Monday, April 26, 2010

Shaq and Stock

I feel confident in saying that at least 95% of basketball fans think that Shaquille O'Neal was a better basketball player than John Stockton. I think Shaquille O'Neal was a better player than John Stockton. But why do I think that? I never saw Stockton or O'Neal play night in and night out during their primes. Is it because of reputation? That doesn't seem like a good enough reason. What about the statistical case? If Shaq was definitely the superior player, he should have superior stats, right?

As you can guess, I didn't pick these players out of the blue. I was reading this post the other day [ed note-I wrote most of this about a month ago, let it sit, then finished it today], and I noticed that O'Neal and Stockton had identical WS%-the statistic that was being used to rank players. Win Shares are one of two popular "player rating systems", PER being the other. But why are player rating systems necessary? What's wrong with regular stats?

Shaq and Stockton are the perfect couple to illustrate the problem. For example, consider the question "which player created more points on offense?" The player who scores 25 points with 3 assists or the player who scores 15 points with 12 assists (per 36 career averages)? Well, if we assume that each assist led to a two point basket, Shaq is responsible for creating 31 points, and Stockton has created 39. But it's not that simple. On an assist, sometimes it is a spectacular pass that leads to an easy score and sometimes a pass leads to a contested jumper that the shooter knocks down. So, let's say that assists are worth less than two points. To be arbitrary, let's count an assist as "one point, assuming we have good reasons for doing so. Then, Shaq has created 28 points and Stockton has created 27. But wait! What about great passes leading to fouls? Which player had a greater % of their shots assisted? Not to mention rebounds, defense, etc.

Win Shares and PER purport to address a lot of these concerns by providing proper weightings, and for the sake of argument, I'm going to assume they do a good, if not perfect, job of resolving many of these problems. So how do they stack up? Before I get to those numbers, I want to bring up the issue of playing time.

Stockton played 19 seasons, Shaq is finishing up his 18th season. Stockton was more durable than Shaq, playing in 1504 games compared to Shaq's 1170. Shaq played more minutes a game, 35 instead of 32. At their peaks, Stockton played 39 minutes a game and Shaq played 40. Overall, Stockton played 47,764 minutes compared to Shaq's 41,166. Both players played 3,000 minutes 3 times, although Stockton played 2,800 minutes 10 times compared to 5 for Shaq.

Stockton leads Shaq in total Win Shares 208-179. Per 82 games, that works out to 12.6 WS for Shaq and 11.3 for Stockton. Per season, that is 10.0 for Shaq and 10.9 for Stockton. Per minute, as we've established, they are exactly the same. Which is preferable? I don't know.

Who was better at their best? Well, Shaq's best was 18.6 Win Shares, while Stockton's best was 15.6. On the other hand, Stockton finished in the top 10 in Win Shares 11 times, compared to 7 for Shaq. Shaq posted at least 13 Win Shares 6 times (probably would be 7 if not for the lockout), while Stockton did so 9 times. Which is preferable? I don't know.

What about PER? In that category, Shaq trumps Stockton 26.6 to 21.8. The advantage was 30.6 to 23.9 at their peaks. Based on this stats, Shaq was clearly the better player. So why don't I write the Win Shares results off as a fluke and tout them as evidence that Shaq was the better player? Because most of PER's known "shortcomings" are in areas that would probably benefit Stockton, like durability and defense (except for steals and blocks, I think). In addition, either PER systemically underrates point guards, or Win Shares overrates point guards. Paul, Magic, Oscar, Frazier, and Billups, the other top PGs by this method, all rank much higher in WS% than in PER. In addition, there are only three point guards in the PER top 50, suggesting that PER may undervalue point guards.

So much for statistics. On balance, the statistics seem to indicate that Shaq is the better player, given his scoring, rebounding, and PER, but there are good reasons to doubt that the statistics actually indicate that, including the uncertain relationship of points and assists, Win Shares, and PER's point guard issues.

But I suspect most fans don't think Shaq is a better player because 25 points and 11 rebounds is that much more impressive than 15 points and 12 assists. It is because Shaq was more dominating. At his best, Shaq could dominate the game like Wilt. We remember him overpowering other players en route to four titles. If he took nights off, or coasted during the regular season, or didn't do the little things when it didn't matter, it wasn't important. After all, he won an MVP, deserved at least one more, and won 4 titles. In Shaq's best season, 2000, he averaged 31 and 15 in the playoffs en route to a championship (although only shooting .456 from the line in those games, and I'm sure I don't have to mention the Portland series). In Stockton's best season, 1989, he averaged 27 points and 14 assists (4 steals, 2 blocks!, 4 turnovers, .601 TS%) while the Jazz were swept in the first round by the seventh seed Golden State Warriors, a series in which 4 of the 6 Jazz players who saw substantial minutes, Thurl Bailey, Bob Hansen, Mark Eaton, and Darrell Griffith, combined for negative Win Shares.

And yet, it is not hard to see how it could have been different, through no fault of Shaq's. If it wasn't for Dick Bavetta and Bennett Salvatore, Shaq would "only" have 2 championships, and Stockton's Jazz had the misfortune of running into Jordan's Bulls when they were at their strongest. And is it fair to blame Stockton for Hakeem Olajuwon destroying the Jazz? No more so than blaming Shaq for failing to contain Chauncey Billups.

I could go on talking about the lack of dominant centers in the early aughts, and the importance of the new hand check rules, but I'm sure you get the point. Why am I writing this? I think I have two reasons; the first being to test the limits of what "total value" box score statistics can tell us. We know that Shaq was better, but there's a (fairly strong) statistical case to be made that Stockton was his equal. On the other hand, I am writing this post as a challenge to our subjective memories. Of course we remember Shaq as the better player, but so many of those memories are based on contingencies outside of Shaq's control: if Portland makes their free throws, if the Lakers and Heat don't get those favorable calls, how is Shaq remembered? Is he remembered as he should be, the most dominating big man of the past dozen years, or is he remembered as the guy who dominated statistically, but missed his free throws, destroyed his team's chemistry, and got dominated by Hakeem? I don't know, but it's worth remembering that many of these subjective reflections are based on extraneous events, and that is where statistics can help us see the larger picture.

So, who is the better player? As a fan, I would say Shaq, as he was better at his peak, which I value over longevity, and yet remarkably durable. And yet, if I were a GM lucky enough to pick between Shaq and Stockton, knowing what I know now, I have to wonder, "Are the half dozen years I get from Shaq before he forces his way out of town really more valuable than the twenty I'll get from Stockton?" In that case, we have a situation in which Shaq is more valuable, but Stockton is more valuable to his organization, and I don't know how to quantify or qualify that.

No comments:

Post a Comment