Be Careful What Analytics You Trust

In the past, we’ve suggested for coaches to create their own analytics to determine the effectiveness of each individual team, system and game performance and, at the same time, be wary of standard methodologies.

Case in point-this past weekend I watched a double overtime college game and when I examined the box score afterwards and focused on the +/- column, I found the following baffling numbers.

Losing team’s Player, A played almost 48 minutes, scored 26 points but was strapped with a -8, meaning, of course, that his team recorded a deficit of 8 points while he was on the floor.

Losing team’s Player, B played just over 17 minutes, scored 2 points but was credited with a +13, 22 points better than Player A.

Player A was clearly the losing team’s best total performer and Player B was very likely the losing team’s 8th most effective player in that game.

But we’re not finished.

Winning team’s Player, A played over 45 minutes, scored 31 points, sent the game into the 2nd overtime with a basket and hit the game winning triple and was credited with a +8 +/-.

Winning team’s Player B played 12 minutes, did not score and was credited with a +17, 9 points better than the game’s MVP.

Player, B was the Winning team’s 7th or 8th most effective yet was credited with a +/- higher than all of teammates by far.

These two examples are extreme and quite easy for a game observer to read through but clearly illustrate the danger of using numbers for numbers’ sake.  The standard NBA +/- doesn’t factor in abbreviated appearances, WHEN a player plays, WITH WHOM a player plays and AGAINST WHOM a player plays.

+/- isn’t the only analytic that can lie.  It’s just an easy example to cite to caution coaches.

If one is searching for a more meaningful +/-, please read Hoop Coach article, “Best Analytic? A Vote for True Game Value and True Season Value”.

Leave a Reply