We think we know how it works, we look at our roster and suspect we don't have the right sort of personnel based on our interpretation of how 2-3 works, and when we run some tests scrimmages we get poor results. Why is that at all discouraging? It would be more confusing if it DID work for us somehow.
You're not wrong, but here's the dilemma... we *think* we know what types of players would make it work, but we don't know for sure. So we'd be encouraging someone (more like, someoneS) to spend 6-7 seasons training a very particular, very expensive type of player, at which point we might realize "oops, we were wrong, it still doesn't work very well". That's a huge thing to ask of the community, with very, very high risk. If people were interested in trying this on their own, with the understanding that it might fail miserably, then that'd be great, but personally, I'd be uncomfortable asking it of anyone. (It's similar to the Logsdon issue... sure, it might be great to have a player like that, but the costs in terms of training time and salary as so extreme that it's not really fair to encourage someone to make a player like that, since we know they're only going to be a marginal NT guy anyway).
As far as the balance between open vs. contested shots, that's a great point and always a huge consideration, and I understand that in general giving up flow to your opponent usually results in giving up more uncontested shots. Intuitively though shouldn't a 2-3 zone that packs everybody inside the arc contest more of everything except 3pter's? With somebody always around the ball, and the OD and SB nuances suggested by your data, wouldn't we actually expect more contested shots (including jumpers)?
You would think, but that's not how it's implemented. As I've heard it explained, OD defends passes, ID defends "post-ups". So in determining whether a pass leads to an open shot (i.e., an assist), it's a PA vs OD evaluation. 2-3 reduces team OD, and ergo causes these passes to more often lead to open shots.
I think the risk side of the risk/reward equation is why people are gun-shy about committing to training something that hasn't been proven yet (since it could embarrassingly fail), but really what is the risk: that we don't win worlds? I think we're already starting to understand that we even with good coaching, GS, community involvement, and training we still aren't a big enough community to use traditional approaches to just steamroll through some of our larger, better organized foreign competition, so I'm not sure we would really be risking all that much by taking some calculated risks based on good data and analysis.
The risk is really to the club team. You give up a lot in terms of training opportunity, and then pay a lot in terms of salary. It's a big personal sacrifice that may or may not produce something positive for the NT.
Also, I should say, I don't think community size is necessarily the reason we've had trouble at worlds. In fact, if you look at the largest user bases in BB, of the top 5 (italy, spain, poland, us, france), only poland has managed to be successful at the NT level. I think part of the problem is similar to the "micronation" issue that has come up-- having a really competitive club league is in many ways counter-productive to having a good NT. (I.e., since all our "rich" teams have such a thin margin for error, it's very hard to make the sorts of sacrifices that would benefit the NT: owning multiple high salaried NT players, training them out of position, etc). Poland has been the exception here, for which they deserve a lot of praise, but they're the exception more than the rule. The other teams that have done the best lately are the countries that have enough teams to generate a big enough based of high quality trainees, but not so big that the top divisions are uber competitive (China, Czeck Republic, Turkey, Slovenia, Chile, etc).