The 2007 exhibition season has now started and fans are finally getting their first glimpse at the players taken in the 2007 NBA draft. And already many are drunk on the “potential” drug.
Although dreaming about potential is fun, a stroll down memory lane suggests that many of the players we think have “potential” today are going to be tomorrow’s journeymen and has-beens.
This past summer I made this point in reviewing the 1991, 1992, 1993, and 1994 NBA draft (you can see these reviews at the NBA Draft page) Today I want to continue this analysis with a look at the 1995 draft.
The 1995 Lottery
My approach in these studies has been to look at the overall career productivity of each player drafted. This analysis will be presented, but I wanted to start with a different perspective.
The draft is first and foremost about the teams that did not make the playoffs. Unless a trade was made, it is these teams that select in the NBA lottery. And it’s the lottery picks that are expected to help turn a non-playoff team around.
Back in 1995 there were 13 lottery picks. Of these, how many can we say truly helped the team who selected the player?
To answer this question, I looked at how productive each lottery pick was for the team choosing the athlete on draft night. For example, Joe Smith was taken with the first overall pick by the Golden State Warriors. After producing at a slightly above average rate his rookie season, Smith was decidedly below average in 1996-97. And after continuing his below average play at the onset of 1997-98 season, Smith was traded to the 76ers.
When we look at the two seasons Smith completed with the Warriors, we see a player who produced 7.8 victories and posted a WP48 [Wins Produced per 48 minutes] of 0.064. Given that average is 0.100, Smith was hardly what the Warriors hoped for on draft night.
The story gets somewhat better for the second pick. Antonio McDyess was selected by the LA Clippers, but traded to the Nuggets. After two seasons in Denver, McDyess posted a WP48 of 0.00. This happened because his Wins Produced in his rookie season of 2.6 was completely offset by a -2.6 offering in his sophomore campaign. Then McDyess went to Phoenix for one season, only to return to Denver for the next four years. The second time around, McDyess was a much better player. In all, McDyess played six seasons in Denver, producing 27.6 victories and a WP48 of 0.108.
I could go on with each draft pick, but that’s a bit tedious. So let’s just turn to Table One, which reports what each lottery pick did for the team choosing the player in 1995.
Table One: The Productivity of the 1995 Lottery Class
As Table One illustrates, only Kevin Garnett, Bryant Reeves, and Corliss Williamson completed more than three seasons with their original team. And only Reeves played his entire career in one spot.
On average, these players only lasted 3.3 seasons (and that number is clearly inflated by Garnett). And it seems that teams basically gave up on these players a bit too fast.
Looking at the Entire Draft Class
That can be seen when we see the career performance of all players selected in the first and second round of the 1995 draft.
Table Two: Looking Back at the 1995 Draft
Ideally I would have created a table with just the career performance of the lottery picks. But I didn’t. Still, let me save everyone the trouble of hunting through Table Two and note that the following lottery picks posted a career WP48 that was better than the mark posted for the team selecting the player: J. Smith, McDyess, Jerry Stackhouse, Rasheed Wallace, Shawn Respert, Ed O’Bannon, Kurt Thomas, Cherokee Parks, and Corliss Williamson. Yes, almost all of these players got better after they went elsewhere (although for many of these players the improvement was slight).
That being said, it’s not the case that many of these lottery picks ever became extremely good players. In fact, of these lottery picks, only Garnett, McDyess, Thomas, and Gary Trent have a career WP48 that is currently above the 0.100 mark. And only Garnett has been substantially above average.
When we look at the entire draft class, we do see a few other players who produced at an above average rate. This list includes Brent Barry, Michael Finley, Eric Snow, Bob Sura, Greg Ostertag (a subject of one of the first posts at the WoW Journal), Fred Hoiberg, Theo Ratliff, and Mario Bennett. Of these, only Barry (the 15th player selected) and Hoiberg (a second round pick I discussed HERE) were able to post a WP48 above 0.200. And if Barry repeats what he did last year, he will be only the second player from this draft class (the first is Garnett, of course) with more than 100 Wins Produced in his career.
A Familiar Story
In looking back at this draft we see a familiar story. Although all of these players were considered blessed with much potential in 1995, many never developed into great players. This story needs to be remembered when we look at the draft picks of 2007. Which players from this draft class will look like great players in 2019?
If great is defined as a career WP48 above 0.200, the answer is “three or less.” At least, when we look back at the drafts in 1991, 1992, 1993, 1994, and 1995 (which you can see at the NBA Draft page), we never see more than three players clear this mark. In fact, none of the players selected in 1993 have posted a career WP48 above 0.200.
Keep this in mind when you look at Kevin Durant, Al Horford, Mike Conley, Jeff Green, Yi Jianlian, Corey Brewer, etc… Right now these players look like they are all going to be “great” (at least to the people who invested a lottery choice in the player). But as time goes by, we will see that most of these players will fail to develop into the great players we envision today. And in fact, if the study of 1995 can be generalized (and I need to look at this for more years to reach any firm conclusions) if these players ever do develop it may happen for some other team. And this thought should be more than an antidote for the “potential” drug noted at the onset of this column.
- DJ
Our research on the NBA was summarized HERE.
The equation connecting wins to offensive/defensive efficiency is given HERE
Wins Produced and Win Score are discussed in the following posts
12 responses so far ↓
Alien Human Hybrid // October 15, 2007 at 9:18 am
Interesting and thought provoking article, as usual. I am really interested in your description of “great” though.
For example, when you think of all-time great players, generally these players are ones that have performed at an amazing level for the vast majority of their careers.
But what of those players who had a 2-6 year run of “greatness”? They may not be all-time greats, but for a period they may have been as productive as any at their position. This seems particularly relevant because when you consider the length of player contracts overlaid across a franchise’s window of championship opportunity – as most teams do not compete for a championship year after year.
Two players immediately come to mind Kirilenko and Penny Hardaway. They may not be all-time greats, but their productivity (or “greatness” if you will) should be understood within the context of their peak abilities, and not only viewed through the length of an entire career.
I’d love to see an all-time peak performer chart- something that lists the top 5 most productive, consecutive three years by a player, and perhaps a break out chart by position. This could be called “The All-Time Dominant Team”, or some similar moniker.
Perhaps that is too much work- but you know this is the kind of thing fans do all the time, so it would be both fascinating and endlessly useful to have your empirical stamp on it.
Thanks for thinking.
-A.H.H.
dberri // October 15, 2007 at 9:37 am
A.H.H.,
I thought of what you were saying last night as I wrote the post. But I didn’t feel like looking up each player’s peak performance. You are correct, though. If we look at maximum performance as opposed to average, we will see more player labeled as “great.” Most player’s, though, will still come up short.
Mark // October 15, 2007 at 4:14 pm
I have a question about “average” players. If I understand correctly, a WP48 has an average of 1 based on minutes on the court. Of course, one assumes that in general the guys on the court are better than the guys on the bench, so I’m inclined to think that a WP48 of 1.0 is actually pretty good.
Doug’s stats page lists 457 players last year. What is the average WP48 for this group? I’m guessing much less than 1.0.
Jason // October 15, 2007 at 6:54 pm
The average wp48 is not 1, but 0.1. This is based on the fact that there is one win in a game. There are 480 minutes available in a game (48 minutes times 10 players–5 for each time–who receive these minutes. If the win is divided up exactly equally for every player-minute on the court, 1/480= 0.0021 wins per minute which is equivalent to 0.100 wins per 48 minutes.
The average score for all players, unweighted is quite a bit below 0.100, though the average weighed per minute played does come out to be 0.1, or something rather close.
dbg // October 15, 2007 at 7:00 pm
Average is not 1 but 0.1, one tenth. A win is 1 and a loss is 0. There are 10 players on each team, so then 1(victory) is divided by 10(players) to give the average.
Actually I don’t have a clue because…I haven’t read the book.
dberri // October 15, 2007 at 7:36 pm
Jason,
Good explanation. Regardless of how you measure wins, it must be the case that wins per 48 minutes for a player is 0.100.
Similar to what Jason says, the average team wins .500 games per48 minutes. And since a team has 5 players, per 48 minutes each player must produce 0.100 wins.
dbg, you have it right also. But you need to read the book (I hear it is pretty good).
Sam Cohen // October 15, 2007 at 8:15 pm
Following up on Mark’s question and the responses, I guess I’m curious to know the median WP48 mark and the 150th best WP48 mark (i.e. the lowest WP48 mark that should start in the league assuming talent was divided perfectly evenly between teams). Is that information readily available? (I’m assuming position adjusted would be most useful, but might not be as available off-hand)
And I ordered my copy of the book last week, so hopefully I’ll be better informed in a few days.
dberri // October 15, 2007 at 8:28 pm
Sam, good question.
I looked at all players who played from 1991-92 to 2006-07. The median for all players is 0.053.
For players who played at least 2,000 minutes, the median is 0.125. For players who played at least 1,000 minutes, the median is 0.095.
Glad to hear you ordered the book. Hope you enjoy.
Jason // October 15, 2007 at 10:19 pm
That’s interesting, but to be expected. Players who play well are scarce but spend much time on the court. Players who don’t produce are quickly replaced, but usually by another lesser talent.
I think the way to state it isn’t that 0.100 is the average player, but 0.100 is the average player production. It must be by definition of what a win is.
dberri // October 15, 2007 at 10:42 pm
Jason,
That’s true. More precisely, 0.100 is weighted average player production. Average production depends on the sample you consider.
Getting Too Much to the Point in Memphis « The Wages of Wins Journal // December 3, 2007 at 9:30 am
[...] Conley can certainly take over for Stoudamire. Again, Stoudamire has been playing since 1995 (a draft I reviewed in October). In his career he had produced 56.9 wins with a 0.097 WP48. In sum, Stoudamire has been an [...]
T-Wolves Minus Garnett Equals the Worst NBA Team « The Wages of Wins Journal // December 10, 2007 at 12:18 am
[...] like Kurt Thomas or Gary Trent; or what if Joe Smith or Antonio McDyess fell to the T-Wolves [see the review of the 1995 draft for the career numbers of these players]? How much would this have impacted Minnesota’s [...]
Leave a Comment