It’s still far too early to be doing this, but I really can’t help myself. It’s hard not to dig into the numbers given the team’s surprising success. I’m curious right now what lineups have worked the best given the uncertainty over Houston’s rotation coming into the season. Checking out nba.com, I found that Houston’s most effective offensive lineup so far has been the quintet of Trevor Ariza, Francisco Garcia, James Harden, Donatas Motiejunas, and Jason Terry. That group has posted an offensive rating of 250 in 1 minute of play. Okay, that was meant as a joke.
There have been tons of lineups that have played together in the single digits, so I’m going to dismiss all of those. The highest ranked Rocket quintet, out of all of the groups that have played at least a decent amount together, is the unit of Ariza, Beverley, Harden, Howard, and Papanikolaou. That group has played 19 minutes together, posting an offensive rating of 155 and a defensive rating of 101.6.
The group of Ariza, Canaan, Harden, Howard, and Motiejunas has played 20 minutes together, posting an offensive rating of 125.1 and a defensive rating of 75.8.
Ariza, Canaan, Harden, Howard, and Jones have spent 29 minutes together, posting an offensive rating of 119.2 and a defensive rating of 77.0.
And lastly, Ariza, Beverley, Harden, Howard, and Jones–the team’s desired opening night lineup–has spent just 25 minutes together, posting an offensive rating of 110.8 and a defensive rating of 92.8.
Ariza, Howard, and Harden will be mainstays in any crunch time lineup the Rockets throw out. But I’m very interested to see which way the data leans over the course of the year with the other two spots. Most of you guys think Motiejunas sucks, but all of the team’s best defensive units last year featured him instead of Jones as the power forward. Then there’s Papa who I am guessing will be a part of the team’s most potent offensive units. Lastly, what is the tradeoff between Canaan and Beverley? Will Canaan’s offensive versatility offset Beverley’s defensive contributions.
Right now, the above sample size is far too small for anything but grins. But in a few weeks, we’ll know enough to know what groups are working best.