APBRmetrics

In another thread there is talk that advanced metrics are subjectively created and not objective. It so happens that as a diversion from other things I am trying to listen to some philosophical podcasts. The first ones on Kant's Critique of Pure Reason. Talk of analytic and synthetic propositions. Hey there might be a link here. Are subjectively created methods a mistake or just the way you have to go about learning about things beyond the physics?

Is anyone reading here knowledgeable of metaphysics (from various authors), applied it to basketball in general or some of the recent discussion or willing to try? Has the Sloan Conference ever invited a philosopher or quoted some? Will it? Has any sports analytic book or article given this a good effort? Does anyone else think it would be worthwhile, important or essential? We study and debate within certain jargon and bounds. Time to reach beyond? I don't know. Maybe I will learn a little more and think / see better or not. Construct and understand or get more lost from reality.

My belief is that you have to have played basketball to know anything about it.
Now that is a strong position and certainly debatable. But a lot of time and energy can be wasted in trying to get an analysis of the subject, if you don't have a clue regarding what makes sense.

Ideally, an analyst's explanation for what happens is going to make perfect sense to the practitioners of the game. Just as a robotics technician doesn't have the surgeon's understanding of a complex surgery, it's the surgeon's opinion that matters. If the technician "gets it", the surgeon can see that and accept the suggestion.

Having a 'feel' for the game is initially more important for the player. If he also understands why something works, that may improve his game. Or it may cause him to over-think it.
The coach is in the position of balancing these 'feels' and 'understandings'. If he was once a player, he's experienced in one and learning the other.

Mike G wrote:Ideally, an analyst's explanation for what happens is going to make perfect sense to the practitioners of the game.

I would suggest that is a pretty questionable statement. If everything science can give basketball, and any other field for that matter, must defer to the opinions and biases of those already entrenched in their views then humankind will be making little progress.

Is the NBA game so different that experience playing division I college, division 3 college, high school or rec league don't inform or much? For a coach, gm or analyst? Can you see and feel accurately or usefully from the stands or a tv monitor and from listening to players and others with different perceptive experiences and / or analytic or synthetic reasoning?

A quick google search shows me that there is a good amount of philosophy of sport material out there but that most of it addresses concerns other than mine. Metaphysics of basketball is a term with some usage but mostly focused on individual mind. I am most interested in metaphysics of team basketball optimization. At quick glance I did not see that phrase used. I did see a paper that asked if seeing an object as a bundle of qualities was the right way to look at it and that might be worth thinking about more. Conceptually and from experience.

v-zero wrote:

Mike G wrote:Ideally, an analyst's explanation for what happens is going to make perfect sense to the practitioners of the game.

I would suggest that is a pretty questionable statement. If everything science can give basketball, and any other field for that matter, must defer to the opinions and biases of those already entrenched in their views then humankind will be making little progress.

What would be ideal?
Outsiders should always tell insiders how to do their jobs?

That's an oddly combative way of looking at things. Ideally people would be objective, without biases either within the game or regarding their advisers.

It might be ideal if any given person was an expert at every aspect of whatever it takes to optimize their job performance. NBA players, for example, would not only be world-class athletes, but also expert basketball strategists, nutritionists (as well as cooks), exercise physiologists, massage therapists, physical therapists, maybe surgeons (for when things go really wrong), athletic trainers, statistician/data analysts(?), and so on. Couldn't hurt if they knew something about marketing, publicity, or contract law either. Unfortunately, I don't think there's enough time in the day. Should basketball analysts have played basketball? I'm sure it would help, but I don't think it's a necessity.

I do agree with Mike, though, that recommendations have to make sense. A player might not have to believe something right off, so long as they can be made to do it until it makes sense, but information has to be communicated and understood to be actionable. If a physicist tried to fix a player's shot by telling him to release it with velocity v at angle X, he isn't going to accomplish much.

Instead of just using video to count subjectively determined good screens or categorizing them and perhaps summarizing raw scoreboard results that follow after or just using rapm to capture the effects of screens (and everything else), one could count considered good screens and use that input to measure rapm estimated impacts of them or do that for every category screens, though you might have to do multi-regression or sets of reggresions for an action with and without the others to separate screen value from value of other major actions. It seems to me that this synthetic analytic approach is more promising than just doing one (and in the case of some simple empiricists aggressively rejecting rapm because of its method or metaphysic feel).

xkonk wrote:... information has to be communicated and understood to be actionable. If a physicist tried to fix a player's shot by telling him to release it with velocity v at angle X, he isn't going to accomplish much.

That's right. The farm extension agent has to learn from the farmers, as much as the farmer is learning from the agent.

A kid in a dorm room may decide players should strive for dunks and free throws, since those yield the best returns. And every player and coach knows those only come about by certain circumstances, and that overplaying those goals will always come with a cost -- turnovers, offensive fouls, etc.

It's a disservice to say things like: "NBA people don't seem to know that a 3-pointer is worth 50% more than a 2-pointer" -- you people are ignorant, and I am knowledgeable, in other words.

We still don't know the upper limit of intelligent, efficient 3 pt taking. In first year of its availability teams averaged 3 attempts per game. Ten years later it was 7 and there probably were players and coaches who thought that was close to the max you could get / responsibly should take. Ten more years to get to 14 per game. At sensible limit or near it? Nope. Gone up another 7 per game on average in last 14 years and the team leaders are 5-7 per game above that. So who was more right, those who advocated more, more, more or those who thought the limit was limited / lower based on their experience? What will the ultimate limit be? The increase rate has slowed some but is still substantial over time.

Mike G wrote:It's a disservice to say things like: "NBA people don't seem to know that a 3-pointer is worth 50% more than a 2-pointer" -- you people are ignorant, and I am knowledgeable, in other words.

I suppose I should take the bait, as this on this board I most resemble this remark.

The point of pointing out that 3 points are more than 2 points is not to highlight the historic inability of NBA GMs, coaches, and, to far lesser degree, players, to do math. It is to make clear that clear and obvious opportunities were (and still are) being foregone. And if it isn't for the ability to multiply by 1.5 (which is wasn't and isn't) then the answer is what?

It is not a disservice to say these things. The primary disservice of the NBA's embarrassing record on this front is to the players whose careers were not allowed to reach their potential and the fans (and players) of franchises that could have grabbed the brass ring, were their front offices and coaching staffs to have had broader (forward-looking) minds and a bit more intestinal fortitude.

Period.

And as for coaches (and front offices) having the power of "feel". Please. Pull another one.

There is no doubt that these folks are, on average, good at what they do locally. No dispute. But for anyone who believes of their superior feelings, what cannot be simply improved by contemporary statistical analysis, I ask for an explanation of one relatively recent event in NBA history, and one that I have referred to before, for its pedagogic value and broad implications: Kevin Garnett moving from the Timberwolves to the Celtics.

And in particular please explain the offers made and not made for him by other franchises in terms of what this suggests about the quality of NBA intuition. Consider this example, then tell me about the ability of those in power in the NBA to "feel" the ability of players, in descending order, to play defense, offense, and then also the anticipated path of aging.

If metaphysics is beyond the physics then it includes concepts such as teamwork, big men, pure point guards, playmaking wings, 3 & D guys, up-tempo play, half court offense, high usage players, star and role players, systems, reads, plays, defense wins championships, 7 seconds or less, comparative PPP, the passing game, double teams, drive n kick, defensive rotations, man to man, zone, rebounders, etc.
It is not a flight to the silly or unimportant.