A grand strategy model?
A grand strategy model?
One of the key questions for team managers / owners is whether to go truly all in trying to win title this season (with spending, use of assets, trades, etc.), next, the one after that or later. Has any team addressed tho choice with a grand strategy analytic model or do they still atop short of that? I am thinking of a model that is built on player salary / contracts / conditions, age & age / production curves, draft pick data / draftee quality / draft selection trends / specific team needs, free market models for each summer and what is known or likely about team effort horizons. There are minimum team salaries and expected maxes by market, owner, team history, etc. How complicated to make it is a resource / ability issue but theoretically one could try to get an analytic model answer to the grand strategy question. How close would the model's recommendation be to actual team choices / behavior?
Re: A grand strategy model?
I'm not sure about basketball, 538 definitely has a baseball model for that, and, though I can't recall particulars, I think I've seen discussions of the whole 'win now or win later' question in several places. To speak to the issue in a numerical way, you need to answer questions about how much marginal wins (or losses) are worth, and what the costs associated with various moves are. Sam Hinkie epitomized the 'win later' philosophy in recent memory, so some of that came up in the discussion about his departure.
We do have existing generic technology for optimization problems, and the resources available to an NBA team are pretty big, so you could probably build a model (or 10) and apply the algorithms to see what they suggest. Of course, all that is in vain if the decision makers ignore it.
We do have existing generic technology for optimization problems, and the resources available to an NBA team are pretty big, so you could probably build a model (or 10) and apply the algorithms to see what they suggest. Of course, all that is in vain if the decision makers ignore it.
Re: A grand strategy model?
With winning a title so difficult, especially outside 6-8 franchises, optimizing one's peak is especially important. If you aren't, then all the analytic driven work at lower levels is being impaired or likely for naught (titles).
Yes 538 did project future titles. I don't recall how much they gave away about the model. Did they run free agent markets?
A prominent ex-team analyst said fairly recently and not for the first time that with a rare exception he never ran or used player rankings. You can't run a grand strategy model imo without it. So either this was an exceptional use- and absolutely a non-trivial, actually huge one- or he didn't run an analytic grand strategy model.
Hinkie understood that the answer to the grand strategy question for the Sixers was "more than 3 years out" initially and to date. At some point it shifts to sooner. The peak should be when a few GOOD draft picks are on tier 1 maxes and many of the later draft picks are still on rookie scale but have grown into positive impact players. And maybe a very good or great vet has joined team. That is probably year 5 of the project at earliest. If the first wave draft picks are inadequate then it will be later... or never.
Yes 538 did project future titles. I don't recall how much they gave away about the model. Did they run free agent markets?
A prominent ex-team analyst said fairly recently and not for the first time that with a rare exception he never ran or used player rankings. You can't run a grand strategy model imo without it. So either this was an exceptional use- and absolutely a non-trivial, actually huge one- or he didn't run an analytic grand strategy model.
Hinkie understood that the answer to the grand strategy question for the Sixers was "more than 3 years out" initially and to date. At some point it shifts to sooner. The peak should be when a few GOOD draft picks are on tier 1 maxes and many of the later draft picks are still on rookie scale but have grown into positive impact players. And maybe a very good or great vet has joined team. That is probably year 5 of the project at earliest. If the first wave draft picks are inadequate then it will be later... or never.