How to Use Critical Thinking with Sports Statistics

Uzu Ometu wrote an article entitled, Is Stephon Marbury Headed for a Second Straight Year in Dipped Statistics? On the whole, the article is somewhat interesting but it contains a fatal flaw. Mr. Ometu uses two common tricks with statistics. He uses multiple endpoints and non-equal comparisons to make a player, in this case Stephon Marbury, appear better than he really is.

Everyone loves to use numbers to prove their point. In the article, Mr. Ometu says, "But the numbers don't lie. Marbury is the only player since Oscar Robertson to average 20 points and 8 assists for his career."

There's a famous saying that goes, "figures don't lie, but liars can figure." Let's start by making it clear that Mr. Ometu is not lying and to the best of my knowledge he is not a liar. His statement is true. However, the comparison makes it seem like Marbury is one of the best players around, because he shares a bond with Robertson, one of the greatest players in basketball history. But are the two players really that comparable, or is this just an illusion, or as the quote above might say, a lie?

Multiple endpoints is the use of one of many valid reference points to look like a single reference point. Here, Mr. Ometu uses 20 points per game and eight assists. But there is nothing magical about those numbers. What if we change the numbers to 19 points and nine assists? Then the only two players in NBA history to accomplish the feat are Robertson and Isiah Thomas. What if we use 27 points and six assists? Then the only two players are Allen Iverson and Jerry West. We can play this game all night long.

Fans and sportswriters use the multiple endpoints game to make a case supporting their favorite player. And that argument can sound mighty impressive on the surface if you don't think too much about it. But I've seen both Marbury and Robertson play, and let me tell you they don't belong in the same sentence.

Now, I only saw Robertson at the end of his career. But Marbury in his prime couldn't hold the jock of the Robertson I saw play in the 1970s. Robertson was a first-team All-NBA pick in nine consecutive seasons. In 1962 he averaged a triple-double over an entire season. These days, it's an accomplishment when a player posts a triple-double in a single game. And Robertson came through in big moments. In 86 playoff games, he averaged 22.2 points per game, 6.7 rebounds and 8.9 assists.

Robertson's assists totals are all the more impressive when you factor in that they were not awarded as liberally in his playing days as they are today. Robertson has five of the top 40 seasons in NBA history for most assists. By contrast, Stephon Marbury has one season in the top 100 and that season ranks 92nd.

When it comes to assists, Robertson's true peers are John Stockton, who has seven of the top 10 seasons in assists and 11 of the top 32 of all-time. Another peer would be Magic Johnson, who had 10 of the top 45 seasons in assists. Despite what Mr. Ometu implied, Marbury does not rank with Robertson. And the comparison is even uglier with points.

The other problem with the 20 and eight link between Robertson and Marbury is that Robertson achieved his totals over his entire career. Marbury's career is still ongoing, meaning the two players are non-equal comparisons. Everybody's career has a beginning, a peak and a decline. But Marbury's numbers do not yet include his decline, so it is not valid to compare his career numbers to a retired player.

Marbury played most of last season as a 28-year old. What if we look at retired players and what they had accomplished through their age 28 season, would anybody else be added to this exclusive (if not particularly meaningful) list?

Yes, indeed they would. This may not be complete, but we would at least double the list with Isiah Thomas and Tim Hardaway, with Tiny Archibald, Bob Cousy and Kevin Johnson just barely missing the arbitrary cutoffs.

The bottom line is that Stephon Marbury has put up some impressive statistics to this point in his career. But those stats are empty, as his teams have yet to make any noise in the playoffs, much less win a title. It's my belief that no team can win with Marbury as its best player. And as far as being one of the elite point guards in the game - don't buy into that hype.

If given a choice, I would rather see Gilbert Arenas, Mike Bibby, Chauncey Billups, Sam Cassell, Baron Davis, Kirk Hinrich, Allen Iverson, Jason Kidd, Andre Miller, Steve Nash, Tony Parker or Chris Paul playing point guard for the Knicks than Marbury. Actually, I'd rather see Jamal Crawford play point and he has the added benefit of already being on the team!

But the ultimate take away is not to accept statistics that a person uses without giving it some critical thought. Or else you might come to a crazy conclusion, like Marbury belongs in the same class as the great Oscar Robertson.