APBRmetrics

Team analytic efforts include staff, computers, video tools, servers, maybe consultants, coffee...

What about a part or full time test basketball team? Former good college (or possibly some pro) players to endlessly run offensive and defensive play tests on to get micro reaction measurements and results on. Do the stuff the coach and real players dont have time for. Maybe eventually after months or years of testing and video capture, the test team could go virtual. But if you want an edge on our teams, this might be a productive one. Does any team currently have anything like this? Assistant coaches and other staff might sim a few plays here and there but I am talking hundreds or thousands of hours of testing.

I guess D league teams owned and run for this purpose can do a lot of this, if run in a real regimented way and given the full video measurement treatment. But are they micro-measured? Possible at home but does any team deployed SportVu tech in a D league gym? Haven't heard that mentioned anywhere yet. And the ability to run the same play a 100 times with variations in a controlled setting would have some unique added value. Of course, can do both approaches- D league and test gym.

Crow wrote:Team analytic efforts include staff, computers, video tools, servers, maybe consultants, coffee...

What about a part or full time test basketball team? Former college or pro players to endlessly run offensive and defensive play tests on to get micro reaction measurements and results on. Do the stuff the coach and real players dont have time for. Maybe eventually after months or years of testing and video capture, the test team could go virtual. But if you want an edge on our teams, this might be a productive one. Does any team currently have anything like this? Assistant coaches and other staff might sim a few plays here and there but I am talking hundreds or thousands of hours of testing.

I guess D league teams owned and run for this purpose can do a lot of this, if run in a real regimented way and given the full video measurement treatment. But are they micro-measured? Possible at home but does any team deployed SportVu tech in a D league gym? Haven't heard that mentioned anywhere yet. And the ability to run the same play a 100 times with variations in a controlled setting would have some unique added value. Of course, can do both approaches- D league and test gym.

What you are describing is essentially being done with the Reno Bighorns and Rio Grande Valley Vipers.
the specific scenario you're describing though would never work though because a) it would be far too costly but far more importantly b) it wouldnt tell you anything. Games are much different than simulations and NBA players are much different than non-NBA players.

It wouldn't be too costly IMO for a $100 plus million / yr operation that NBA teams are to spend a few hundred thousand to maybe a mil. on this tops (increasing R&D spending by less than 1% should not be considered too outlandish to consider). A championship should be worth $50-$100+ mil.

You'd have to have the right guys making the right effort (bonus structure for performance?) with intent of offense and defense hidden from each other by mixing the plays up for at least some of the testing, but I think you could learn useful info from the test team approach that would be difficult or impossible to get otherwise. It comes down to the intelligence of the test program operators / analysts. If I am wrong and it is worthless and no way to make it useful at all (really doubt that), then shut it down after a month or two. Don't know if dont try and dont care to try for that elusive edge. The test team could both test things for later trial by D league team and NBA team and be a shop for fine tuning tiny details of things that get tried and look good with those teams. Make it a loop. Coaches do this stuff in their heads or on paper. This would be more real than just that alone.

Crow wrote:Team analytic efforts include staff, computers, video tools, servers, maybe consultants, coffee...

What about a part or full time test basketball team? Former college or pro players to endlessly run offensive and defensive play tests on to get micro reaction measurements and results on. Do the stuff the coach and real players dont have time for. Maybe eventually after months or years of testing and video capture, the test team could go virtual. But if you want an edge on our teams, this might be a productive one. Does any team currently have anything like this? Assistant coaches and other staff might sim a few plays here and there but I am talking hundreds or thousands of hours of testing.

...

Basically, that's the reason that a NBA team hires staff with basketball experience. (This is the whole 'what is analytics' thing that keeps coming up at Sloan.)

There's also the issue that the NBA is way off on the tail end of a bell curve. I don't think you could find a Lebron, Shaq or Steph Curry substitute.

The thing that I always thought could be useful is an animatronic puppet on a crane fed with motion capture data...

You are not going to be able to test for or against everything but you might be able to understand where to improve footwork or spacing on 5-10% of typical plays, test novel out of timeout plays, etc. 1 pt per game in NBA is worth millions. Sometimes a single point is worth much much more than that. Stay within the norm or go beyond... The beyonders tend to do well, maintain that stuff and look for the next source of edge.

Crow wrote:You are not going to be able to test for or against everything but you might be able to understand where to improve footwork or spacing on 5-10% of typical plays, test novel out of timeout plays, etc. 1 pt per game in NBA is worth millions. Sometimes a single point is worth much much more than that. Stay within the norm or go beyond... The beyonders tend to do well, maintain that stuff and look for the next source of edge.

Sure, if we go by salary cap per win, a marginal extra point per game is worth about $10M over the course of a season (and probably more in revenue), but how sure can you be that you'll actually realize that much of a gain from a big elaborate program like that? (It probably wouldn't take that much by NBA standards to run the research project you describe, but it's still a big pile of money.)

The claims from Sloan are that the difficulty in implementing change is more about the institution than the information. The questions generally aren't "is it good to take 3 pointers and layups?", but "how do we communicate better?" I don't think running this sort of simulation will provide actionable credibility to the research, or give it a more transparent or persuasive presentation.

Crow wrote:Team analytic efforts include staff, computers, video tools, servers, maybe consultants, coffee...

What about a part or full time test basketball team? Former good college (or possibly some pro) players to endlessly run offensive and defensive play tests on to get micro reaction measurements and results on. Do the stuff the coach and real players dont have time for. Maybe eventually after months or years of testing and video capture, the test team could go virtual. But if you want an edge on our teams, this might be a productive one. Does any team currently have anything like this? Assistant coaches and other staff might sim a few plays here and there but I am talking hundreds or thousands of hours of testing.

I guess D league teams owned and run for this purpose can do a lot of this, if run in a real regimented way and given the full video measurement treatment. But are they micro-measured? Possible at home but does any team deployed SportVu tech in a D league gym? Haven't heard that mentioned anywhere yet. And the ability to run the same play a 100 times with variations in a controlled setting would have some unique added value. Of course, can do both approaches- D league and test gym.

What are the specific questions you are trying to answer?

That would be up to the team, the experts; but I'd think you could learn something beyond current knowledge by precisely measuring and adjusting pick n roll movements and angles, how to best move the ball from side to side or max opportunity to steal it on swings, timing on cuts, angles on help d, etc. Stuff coaches know about and study already but probably without the same level of precision and documentation or as massive an effort at micro-adjustment to find optimal.