BuzzerBeater Forums

let me get your depth chart straight

rows are different position PG, G, SF, PF, C from top to bottom
columns are different depths, starter, backup, reserve from left to right
different letters correspond to different players

A A A
B C C
D E E
F G H
I J K

now as i understand it in the fourth quarter A,B,D,F and I were taken out as the blowout logic dictates. So then in looking to fill the PG spot, first the algorithm picked H, and then rotated the rest of the PG time between G and J? is that correct?

So i'm guessing at first J played center, then got tired, got taken out for K at center, but at some point became a better option at PG than H and so got subbed in there, even though he was still not rested enough to be a better center than K.

does that jive with what happened?

Last edited by BB-Forrest at 4/3/2008 3:09:06 PM

and which game are you talking about? as in your most recent scrimmage, it seems that after your starting point guard was taking out at the start of the 4th quarter.. the same two guys swapped back and forth.

and which game are you talking about? as in your most recent scrimmage, it seems that after your starting point guard was taking out at the start of the 4th quarter.. the same two guys swapped back and forth.

I am talking about my most recent scrimmage.

To use your chart: in the beginning of the fourth quarter, A got replaced by H. Subsequently, the game proceeded to rotate H and K at the PG position for the remainder of the quarter. The remaining 4 players (all backups) played the entire fourth quarter.

I don't see a reason why this should happen: H was, allegedly, the freshest player on the floor, and K is a C in the depth chart and C by suggested position. Why didn't the GE rotate K at the C position?!

because J was a better center than K throughout the entire 4th quarter... exaggerated by the fact that J is higher on the depth chart than K. Whereas K is on an even playing field with H (and everyone else) in terms of depth chart at PG...they are both not on it. If you had even a single other PG in the depth chart, they would have played over both H and K... and K probably would have gotten very little or no PT at all.

Remember the coach isn't asking... how can I get time for K. The coach is asking.. ok.. how do i put the best possible team out on the court, given their ratings/energy and modulation by the instructions for my GM. And I think that is the question the coach should be asking.

Last edited by BB-Forrest at 4/3/2008 6:36:19 PM

I guess the explanation is fair enough. Which of course doesn't make it less bizarre (or annoying ;)) in terms of end result.

I guess that the bizarre part comes from the fact that even though the coach is generally trying to rotate the best players, the blowout rule causes him to unconditionally rest the starters. Which might in effect field a better squad ;)

I understand what the case in point is, but monitoring the minutes of players is not unheard of, and just about any NBA coach that's worth his salt does that one way or another. That is, "how can I get time for K is a very legitimate coaching question. I don't see any reason why it shouldn't be hard coded in BB's GE.

To put it another way, when the other guys are getting blown out, the coach has no business worrying about putting the best team on the floor. If he did, he wouldn't be benching the starters anyhow. I imagine his primary concern should be playing as few players out of DC position as possible.

I guess I am going to stop here, because I am rambling. Cheers

so its very easy to take specific players out of consideration within the context of a "put the best team on the floor" sort of algorithm.

its harder to write a whole separate algorithm based upon equalizing minutes amongst backups and reserves. (but maybe we should do that)

its even harder to write an algorithm which both tries to put the best team on the floor and distributes minutes... for how do you quantify that tradeoff.

fwiw its not clear to me that a nba coach really pays close attention to minutes. He pays closer attention to how tired a player is and what his foul situation is, and how he is matching up with the opposing team.

it is precisely because he is not thinking "oh i need to get so and so some minutes" that you see this big difference in minutes between starters and backups generally... cause once someone has moved in the coaches mind to being better the coach leans towards having him in the game if he isn't too tired, having foul problems, or he needs a different sort of lineup for some reason. A smart coach might think ahead and say.. oh.. if i don't give him a rest now he's gonna get tired at the end of the game so i better do that.

i really don't mind debate... i've just spent time thinking about this problem a lot so I tend to have a lot of opinions.

Last edited by BB-Forrest at 4/3/2008 7:35:21 PM

Dammit, there's a thread somewhere from back around Christmas where BB-Charles (I think... maybe it was you, though?) and I had a productive discussion on this topic, and where there was even agreement to try to change a few things, although I'm not sure if anything came of it.

I can't find it... if you have any super BB-tools to find that thread, it might be helpful to this discussion.

I am basically just tossing stuff out there, and some of it might hopefully be useful.

I agree with most of what you say, but do have in mind that in a blowout you are effectively playing a quarter with just 7 guys. It takes very little to get to the point where the coach starts doing funny stuff, so essentially the reserve assignment becomes moot.

It is really a great job what has been done to the game engine so far. Of course, it is not perfect, but I think it is already quite realistic.

In my opinion, changes should be directed only to make the game closer to real basket and not to solve the managers problems. It is the manager who should adapt his/her decitions to the game and not the opposite. The amount of minutes only exists because of the BB training mode. Nothing to do with real basketball, so I don´t think making it easier should be a priority. Unrealistic line-ups like having only one player for one position are not supposed to lead to logic results.