Tuesday, March 24, 2009

Age vs. Ability in the WNBA

The WNBA Veterans have beaten the clock.

I've been very much interested in WNBA age scales. An age scale is a graph with the age of the player on one axis (easy to determine) and some talent metric on the other axis (hard to determine). The idea is that a player's performance can be projected across time.

If you look at any of baseball's proposed age scales, those graphs will be bell curves. As the "average player" ages, he becomes better and better until he peaks at age 28 and reaches the top of the little hill on the graph. Then, the player slowly declines as "it's all downhill after age 28" with all the little injuries and better players attriting the player's natural talent.

There are a lot of 28 year olds in baseball. There aren't so many 38 year olds, and this age-related athletic decline explains why.

Figuring out the age of WNBA players was easy. Now, I just needed a metric to evaluate them. I decided to use the Wins Score metric. (If you want to know what "Wins Score" is, just follow the link.) I like the metric better than I like the WNBA's official player metric - "Efficiency".

I limited to this analysis to ages where at least 50 seasons were played. For example, there are only 20 seasons from 36-year olds, so we won't look at an "average" 36 year old. Not enough data.

I expected a bell curve - what I got was shocking.

Player AgeMean Wins Score

Holy crap. The values imply that as a WNBA player gets older, she just gets better and better and better...on the average. The values also implies that the best WNBA players - the ones who could most influence an arithmetic mean, or simple average - are on the upper end of the age scale.

I tried this analysis on John Hollinger's PER. I weighted PER by minutes played. Pretty much I got the same result, except that the numbers were less eye-popping as the range of values for PER is tighter for that of Wins Score - a great PER is above 30; a great Wins Score is above 300.

If you look at ages for which I didn't have 50 seasons, the values get even higher. At 34 the average Wins Score is 63.28; at 35 it's 65.92! Only at age 36 do we see a decline.

Since this flies in the face of what we'd expect - that the best players should be about 28 or 29 and then fall apart as they get older - we need some sort of explanation.

Here are some hypotheses:

1) The WNBA coaches and GMs are excellent managers of talent. They simply weed out every player who isn't good and keep weeding them out. (Okay, you can stop laughing.)

2) Marginal players become discouraged by sitting on the bench and leave the WNBA before they hit 30. Possible. But I don't know any WNBA player that didn't love the game so much that she wouldn't play it even for limited minutes.

3) Financial pressures: I think this a great reason. Note that there are two jumps, one at age 23 and another at age 24. The jump disappears at age 25. My understanding is that after three years, players stop being rookies and stop getting paid with the rookie scale. Anyone in the WNBA who hasn't shown a coach or GM something after three years gets weeded out of the league. After 25, you are an official WNBA survivor.

Since WNBA veteran salaries are in a narrow band - between $50K and $100K at the uppermost - there are different pressures than in multimillion dollar pro sports. When a NBA player is being paid $10 million a year you play him even if he isn't any good. Those kinds of pressures don't exist in the WNBA. Since the star factor is also diminished - WNBA players aren't national superstars - there's no pressure to play a popular player whose skills are diminishing. (Look at what happened to Sheryl Swoopes in Seattle. If it had been Sam Swoopes playing for the Oklahoma Thunder that situation would have been handled much differently.)

Furthermore, since it's very hard to establish a career on just $45 K for a limited number of years, there's more pressure on players to bail out and do something more lucrative.

I don't know if hypothesis #3 is the truth, but I think it comes closer to the truth than the other two explanations. "Locked into six years at multimillion dollar contract" disappears. "Can't waive her because she's a national superstar" disappears. In short, they don't pay or play you in the WNBA at age 33 unless you deserve to be paid and played.

4) Tighter band of athletic skill. When I say "tighter band", I mean that the deviation in athletic skill is smaller than it is in MLB or the NBA. That deviation comes from gender. The most athletic player in the NBA is probably more dominant over his average counterpart. Whereas the difference is smaller between the WNBA's average player and its most dominant player. Men are taller and have more muscular power than women do...and when that power disappears with age, male players are more substantially diminished. This leads to a discussion about....

5) Game expertise. In the NBA, it's much more natural for one player to take over the game with superior athletic ability, but I suspect it's more difficult in the WNBA. Therefore, knowledge of the tiny idiosyncracies of the game becomes more valuable.

Which refs can I work with a smile and a wink? (Becky Hammon)
Which rookie can I provoke into doing something stupid? (Plenette Pierson)
How many game situations like this have I seen before? (Vickie Johnson)

Those 35 year old WNBA players become as wily as jungle cats. They'll outsmart you. It's one of the only ways to gain an advantage in the WNBA where you can't use natural strength to have your way with your opposition (at least, until Brittney Griner shows up).

Furthermore, once the Allen Iverson type player in the NBA loses a step...he's lost. He's always ignored fundamentals for athleticism and his decline becomes that much sharper when he gets older. Since WNBA players don't have that advantage, they have to know the game to a greater degree than their male counterparts.

(* * *)

All in all, it adds up to one fact. WNBA players have beaten the clock. They're like fine wine; they just get better with age. However, if genetics catch up, we'll see the stars of tomorrow begin to dominate the game with athleticism.

Update: To see how wrong this post really was, go here to the update.


pilight said...

The problem is that you're only considering players who are still in the league at each given age. The only players who make rosters at 35 are those who are above average. Most players are long since out of the league. 37 people who have played in the WNBA turn 35 in 2009. No more than six (Milton, K Smith, Dydek, S Johnson, Sam, Penicheiro) will make rosters. Unless you factor in the non-production of Texlin Quinney, et al, your numbers will be skewed.

pt said...

Oh, definitely, but it seems to me that this doesn't seem to be what happens in non-WNBA leagues. Underperforming players manage to remain in MLB, as least as far as I can tell, due to financial or other considerations.

It might be useful to post an "attrition table", which indicates how many 33 year old, 34 year old, 35 year old seasons there are. Those numbers definitely rise to age 23 and then fall every successive year after 23 (i. e. there are fewer 24 year old seasons than 23, fewer 25 than 24, etc.) I'll post the table when I get a chance.

pilight said...

It does happen in other leagues. Hollinger and the other stat heads have ways to adjust for it. Bill James discovered the same effect 30 years ago.

pt said...

pilight, I'm going to give this some more thought. I found somewhere else another way to approach this problem, and using the approach we might get our expected answers after all. It involves, say, comparing the 23 year old cohort to their performance at age 24. I'll post something when I have some time to work things out.

Rebecca said...

I think there's also a bias because of the way the league started. If you look at those great Houston teams, for example, Cynthia Cooper was 34 years old before she played her first WNBA game, Tammy Jackson was Arcain was 28, and Swoopes was 27 (Thompson was the oddball, a pure rookie). For the Liberty, T-Spoon and Sue Wicks were 30, Sophia Witherspoon was 27, Kym Hampton was 34, even VJ was 25. Pettis, Gillom, and Timms for the Merc; Mapp, Stinson, and Bullett for the Sting; Braxton, Edwards, even Jones for the Rockers. And so on and so on. And then you had the teams that built on the ABL players, like the Sparks and the Monarchs.

I don't think this metric can be accurate yet, in other words.

pt said...

Rebecca, pilight:

Check the new post above. :D