2012 NBA Playoff Predictions Using Lineup Data

| by

We recently observed that adding lineup data to some basic overall team info gives us stronger predictive power for the postseason. In the last 4 years, anytime a team has been better than its postseason opponent in Pythagorean Wins — based on total points scored throughout the year — and has had a better high-minute lineup, the team has won over 83% of the time.

Before we jump into what this method forecasts for the 2012 playoffs, a quick note on the all the hectic shuffling in this shortened NBA season. The Knicks have a new coach (Mike Woodson). The Celtics have a new look (Avery Bradley and Kevin Garnett at center). The Lakers have a new point (Ramon Sessions). The Magic have no Dwight Howard. And anything else you can think of. So with that said, what do the post-all-star NBA standings?

Recency has never been a great predictor of postseason play, so Spurs fans shouldn’t bank too much on it and Mavs fans shouldn’t entirely panic. (OK, Mavs fans should probably panic.) This does paint a clearer picture for all of those musing about who is “hot” entering the postseason. Please remember, that the last few weeks of the season involve all kinds of weird rest, lineups and practices (“tanking”), which explains why this isn’t an incredibly strong predictor.

Moving on to the top adjusted lineups, here they are for the 16 playoff teams. The qualifying minutes was pro-rated for the 66-game season to 181 minutes. Any team without a qualifying lineup (see: Orlando) has it’s next best unit inserted. Lineups courtesy of the great basketballvalue website...

Get the rest of this article over at Back Picks.