Yesterday morning Brendan Nyhan asked me if I knew Sports Illustrated had once again cited The Wages of Wins. Since my copy of Sports Illustrated doesn’t arrive until Friday afternoon, I had to tell Brendan that I did not know this.
But by yesterday afternoon I learned that Brendan was indeed correct. In the afternoon I pedaled home (I live in California so I ride my bike to work every day) to take my daughters to gymnastics. And as I watched my girls do things that looked like they would hurt me, I got to read my copy of Sports Illustrated (which had arrived after lunch).
Before I got to where The Wages of Wins was mentioned, I first had to read Tom Verducci’s description of the Detroit Tigers trade for Miguel Cabrera and Dontrelle Willis. After reading the details on a deal that might allow the Tigers to contend for a title next year (baseball is hard to predict so that is a big “might”), I then turned to an article by Jack McCallum on Chris Paul and Deron Williams (which I later learned was posted on-line on Tuesday). Towards the end of this article is the following two paragraphs:
Indeed, one can collect several not-for-attribution votes for Williams, based on his superior shooting ability and size (the latter giving him an advantage on defense) and the 17-game playoff experience he got last season. A coach who observed both of them during the U.S. national team trials in Las Vegas over the last two summers — Paul participated in ’06, Williams in ’07 — says that Williams seems closer to stardom. “Chris was kind of like everybody’s little brother in camp,” says the coach. “Deron was willing to learn, particularly from Jason, but he never acted like he didn’t belong.”
On the other hand The Wages of Wins, a stats-based website, devoted a long recent post to comparing the two, concluding that Paul is the superior player. Paul’s assist-to-turnover ratio was 3.28 through Sunday, better than Williams’s (2.37), Kidd’s (2.60) and Nash’s (3.27); that doesn’t sound very loose-cannonish.
Thus far, Paul and Williams have been the subject of three columns at the WoW Journal, with the latest posted about a week ago.
None of these noted assist-to-turnover ratio, but one cannot expect McCallum to explain Win Score or Wins Produced in a magazine article. I made an effort in VIBE magazine (in an article that appeared in November), and I can tell you it’s very difficult to explain these metrics in a few words.
Of course, in this forum we can offer more words (and I have offered many). Consequently, we can tell stories with the Wages of Wins measures.
The strict comparison between Paul and Williams, though, has been made. In an effort to say something new, I wish to build upon a table in the McCallum article which compares Williams and Paul to every point guard taken with one of the top four picks in the draft since the lottery was instituted in 1985. McCallum’s table – which you can see here – looks at each guard’s per game production of points and assists per game in their third NBA season.
I thought this was an interesting comparison, but wished to see more. Table One reports what each of these guards did their second and third season with respect to all the box score statistics. And of course, the summary Win Score measure.
When we look at performance in year two, we see that Chris Paul’s Win Score per 48 minutes was 10.2. Of the thirteen point guards listed, only Steve Francis bested this mark (BTW, this was the best year of Stevie Franchise’s career). Deron Williams was above average with a mark of 6.8, but eight other point guards in this sample were more productive in their second season.
This season Paul has appeared in 21 games while Williams has played in 24 contests. When we look at what each of these players have offered thus far in 2007-08, we see that both guards have improved upon 2006-07. Paul has posted a Win Score per 48 minutes of 13.0, which clearly leads the field. Williams does not come close to Paul, but his per 48 minute mark of 8.6 places third on this list.
The key difference between Paul and Williams is what each player does with respect to the possession stats – rebounds, steals, and turnovers. If we calculate net possession [rebounds +steals – turnovers] (BTW, did I just make this stat up?), we see that Paul has a mark of 4.7 per 48 minutes. Kidd, Payton, and Francis actually did better with respect to net possessions, but Paul’s superior shooting efficiency and accumulation of assists trumps the small advantage these three have in net possessions.
Turning to Williams we see a net possession mark of 0.8. None of the point guards listed did worse than this in their third season. Despite such poor numbers with respect to rebounds, steals, and turnovers, Williams is able to post a relatively high Win Score because of a very high level of shooting efficiency and solid assist numbers.
Let me try and summarize what we have learned. Williams is a very good point guard. Relative to other high lottery picks, he has looked very good in his third season. Still, Paul appears to be from another planet. The only place where Paul is consistently below average is with respect to blocked shots. The average point guard blocks 0.3 shots per 48 minutes, or one shot every 184 minutes. So far this season Paul has only blocked one shot in 784 minutes.
So if your defense is predicated on your point guard blocking shots, Paul is not going to work for you. But if you want a point guard who can hit his shots, get assists, rebounds, and steals; then Paul is the choice. And if Paul continues to produce, before his career is over he will most definitely rank among the all-time greats at this position.
Our research on the NBA was summarized HERE.
Wins Produced, Win Score, and PAWSmin are also discussed in the following posts:
Finally, A Guide to Evaluating Models contains useful hints on how to interpret and evaluate statistical models.