WSU BASKETBALL: An Attempt To Quantify Chemistry

As statistically inclined fans and writers, we at CougCenter sometimes get accused of relying too much on numbers. Part of my personality is that I really like things that are concrete -- it's just the way I'm wired -- so drawing a deeper understanding of the sports I enjoy from statistical analysis is something that appealed to me from the start. And if there are two ways to interpret something, and there's a stat that seems to explain it, I usually lean that direction.

Heck, who am I kidding. I practically sprint in that direction.

My response to those who criticize my use of statistics is always the same: The best analysis uses both eyes and numbers. (And I'm in pretty good company in that regard.) As such, unquantifiable things such as "chemistry" and "leadership" are always tricky subjects for me. On the one hand, there's no good way to measure either one, and it seems one would have to be awfully close to the players involved to have a credible opinion on them. However, on the other hand, I cannot deny their existence.

This is the dilemma I find myself facing in terms of trying to assess what kind of impact the return of Faisal Aden, Mike Ladd and Abe Lodwick will have on the WSU basketball team, which has been playing fairly incredible ball as of late. Their dissection of Santa Clara on Sunday was remarkable. I know most people didn't see it, since it was only available online via Cougar All-Access, but trust me when I tell you that it was every bit as impressive as the final score.

Some have said the improved performance is the result of improved "chemistry." I think what people mean when they say that is "the seeming ability of the team to work together better on the floor, leading to better results." The passing seems crisper and more productive on offense. But if it seems that way, shouldn't there be some sort of tangilble evidence for that? Beyond just the scoreboard?

I went searching, and found some interesting things.

Ken Bone said before the season that without Klay Thompson and DeAngelo Casto, this team was going to have to use excellent teamwork to exceed the seeming sum of its parts. The Cougs weren't going to be able to play the way they had and expect to succeed because they no longer have an individual talent like Thompson on this roster. On offense, that means running the offense in such a way that baskets are easier to come by. In that regard, I figured assists would be a reasonable proxy for "teamwork."

I prefer assist percentage for this sort of analysis -- on what percentage of made baskets did the Cougars earn an assist? It's better than raw assists because it takes the number of possessions (which can skew the overall numbers) out of the equation. If there really is a connection between offensive success and sharing the ball, there should be some sort of mathematical connection between the team's offensive efficiency and assist percentage.

Turns out, there's not much of one: There's only a +.28 correlation between the two. Strong correlations approach 1.0, so .28 really is tantamount to nothing, telling us that the number of baskets on which the Cougars assist really has no impact on the team's points-per-possession efficiency. There's also no correlation between assist percentage and effective field goal percentage (+.11), so you can throw the idea of it leading to a bunch of easier baskets out the window, too. So, then I thought: Well, what if it's leading to more trips to the free throw line? Then there would be a correlation between true shooting percentage and assist percentage. There's nothing there, either -- just +.21.

"BUT!" you shout, "I've SEEN it with my own two eyes!" And I agree -- I've seen it too. On the floor, there's clearly a difference. So I kept looking, running correlations between offensive efficiency and all sorts of stats. Most of it turned up nothing.

However, it turns out there actually was one stat that bore a reasonably high correlation to offensive efficiency: Assist-to-turnover ratio. At +0.78, it's actually the only stat (other than the shooting stats, which are predictably high) that exceeds the critical value (for Pearson correlation coefficients) of 0.632 for eight degrees of freedom at the alpha level equal to 0.05. It's the only one that's statistically significant.

You might be thinking to yourself, "Well, duh -- getting assists and taking care of the ball would have a high correlation for any team." Not true! We only need to look back at last year's team to see a squad whose correlation between offensive efficiency and assist-to-turnover ratio was insignificant -- just +.42 for those guys. The reality is that no, not every team needs to generate assists and limit turnovers to be successful. Think of Kentucky, where the offense is predicated to high degree off the dribble drive. And think of fast break teams such as Washington, where turnovers aren't the end of the world when you're trying to run with the ball. And last year's Cougars? The scoring prowess of Klay Thompson can make up for a multitude of issues.

Please, coaches in the audience, don't flog me. Of course you want to generate assists, and of course you want to limit turnovers -- any team is going to be better when it does those two things. I'm just saying that for some teams, it's simply not critical.

Which brings us back to this year's Cougs, for whom these two factors do seem to be critical. The more I thought about it, the more it seemed to make a lot of sense for what we've seen on the floor. Early in the year, there were some major problems sharing the ball. They actually were generating some assists, but they were throwing away a startling number of possessions to do it, and it was short-circuiting any advantage gained from the assists. Because they don't really have guys who are going to take over a game on their own independent of their teammates, execution the offensive system is paramount. Put simply, the increased assist-to-turnover ratio suggests the presence of such execution, and our eyes confirm it.

So now that we know what really seems to be driving this increased efficiency, the question is why it's improved. Of course, obvious answer is obvious: A guy who is poor at sharing the ball, recording few assists and turning it over a bunch, hasn't played in the past three games.

While I typically will eschew simple, obvious answers in search of more nuanced explanations, in this case, I think it really is probably best explained by the absence of Faisal Aden. This is a case where the numbers back up what you see on the floor with your eyes: He dominates the ball, using 30 percent of possessions when he's on the floor, but his assist rate (assists per 100 possessions) is just 19 and his turnover rate is 26. It's why, despite a personal effective field goal percentage of 55 -- really pretty good for a guard -- his offensive rating is just 97. That's not terrible, but when a guy is using that many possessions with an offensive rating under 100, it's bad news for an offense. (Just ask Washington.)

All those possessions that were being used by Aden are now being used by guys who are much more efficient -- notably Reggie Moore (who has increased his productivity immensely the last two games), Charlie Enquist, Dexter Kernich-Drew and Patrick Simon. And guys who turn the ball over less than Aden (Marcus Capers, DaVonte Lacy) are doing more of the ball handling.

How Ken Bone decides to use Aden when he returns is anyone's guess. But as Bone said, he's going to need to fit in with the good thing the team has going. Let's hope he has the guts to follow through if the team's "chemistry" is disrupted.

Huge hat tip to Craig, who checked my math and made me sound smart. Advanced stats pulled from kenpom.com and statsheet.com. Thanks guys!

-----

Want more? Get analysis delivered right to your email inbox with Cougar Sports Weekly. Also follow us on Twitter and Facebook:

X
Log In Sign Up

forgot?
Log In Sign Up

Forgot password?

We'll email you a reset link.

If you signed up using a 3rd party account like Facebook or Twitter, please login with it instead.

Forgot password?

Try another email?

Almost done,

By becoming a registered user, you are also agreeing to our Terms and confirming that you have read our Privacy Policy.

Join CougCenter

You must be a member of CougCenter to participate.

We have our own Community Guidelines at CougCenter. You should read them.

Join CougCenter

You must be a member of CougCenter to participate.

We have our own Community Guidelines at CougCenter. You should read them.

Spinner.vc97ec6e

Authenticating

Great!

Choose an available username to complete sign up.

In order to provide our users with a better overall experience, we ask for more information from Facebook when using it to login so that we can learn more about our audience and provide you with the best possible experience. We do not store specific user data and the sharing of it is not required to login with Facebook.

tracking_pixel_9347_tracker