The recent Basketball Hall of Fame induction ceremony featured two of the great centers of the 1990s, Patrick Ewing and Hakeem Olajuwon. With a third, David Robinson, sure to join them when he becomes eligible next year, it seemed like an ideal team to consider the question of who has been the top center of the past quarter-century, a discussion that in addition to those three also prominently features the last great '90s center still playing, Shaquille O'Neal.
Now, if this idea sounds unoriginal, that's probably because while I was pulling together the numbers I needed, Dave Berri looked at precisely the some question on the Wages of Wins Journal. Since my numbers are different than Dave's and I approached the question a little differently, I think there's value in the contrasting perspectives.
My reference point for evaluating the four great centers is The New Bill James Historical Baseball Abstract, in which the godfather of sabermetrics used his Win Shares rating system to rank the top 100 players of all time as well as the top 100 players at each position.
James uses six factors to determine his final rating:
- Career Win Shares total
- Win Shares in three best seasons
- Win Shares in five best consecutive seasons
- Career Win Shares per season
- Timeline adjustment
- The subjective element
For our purposes, two of these six factors are essentially irrelevant. What kind of timeline adjustment is appropriate for the NBA, if any, is a debate all its own. Since all four players in question were roughly contemporaries, it's meaningless in this case. I also chose not to use a per-season rating because by accounting for replacement level with WARP in a way Win Shares does not, I don't think it's necessary to discount lengthy careers. I'll go ahead and report the numbers anyway, should you disagree with me on that point.
Let's take the remaining factors individually.
Career WARP Total
Here are the year-by-year totals:
Player 8485 8586 8687 8788 8889 8990 9091 9192 9293 9394 9495 9596
Ewing 5.3 7.2 14.0 15.9 20.0 16.7 17.7 13.4 17.2 14.0 11.8
Olajuwon 15.8 16.3 19.0 17.7 22.1 21.8 14.2 18.1 25.5 22.8 19.0 17.5
O'Neal 19.3 24.9 21.4 12.3
Robinson 23.3 24.8 20.3 22.3 28.0 24.8 25.9
Player 9697 9798 9899 9900 0001 0102 0203 0304 0405 0506 0607 0708
Ewing 14.3 4.5 8.5 5.4 1.1 1.1
Olajuwon 14.3 8.4 16.0 2.9 7.5 3.1
O'Neal 13.9 15.8 21.9 26.5 23.5 18.8 19.0 15.8 16.8 10.1 4.8 4.5
Robinson 1.2 19.7 18.8 16.8 15.7 12.2 7.4
I can't remember the last time I've considered it, but O'Neal's early Lakers career was something of a disappointment. He fought a variety of injuries and missed an average of 27 games over a three-year stretch that included his last season with Orlando. At the same time, as is much more documented, the Lakers were considered underachivers until Phil Jackson's arrival. Hakeem Olajuwon had his own down stretch where he was all but traded by the Rockets before recommitting to Houston and becoming an MVP and a champion.
Despite being the best and most experienced NCAA player of the group, Ewing took longer to adjust to the NBA, a process that wasn't aided by a knee injury during his rookie season. When we consider that Robinson was already 24 during his rookie campaign after two years in the Navy, O'Neal's rookie performance--at age 21--has to be considered the most impressive debut.
One technical note: I projected totals from the 1998-99 lockout-shortened season to 82 games. With that, the totals:
Though O'Neal has an outside chance of catching him with a resurgent year or two, Olajuwon is the career leader, with Robinson not far behind. This is the first sign that while Ewing was a great player and a sure Hall of Famer, he may be outmatched against this competition.
WARP in Three Best Seasons
Ewing: 20.0 (1989-90), 17.7 (1991-92), 17.2 (1993-94)
Olajuwon: 25.5 (1992-93), 22.8 (1993-94), 22.1 (1988-89)
O'Neal: 26.5 (1999-00), 24.9 (1993-94), 23.5 (2000-01)
Robinson: 28.0 (1993-94), 25.9 (1995-96), 24.8 (1994-95/1990-91)
The four elite centers won just three MVP awards, one apiece for Olajuwon, O'Neal and Robinson. From the perspective of WARP, they were robbed. One of the three rates as the most valuable player in the league eight times in the 11-year span from 1990-91 through 2000-01. Robinson earned four WARP MVPs, O'Neal three and Olajuwon one.
It's hard to argue against the MVPs won by Michael Jordan over Robinson in 1990-01 and 1995-96. However, O'Neal pretty clearly should have won more than once in the late '90s and early '00s when he was the league's dominant player. Allen Iverson had a great season for a great team and is a wonderful, unique player, but O'Neal was the most valuable player in 2000-01, and it wasn't particularly close.
The funny note is that neither Olajuwon won their actual MVP in their best-rated season, instead winning the following year. In fact, Robinson's MVP campaign is tied for his third-best in terms of WARP.
Ewing's best season would rank as the worst of the first seven years of Robinson's career. It's safe to say he lags again in this category.
WARP in Five Best Consecutive Seasons
Player Total Years
Robinson 121.2 1992-96
O'Neal 109.7 1999-03
Olajuwon 102.9 1992-96
Ewing 85.0 1990-94
Here, Robinson's consistent brilliance really shines through. He rated as worth at least 20 Wins Above Replacement Player all five seasons and seven straight overall. O'Neal's longest stretch of 20+ WARP seasons was three years; Olajuwon never did it more than twice consecutively.
Career WARP per Season
Along the lines of what James did, I divided each player's career WARP total by games played and multiplied by 82 to get an "average" 82-game season's value.
By this measure, Olajuwon, who played nearly 200 more games than O'Neal and about 250 more than Robinson, really gets knocked downward. However, as I said earlier, I think we've already largely accounted for this by focusing on peak value. Olajuwon's later seasons didn't really add much to his total.
And...here's where it gets interesting. We can add the three numbers-based categories together to rank the players, with some adjustments, to put them on the same scale. Naturally, I divided the best three seasons by three to get the average and the best five-year stretch by five. I decided to divide the career WARP totals by 10 because that ends up producing numbers similar to the best single-season values.
Strictly by the numbers, here's how the four centers rank:
Player Total Career Best3 5year
Robinson 76.6 26.1 26.2 24.2
O'Neal 73.8 26.9 25.0 21.9
Olajuwon 72.3 28.2 23.5 20.6
Ewing 54.1 18.8 18.3 17.0
James lists seven subcategories under the subjective factor:
- Statistically undocumented portions of a player's career.
- Inequalities in the caliber of competition.
- World Series performance.
- Positive or negative leadership.
- Clutch performance.
- Special contributions of the player undefined by the statistics.
- Defensive value beyond that accounted for in the rating system.
For the most part, I don't think these are of particular importance. James might argue that Robinson deserves credit for the two seasons he spent in the Navy prior to joining the NBA; players who were sent off to war in the '50s are credited by James for their missing seasons. Then again, Robinson's voluntary peacetime enlistment isn't exactly analogous to being drafted during the midst of World War II.
In terms of leadership, all four players have had good moments and bad ones. O'Neal could be dinged for his feuding with Kobe Bryant in Los Angeles. Robinson, as upstanding a human being as you'll find in the NBA in most regards, found his soft touch ineffective when confronted by Dennis Rodman's antics. Olajuwon had the aforementioned rough stretch in the middle of his career, and Ewing's absence ultimately gave name to a theory that teams play better without their overrated stars. I'm going to leave this area alone; feel free to disagree.
All four players were superior defenders, as is reflected in the statistics. The one player I feel could be slightly misrated is O'Neal late in his career, when he was vulnerable defensively against the pick-and-roll because of his limited mobility, something that can't be readily picked up from the defensive numbers tracked at the individual level. I'll penalize him 0.5 points, the equivalent of five WARP over the course of his career.
This leaves us with the playoffs, and a problem. James could largely write off the postseason because, for much of baseball's history, it consisted of seven games between two teams. Making it to the World Series was as important as how a player performed once there. Not so in the NBA playoffs. O'Neal has played in 203 postseason games, the equivalent of two-and-a-half full seasons. Even Robinson, the low man in the group with 123 career playoff games, has about 1/8 as many playoff games as regular-season games.
Playoff performance counts in the NBA--a lot. Unfortunately, one area where APBRmetricians have been deficient is not devoting enough time to studying playoff value. This column aside, rating careers isn't usually my interest, so I've spent virtually no time on the playoffs.
It doesn't take advanced analysis to realize this: David Robinson struggled in the playoffs, while Hakeem Olajuwon picked up his game. Olajuwon averaged 25.9 points and 11.2 boards with a 56.9 percent True Shooting Percentage in postseason games, while Robinson averaged 18.1 points and 10.6 rebounds, his TS% slipping to 54.8 percent. What I remember from the Sonics facing the Rockets most was that Olajuwon was indefatigable in May and June, averaging a remarkable 39.6 minutes per game. Those five to eight minutes a night he spent on the bench seemed so crucial to try to take a lead. (By contrast, Robinson averaged 34.3 mpg in the playoffs.)
One of the most comprehensive studies of playoff performance was done by John Hollinger in the Pro Basketball Prospectus 2003-04. In it, Hollinger uses Game Score per 40 minutes, less sophisticated than his usual PER, to assess how much players improved or declined in the postseason, finding that the average player loses 8.6 percent of their per-minute productivity because of the increased level of competition. How did the four centers in question fare?
Player Pl GS/40 Change
O'Neal 18.05 - 5.11%
Olajuwon 16.96 + 5.78%
Robinson 14.86 -13.90%
Ewing 11.82 -14.10%
Because this was a mid-career snapshot for O'Neal, while the other players were already finished, he comes out looking better in raw terms. Presumably, weaker postseason runs the last few years--even including a title in Miami--have dragged him back to the pack. For Olajuwon to have improved his per-minute rating nearly six percent in the playoffs, given the average 8.6 percent decline, is pretty remarkable. Meanwhile, the numbers confirm a slide for Robinson, as well as Ewing.
If we take Hollinger's postseason rates, apply them to each player's career regular-season WARP rate per 82 games and multiply them by career playoff games and adjust for minutes per game to estimate postseason WARP, we find the big winner is in fact O'Neal.
Player Pl WARP
Even though Olajuwon picked up his game dramatically in the playoffs, it still wasn't quite enough to make up the per-game difference with O'Neal. When we consider O'Neal's enormous advantage in career games played, he was the most valuable center of the group in the playoffs, which seems reasonable given his four NBA championships and six Finals appearances.
If we weight the estimates of playoff WARP at the same rate as career regular-season WARP, we can come up with a final ranking. Remember that O'Neal is also carrying the 0.5 penalty for his defense.
Player Total Career Best3 5year Play
Robinson 79.3 26.1 26.2 24.2 2.8
O'Neal 78.6 26.9 25.0 21.9 5.3
Olajuwon 76.1 28.2 23.5 20.6 3.9
Ewing 56.2 18.8 18.3 17.0 2.1
While Robinson is the leader, O'Neal could easily catch him. It would require about seven WARP the rest of his career; O'Neal rated as worth 4.5 last season and 4.8 the year before. Olajuwon is close enough in third that backers who want to put more weight on his career and postseason value could certainly make an argument in favor of the other center from the Lone Star state, though it's much more difficult to argue for Olajuwon vis-à-vis O'Neal. Despite all that, my ultimate conclusion is that Robinson has been the best of the group of elite centers in the NBA the last quarter-century.
Kevin Pelton is an author of Basketball Prospectus.
You can contact Kevin by clicking here or click here to see Kevin's other articles.