Finding the formula for team chemistry…or dynamics…or cohesion…or something…

. Thursday, August 28, 2008
  • Agregar a Technorati
  • Agregar a Del.icio.us
  • Agregar a DiggIt!
  • Agregar a Yahoo!
  • Agregar a Google
  • Agregar a Meneame
  • Agregar a Furl
  • Agregar a Reddit
  • Agregar a Magnolia
  • Agregar a Blinklist
  • Agregar a Blogmarks

A few weeks ago, PT from the Pleasant Dreams blog commented that “chemistry” didn’t seem like quite the right word to describe what I’m looking at since that implies that players don’t get along with each other. But I recently got some clarification on that.

Eric Musselman, a former NBA coach who blogs about basketball and observations on sports coaching had a post last Friday entitled, “Finding the formula for team chemistry”. He cites an article from Pat Bloom, the head baseball coach at the University of Wisconsin- Stevens Point, that separates chemistry into two types – task cohesion and social cohesion. Uh oh… more terms…but the explanation is insightful:

Task cohesion "refers to a team’s ability to function as a collective unit and perform effectively on the field. If your team has a high level of task cohesion, meaning that they play well together and remain united in the pursuit of the team’s goals, then they are more likely to enjoy success."

Liking each other, simply being friends and enjoying hanging out together, i.e., a team with high social cohesion, "means very little in the way of predicting your team's performance." In other words, just because your team gets along doesn't mean they'll win any games.

In fact, according to Coach Bloom, "it has even been found that teams who are high in social cohesion play worse as a team."

But a team with high task cohesion isn't guaranteed to succeed. However, there's good news for basketball coaches, according to Coach Bloom.

"For team sports like basketball and ice hockey, where players’ movements and verbalizations must be highly interactive and coordinated to achieve success, it has been found that greater levels of task cohesion relate to greater team success."
It would be interesting to find a way to read more about the theory that teams that are high in social cohesion play worse as a team, but that’s besides the point for my current task. The point here is that finding ways to measure task cohesion is of huge importance to basketball analysis.

So during the break, I spent some time looking more deeply at the “team dynamics” ratings I have used in the past to make them more useful for analysis. Here’s what I came up with.

Deriving a formula

Originally I started my analysis with the concept of synergy, which was a metric created to say something about a team’s offensive cohesion by adding their assisted field goal percentage to their true shooting percentage. I then looked at a team’s opponent’s synergy to get a sense of their ability to disrupt the opponent’s synergy. Thus, I used synergy differential to measure defensive and offensive task cohesion.

However, synergy differential didn’t seem to do much to explain why a team won and lost, so I added other statistics based on things that seemed to explain why certain teams succeeded despite having low synergy – offensive rebounding rate and turnover rate.

So then you have the following formula, which I have called "Team Dynamics":

(Team synergy – opp. Synergy) + off reb rate – turnover rate = team dynamics rating

Essentially, those numbers described the core principles of basketball: the ability to move the ball to create high percentage scoring opportunities, the ability to disrupt the other team’s offense, and the ability to manage possessions effectively. In terms described by Bloom above, team dynamics are used to measure the degree of cohesion for a given team.

However, that still didn’t fully describe why teams win or lose – unfortunately there were six games in which the team with the higher team dynamics rating lost. And so I tweaked the formula again.

Adding the fifth factor of team dynamics

You may notice that the formula bears a striking resemblance to Dean Oliver’s Four Factors. However, there was one factor I left out as I “derived” the formula from common basketball sense – free throw rate.

Free throw rate is a ratio of free throws made to field goals attempted. In plain language, it describes the percentage of their offensive production that comes from the free throw line. An opponent’s free throw rate says something about how many free throws a team is allowing, which is also a proxy to how much they are fouling.

When I added that to the existing formula it described the winners and losers in every game except one -- Los Angeles vs. New York on Friday, July 25th (Los Angeles had a huge fourth quarter and a massive offensive rebounding advantage in that game). But that also required an additional change – instead of using true shooting percentage (which includes free throw and three point shooting), I switched to effective field goal percentage (which just weights three point shots more heavily).

In addition, I realized that looking at synergy differential over the course of a season was not that useful – synergy will vary from opponent to opponent depending on match-ups, so it’s more useful to just know a team’s average differential and compare it to the opponent they’re facing. So here’s the new formula:

Team synergy + off reb rate + free throw rate – turnover rate

But now there’s an additional problem – it only measures offensive cohesion. But there’s a very simple solution to that based on what I’ve done with game analysis – it’s simple one team’s offensive cohesion vs. the other’s. So the opponent’s offensive cohesion really tells us quite a bit about a team’s defense. So over the course of a season, if we look at a team’s offensive cohesion and their defensive cohesion based upon their opponents’ statistics, we get an offensive and defensive cohesion rating.

Why is this useful?

The reason I find this useful is that it’s not just a simple way of comparing teams to see how they’re playing, but it also allows us to say something about why a team is playing well or poorly on both sides of the ball.

When we want to know what happened in a game or even in a given quarter, these numbers give us the opportunity to actually tell a story of what happened beyond the final score and why a team won or lost. For a basketball geek like me, it’s extremely helpful just to get an idea of what makes each team tick.

So here are the team dynamics numbers for each team (their overall offensive and defensive numbers are on today’s other post):































































































































Team Dynamics
 TeamSynergyOReb RateFT RateTov RateDifferential
 Seattle Storm70.4730.98%26.32%18.30%+8.82
 Connecticut Sun72.7831.19%23.90%17.62%+8.29
 San Antonio75.2725.29%27.53%19.39%+8.26
 Los Angeles72.9132.72%24.53%20.93%+7.76
 Detroit Shock71.6634.21%22.31%17.66%+4.99
 Chicago Sky71.1931.29%26.88%18.91%+3.88
 New York Liberty73.4528.22%24.57%19.52%+3.15
 Sacramento66.2233.15%28.72%19.98%+2.27
 Houston70.5733.68%24.68%20.91%+1.21
 Minnesota70.4730.98%26.32%18.30%+1.14
 Indiana67.0928.62%21.13%21.02%-3.75
 Washington69.4232.45%19.24%23.02%-7.30
 Phoenix70.3630.83%25.24%16.39%-10.87
 Atlanta65.6828.20%24.64%20.42%-26.06






The San Antonio Silver Stars are an excellent example of how these statistics are helpful. They currently have the best record in the league while having a below average offense statistically and above average defense. The easy explanation is that they are just a very good defensive team. However, that doesn’t tell the full story.

You can’t really say they’re just winning with defense when they have two of the league’s top ten scorers (Sophia Young and Becky Hammon) and a third (Ann Wauters at #21). But when we look at their team dynamics, we see that they have consistently had the best synergy rating of any team in the league. In other words, they not only have a versatile set of scorers, but they also move the ball extremely well, which makes them difficult to defend.

When we can look at teams in terms of strengths and weaknesses it only enhances our ability as fans to talk about and understand what makes our team great…and vulnerable. It helps us analyze player transactions and perhaps even matchups.

A preview of tonight’s match-ups

So I’m going to stick my neck out a little and try to make some predictions about two games tonight that I plan on watching: Chicago vs. New York and San Antonio vs. Phoenix.

Chicago vs. New York

The playoffs will start early for me this season – the playoffs to decide my favorite team, that is.

The Liberty and Sky will play three times before the end of the regular season and by the end, I should have a good idea of which team is my favorite. As I’ve implied previously, I’d say I’m leaning toward the Sky, but there’s just something I love about the Liberty’s style of play…

Anyway…

I’ve already noted that the key players for Chicago in this game will be Sylvia Fowles and Jia Perkins – a post presence and a perimeter scorer, two positions that are actually a strength for New York. Fowles vs. Janel McCarville and Perkins vs. Loree Moore/Leilani Mitchell will be exciting to watch.

But based on these team dynamics Chicago is the more efficient offensive team whereas New York is the stronger defensive team. In fact, these are two of the teams whose numbers don’t at all reflect their records – New York is over-performing its numbers and Chicago is under-performing its numbers. But there’s a story even there.

The key appears to be turnover percentage – Liberty opponents have committed the third most turnovers in the league whereas the Sky’s limited success is predicated on playing a safe brand of basketball. In addition, the Liberty are a poor offensive rebounding team, but the Sky allow the third highest offensive rebounding rate in the league.

In other words, the Liberty have won games despite poor offensive rebounding but the Sky allow a high percentage of offensive rebounds. And the Liberty win by creating turnovers while the Sky thrive on playing it safe and waiting for shots. Then there’s the x-factor of Fowles who should influence the game on both ends of the floor.

So my pick? Right now, I really like the Sky’s lineup and think Fowles will help them keep the Liberty’s offensive rebounds down. But offensively, I’d imagine that McCarville and Catherine Kraaveld could force turnovers from Fowles and Dupree that will minimize their effectiveness and force the Sky’s perimeter players to win the game.

Even as well as Jia Perkins is playing, I don’t think she’s the type of player who can singlehandedly win a game. Also, it’s worth keeping in mind that the Sky have lost seven consecutive road games while the Liberty have won six straight games in Madison Square Garden.

There is hope for the Sky -- IF the Sky use Fowles as a key component of their offense to take pressure off of Candice Dupree and Jia Perkins AND Fowles can respond well to the inevitable Liberty double teams then MAYBE the Sky could pull off the upset. But I see it as extremely unlikely tonight.

Edge: New York

San Antonio vs. Phoenix

I'm excited to see how Diana Taurasi plays in her first game back, so I'm watching this one as well instead of the Sparks-Monarchs game which should also be pretty good.

Phoenix needs to set the tone for the rest of the season and this game would give them a huge energy boost. But really, San Antonio is the worst possible match-up for the Mercury – they play ball control offense and solid team defense.

Los Angeles exposed San Antonio’s one key weakness – their opponent’s offensive rebounding. But without Tangela Smith, it’s unlikely that the Mercury will be able exploit that weakness, even though they are an average offensive rebounding team.

Even if Taurasi and Pondexter go off for huge games, it’s unlikely that the Mercury will find a way to stop both Wauters and Young from dominating on the boards and extending possessions.

The way the Mercury could win? If Taurasi scores 30+ points and they find a way to contain Young with the Rover defense, the Mercury have a chance. But with the season Young was having prior to the break, it’s unlikely that will occur.

Edge: San Antonio