Thursday, December 18, 2008

Another Offensive Metric: Team Field Goal Percentage

I still have a couple of Euroleague games to report, but work has taken a lot of my time - it's the ol' Christmas rush. So here's a bit of a diversion.

There's an interesting post on the Basketball-Reference Blog that writes about the various measures used to measure how important a player is to the offense of the team. There are a lot of ways to do this: by usage, by boxscore measures, by a new metric called "touches" which measures how often a player has the ball in his/her hands, etc.

The post introduces an "old-school metric", as they call it: the percent of team buckets the player played a direct role in. The formula becomes:

percent = (player FG + player assists) / (team FG)

Given that we accept the above, here are the WNBA leaders:

Top 20 Players in Team Field Goal Percentage for 2008

Sue Bird, Storm - 38.84 %
Becky Hammon, Silver Stars - 36.85 %
Candace Parker, Sparks - 36.50 %
Diana Taurasi, Mercury - 35.69%
Jia Perkins, Sky - 34.79%
Cappie Pondexter, Mercury - 34.53 %
Alana Beard, Mystics - 34.46 %
Seimone Augustus, Lynx - 33.05 %
Betty Lennox, Dream - 32.18 %
Candice Dupree, Sky - 31.92%
Lindsay Whalen, Sun - 31.81 %
Ashja Jones, Sun - 31.50 %
Sophia Young, Silver Stars - 31.30 %
Katie Douglas, Fever - 31.28 %
Katie Smith, Shock - 30.29 %
Ticha Penicheiro, Monarchs - 29.01 %
Lisa Leslie, Sparks - 28.73 %
Ann Wauters, Silver Stars - 27.86 %
Shannon Johnson, Comets - 27.73 %
Ivory Latta, Dream - 27.60 %

For example, Sue Bird of the Seattle Storm was involved in roughly 38 percent of her team's total scoring - either through scoring shots herself (FG) or giving the ball to someone who then scored (Assists). If you take a team, and add every players Team Field Goal Percentage, you'll get a number greater than 100 percent for the whole team. This is to be expected, as a field goal can be credited to two people - the person making the shot and the person finding the open player.

Of course, there are some points of contention:

a) Are assists really worth as much as field goals? Furthermore, all assists get rated at equal difficulty. (Then again, all field goals get rated at equal difficulty as well, so that's not much of an argument - all "home runs" in baseball get rated at equal difficulty for a baseball in offensive evaluation.)
b) It's an additive metric - it rewards players for overall contribution, so players that have played in more of their team's games get higher percentages than players with injuries.
c) It doesn't take into account the team's offensive scheme, although it nicely equalizes for offensive pace.
d) It doesn't take into account fastbreak opportunities or defensive blocks/steals/offensive rebounds that might have been just as important in setting up a field goal.
e) In general, the metric punishes players who don't create assists, so centers get a raw deal. Janel McCarville, the Liberty center, leads all Liberty players with 26.07 percent. But can you say that McCarville is relatively less important to the Liberty than Bird is to the Storm or Augustus is to the Lynx?

Given the above in mind, we look at the Dream.

Values for Atlanta Dream Players

Betty Lennox - 32.18 %
Ivory Latta - 27.60 %
Iziane Castro Marques - 16.54 %
Tamera Young - 13.52 %
Kristin Haynie - 12.51 %
Katie Feenstra - 9.72 %
Erika de Souza - 6.26 %
Stacey Lovelace - 5.69 %
Kasha Terry - 5.14 %
Alison Bales - 4.58 %
Kristen Mann - 3.57 %
Camille Little - 3.35 %
Ann Strother - 2.12 %
Chioma Nnamaka - 0.89 %

(Bold players started the season but were traded later.)

The only surprises for me were Kristin Haynie's 12.51 percent - I didn't know she had that much value in terms of field goal creation. Erika de Souza has a low value, but she was injured much of the season. Katie Feenstra looks great, until you realize that Kasha Terry and Alison Bales are mid-season acquisitions and their numbers compare just as well.


Anonymous said...

WHy not divide these numbers by the percentage of total minutes played by each player (with 100% equaling 34 games x 40 minutes) as a way to then scale the math to eliminate the playing time factor? I would then still nix anyone under avg 10 minutes per game though probably....

pt said...

That would probably be an interesting stat in itself: a ratio of "percentage of team involvement in shots/percentage of minutes to team minutes". It might be simpler, though, to do a "shots per minute" type of metric.

I'll try to work on this, and add a link in the comments to the final results, but not create a new post for it.

pt said...

Anon, here's the new stat: "percentage of offensive involvement divided by percentage of minutes played".

This will provide a value somewhere between 1 and 2 in most cases. For example, if Jane Doe is involved in 30 percent of her team's field goals, and plays 15 percent of her team's minutes, her above ratio value would be 0.30/0.15 = 2.00.

The following pic is a list of the top 20 players in the WNBA. Tasha Humphrey's numbers aren't quite right, since she's played for two different teams, but I'll bet her true value is close to the one stated.