clock menu more-arrow no yes

Filed under:

Team Synergy Part 2: Ball Movement and Team Dynamics

After going through an amateurish statistical analysis of ball movement in the WNBA and NBA, I found that it might be valuable, but only one piece of a bigger picture.

I started out by looking at assisted field goal percentage (a/fg) and from there found a synergy statistic that seemed to have some descriptive (though not explanatory) value in terms of a team’s style of play.

Although most successful WNBA teams seemed to move the ball better than their opponents in 2007, there were still some conceptual flaws that might explain a few oddball results.

First, synergy differential doesn’t take turnovers into account and most coaches know that turnovers are extremely harmful. Second, it doesn’t take into account rebounding, something that many good teams do well – NBA and WNBA. Third there’s the problem with using assists as a means of analysis for anything – assists are assigned subjectively and are therefore inherently flawed.

So after reading through Dean Oliver’s Four Factors, I decided to explore the possibility of integrating turnovers and offensive rebounds into the formula. The results are pretty interesting, though basic, and I think worthy of further examination.

So here are the WNBA results integrating offensive rebounds, turnovers, and a look at true shooting percentage instead of field goal percentage. (NBA results were interesting, but not as good as the WNBA stats…so I’ll leave that up to an NBA APBRmetrician).

(Again, if you don’t want all the explanation and wish to just cut to the chase, scroll down to the section titled: "The Final Formula")


Offensive rebounding: Making up for missing shots?

The first thing that occurred to me was that I needed to find a place for offensive rebounds. Last year’s Sacramento Monarchs did not score very high in synergy differential. But they did score well in offensive rebounds.

Following Jeff Fogle’s example, I figured if I could add another percentage to the existing synergy differential formula, it would remain reasonably balanced. So consistent with Dean Oliver’s Four Factors, I decided to take a look at offensive rebounding percentage. I just added the two together. Here are the WNBA results for that equation:

.nobrtable br { display: none }




Team
Syn Diff


Off Reb Rate




DET


4.19%
32.80%


IND
3.57%
29.90%


SAC
-3.91%
38.60%


CON
2.42%
29.60%


SEA
-0.25%
31.10%


SAS
2.09%
29.40%


HOU
-0.50%
34.20%


PHO
6.86%
22.80%


CHI
-2.13%
31.20%


MIN
-4.21%
33.20%





So in comparison to the synergy score results, these results have 7 of the 8 playoff teams in the top 10. However, the order is skewed as Phoenix is the 8th rated team, New York is still left out, and Houston is somehow in the top 5.

And looking back at four factors it seemed likely that turnover percentage could help balance things out.

Ball security is vital

It’s pretty much common knowledge that turnovers are bad. But the tricky thing with turnovers is that teams that play up tempo basketball are bound to make more turnovers just as a function of having more possessions per game. So turnover percentage accounts for that by looking at the percentage of possessions that end in a turnover. I subtracted that from the other numbers and came out with some pretty nice results.

.nobrtable br { display: none }




Team
TOV%


SynDiff+OReb%+Tov%




DET


20.44%
16.54%


PHO
16.60%
13.06%


CON
19.57%
12.45%


SAC
22.88%
11.81%


IND
22.13%
11.35%


SEA
19.85%
11.00%


SAS
21.12%
10.37%


HOU
23.48%
10.23%


CHI
18.94%
10.14%


MIN
20.12%
8.87%





So again, with the exception of the Liberty still ranked low (11th), this seems like a more accurate look at the WNBA in 2007.

However, this still uses FG% whereas Oliver recommends effective FG% which takes three point shooting into account. Oliver also has free throw shooting in his four factors. So could that be part of the solution?

How valuable are free throws and three pointers?

Since Oliver’s formula takes free throw shooting and 3 point shooting into account, I decided to try adapting the original synergy formula by using true shooting percentage instead of field goal percentage. True shooting percentage is helpful because it takes three point shooting, free throws, and two point shooting into account in one number.

So I replaced field goal percentage with true shooting percentage in the original synergy formula and came up with the following results. And for the sake of easy labeling, I'll call the total "teamwork rating" (which just sounds better than a linear stat formula).

.nobrtable br { display: none }




Team
SynDiff (TS%)


Teamwork Rating




DET


3.24%
15.59




PHO
7,25%
13.45




CON
3.00%
13.03




SEA
1.48%
12.73




IND


4.58%
12.35




SAC


-4.33%
11.39




SAS
2.86%
11.14




CHI


-2.86%
9.40




HOU


-2.42%
8.30




MIN


-4.89%
8.19







So still can’t find a way to get the Liberty into the top 8, and the order of the 7 playoff teams is now further from the standings, but I think the differences in the ordering are small...and perhaps justifiable.

The Final Formula: Explaining Teamwork Rating

What I like most about this formula is that (as of now) it’s relatively simple to evaluate a team just looking at a box score: Team synergy – opponents’ synergy + offensive rebounding percentage – turnover percentage.

In the end, it's meaningful in terms of understanding the quality of a team and they're style of play.

It’s also pretty easy to describe in plain language an approximate "formula for success" in the WNBA: if you can move the ball better than the other team while taking care of the ball and rebounding missed shots, you have a pretty good shot of winning. And then there are teams (like the Monarchs or Mercury last year) who can make up for deficiencies by doing one thing extremely well. It gives us a language to talk about what’s going on with a team.

For now, I'm calling this a "teamwork rating" because I think each one of these statistics is a function of good teamwork in addition to individual play. With a strategy that fits the personnel well, a team can rate highly in synergy differential, offensive rebounding rate, and turnover percentage.

While I don’t have a way to evaluate individual contributions to the team chemistry, this might be a start. I’ve been pretty interested in finding a way to understand team chemistry for some time and this might be a start – it describes the components necessary to build a contender. Efficient scorers, good rebounders, and good ball handlers. A good team will have players that can fill all those roles and complement (or mask) weaknesses.

For example, you could analyze trades in terms of what the incoming player adds compared to the loss from the outgoing player and guess how that might impact the team. You could preview match-ups by looking at what to take away from teams.

It’s not perfect, but I think all of the remaining flaws based on last year’s statistics can be accounted for as follows:

1) The Liberty were a playoff team, but also had a losing record. So these factors do explain the success of a winning teams to some extent. The Liberty also had what seems like an erratic season, which could account for poor regular season numbers.

2) If you look at Pythagorean wins/losses from last year, these results are relatively similar, but Pythagorean wins/losses have a better predictive quality. The trade-off is that with my assessment of teamwork (or whatever you want to call it) you can also look at a team's strengths and weaknesses as they pertain to regular season performance. As Oliver says about his Four Factors, it can be used as a scouting tool for us fans to pinpoint how to beat a given team.

3) There is probably never going to be a "holy grail" formula because real life has things like injuries, trades, hot streaks and slumps that can’t really be accounted for. So one thing I like about this approach to teamwork is that it looks at enough team variables to both describe a team’s style of play and tell us something about its effectiveness. When I look at a team like Seattle for example, if Sue Bird and Lauren Jackson had stayed healthy, might they have been second in the West? It’s reasonable to say so given how dominant Jackson was last year.

Of course I only looked at one season of data and given how short WNBA seasons are, that’s probably not enough to make any major claims. But for now, I’ll go with this.

So here are this year’s rankings as of this past Wednesday (I updated numbers then and haven’t since…so nobody had played more than 19 games at that point):

.nobrtable br { display: none }




Team
"Teamwork rating"




DET


23.45%


PHO
19.56%


CON
18.19%


SAC
17.03%


IND
16.73%


SEA
14.87%


SAS
13.24%


HOU
13.20%


CHI
11.35%


MIN
11.07%


PHO


6.85%


WAS


4.24%


ATL


3.27%


SAC


2.14%





The reason Chicago ranks so high is likely because of Sylvia Fowles' early contribution and having a weak strength of schedule thus far – Fowles was huge in terms of field goal percentage and offensive rebounds. So her five games probably still have a large effect on the team’s data.

Obviously, the Sparks' chemistry is pretty volatile to say the least, but when they get hot they’re unstoppable. Statistically, they are by far the best team in the league. This "teamwork rating" probably only reinforces my belief that Shannon Bobbitt should start for the Sparks, although Kiesha Brown is the better scorer – Bobbitt does a lot to keep the ball movement and has the ability to disrupt opponents’ offense.

The Mercury's low score is probably due to the fact that they were absolutely dreadful the first 8 games of the season. They have since recovered from that on the court, but not statistically...

The Dream’s low score is due to ranking last in every single stat evaluated. You have to wonder if Alison Bales can at least boost their offensive rebounding and field goal percentage to help them win.

As for those three games I looked at this week – Mercury vs. Comets, Lynx vs. Dream, and Monarchs vs. Sparks – if the winning team didn’t have an edge in synergy differential, they did have an edge in turnovers and/or offensive rebounds. In addition, they all had at least one player who was able to score a lot of points very efficiently – Diana Taurasi, Betty Lennox, and Nicole Powell. So good scorers are valuable when shooting efficiently.

As with the rookie and point guard rankings, I’ll try to come back to this later in the season, probably around the Olympic break once there’s a little bit more data.

For now, any feedback is welcome – I like this because it ended up matching the things I value in basketball…but does it work for you???

Transition Points:

One way to look at team chemistry is to analyze combinations of playing styles on a given team. D Sparks at the Arbitrarian blog has already done this for the NBA and offered to do so for the WNBA. In a few days (or so) I'll try to combine these two approaches to make sense of what combinations of players are most effective and what teams may be missing. Again, it won't be exact science, but hopefully interesting.

I haven't come across current true shooting percentages for individual players, but if you want to look at last year's percentages, check out basketball-reference.com. Last year, Lauren Jackson topped the league with a true shooting percentage of 63.3%. Penny Taylor came in second at 62.2%. One of those two has had the top figure in the league since 2004 -- Taylor in 2004 & 2005, Jackson in 2006 & 2007.