Posted by brianxavier on 2/28/2013 4:46:00 PM (view original):
to clarify, I mean that when the program runs it's algorithm and is trying to figure out the result of a play:
does it take all 5 players into consideration when deciding if the team either: scores (perhaps being fouled), turns it over (via bad play, steal, whatever) or shoots but misses (rebound algorithm starts)... and then the result is posted in the play by play taking into account distro, IQ, skills of the players, etc.
Or does the program somehow consider: player A has the ball --> run program to determine next step which could be --> pass, shoot, make shot, turnover; player b has ball: pass, shoot, turnover, etc....
I thought you [Gillispie] had said in the past its really just a complex algorithm that determes 3 or 4 different outcomes rather than breaking down each part of an actual possession. The main outcomes are always either: turnover (steal, bad play), shoot (make or miss, then rebound algorithm), foul. I know you talked about it when discussing whether assists are just window dressing assigned after the fact rather than something more meaningful associated with passing skill...
If however the program goes step by step via player to player (much more complex IMO) then each individual player would be more important because a very weak player could more likely disrupt the play.... hence I am screwed with a new PG that don't know jack about the flex offense!
I am no mathematician. I am probably naive how complex these equations could be... perhaps it doesn't matter, but I was just curious what others thought.
we basically know for sure (via admin confirmation) that there is NOT passing and dribbling the ball around in the half court, that kind of stuff. its from a write up from old admin, i think, that the TSF (turnover, shot, foul) decision is known to us. a player starts with the ball, and that player certainly impacts the outcome of the decision in a big way. however, i wouldn't take this to mean that the rest of the team doesnt factor in, too. there are definitely some equations that take into account the whole team (often weighted, some players counting for more), and some that are player driven. it *used* to be the case that a player's shooting % was based on him and him alone (and opponents, but no team mates), but that has been changed per seble - at the time, it was not made clear what all equations team mates were added to, so i can't give you a for-sure example that is only based on the player.
however, given the complexity of the state transitions in the sim (which are not very complex, they are overly simplified to the point it degrades the realism, for those who look closely), you basically have to have complex equations. you pretty much need to throw in the kitchen sink, because with limited states, there are limited equations, so you dont have the option to have a bunch of simple equations instead of fewer, more complex ones.
so, answering your question, its really both. the program considers, player A has the ball, run the function to determine the next step - but the equations within that function definitely operate on more than simply the player with the ball (not necessarily every equation, but some and probably most of them). stuff like rebounding, defense, shooting - those things all take into account teams, not single players, and as that is most of the big stuff, its probably safe to just assume the team factors into everything. even if its not directly factoring in, it usually will indirectly (like, the team ratings/properties go into deciding if you get to the halfcourt offense TSF decision, and who gets the ball - even if after that point, there are equations based on the guy with the ball alone, the chance of him getting there was team-driven, so that will indirectly play into things like, how many points he scores, etc - even if it didnt factor into fg% (which it does, now, after seble's change a few years back)).