Back in January I wrote a post about play types and how they affect a team’s overall offensive efficiency. Now being a lot deeper into the season, I took a loot at each of the 9 WBBL teams individually and compared their play type frequencies with their offensive efficiency to come up with correlation values.
For a reminder: There at 10 different kinds of play types that are tagged in the game of basketball according to Synergy. Those play types are: Transition, Cut, Post Up, Off Screen, Hand Off, P&R Roll Man (in a pick and roll where the possession end in the roller’s hands), P&R Ball Handler (in a pick and roll where the possession ends in the ball handler’s hands), Offensive Rebound, Isolation, or Spot Up. Play types are tagged based on how the play ends. Another reminder: a play ends when a shot is made, a shot is missed, a turnover is made, or a foul is committed.
So, what’s the point of coming up with correlation values? Why does knowing the relationship between a team’s play type frequencies and a team’s efficiency even matter? Well, let’s think about it like this… If there’s a play type that has a very strong positive relationship (Correlation value > 0.8) with that team’s offensive performance, then as an opponent you could know which play types to prevent that team from executing.
After comparing each team’s play type frequencies in each game with their corresponding offensive efficiency (Points Per Possession), I came up with the below correlation values:
From a basketball analytics standpoint, if a correlation value (aka R value) is 0.50 and greater or -0.50 and lesser, then I would say that those are great enough values to show a moderately strong relationship. The values that are in-between 0.50 and -0.50 aren’t very meaningful.
For example, Sheffield has an R value of 0.89 with the P&R Roll Man play type. This means that in games that there was a higher frequency of plays ending in the roll man’s hands, the more efficient their offensive performed. But, in games when they less frequently ended plays in the roll man’s hands, the worse they performed. As an opponent to Sheffield, this can tell the coach/team to try to prevent Sheffield from getting the ball into the roll man’s hands. If they do this, it’s statistically probable that Sheffield won’t perform as efficiently on offense.
Negative relationships have the same idea. For example, Brixton has an R value of -0.60 with the Isolation play type. This tells us that the more frequently they have isolations, the less offensive efficient they perform and the less frequently they have isolations, the more efficient they perform. As an opponent using this data, you would be inclined to force Brixton into isolation situations. As for Brixton, you would be inclined to avoid isolation situation at all costs.
You’ll notice that some teams don’t have very meaningful relationships (all R values are in-between 0.5 and -0.5). This just means that the frequency of play types isn’t a great determining factor in their team’s offensive efficiency.
For example, Manchester doesn’t have very strong R-values, so it’s telling us that statistically we can’t really pin point any certain play types that are contributing to their offensive efficiency.
To conclude, it’s not to say that stopping teams from doing certain play types will automatically make your team win more games. It is interesting though that the 9 teams differ greatly with which play types seem to show correlations. The fun thing about basketball is that so many factors go into performing well and winning games. This analysis on play types is just a solitary way to improve a team’s statistical probability of decreasing the offensive performance of their opponent.
Until next time,
“Life it is not just a series of calculations and a sum total of statistics, it’s about experience, it’s about participation, it is something more complex and more interesting than what is obvious.” – Daniel Libeskind