes he grabs 18 rebounds. Perhaps he should be given more of a chance to play.
By the way, where did you find last season’s preseason numbers? I would not mind seeing these.
Ed Devere
October 31, 2006
i meant last season: last regular season, not last season’s preseason