Developement Idea: Ends Focused Roster Management

[albionmoonlight]

It seems that a lot of complaints about gameplanning/roster shuffling take the form of "I am trying to get my team to do [X], but I can't find a gameplan that will make them do it."

X here is defined as something like "get my running backs to split carries 50/50" or "blitz about 25% of the time."

One of the problems into which we run is that it can be very hard, working from the bottom up, to do something simple like get a 50/50 split on RB carries. Between playing time, endurance, formation usage, depth charts, etc. we can only make educated guesses about how our bottom up approach will actually play out in a game. I know that sometimes I would have the game play out just how I wanted, only to have the next week completely bullocks up without me changing the gameplan at all. This isn't that suprising since the gameplan is based on percentages.

So, I am just wondering how hard it would be to design a sim (football or otherwise) where you did not imput a complex bottoms-up set of directions, but instead told the game what you wanted to happen in a more simple form.

For instance, you tell the game that you want player A and player B to split playing time 50/50 (or 60/40 or 70/30, etc.). Or you tell the game that you want player X to play the first half, and player Y to play the second half.

This approach, of course, seems much easier when it comes to playing time than it does when it comes to play calling. I don't know if there is much functional difference between instructing a team to take between 15-17 three point shots than instructing them to shoot 3s 40% of the time.

But, at least when it comes to playing time, what are the advantages and disadvantages to the end-focused system? As gamers, we think in terms of "I want my young DE to play about a third of the snaps." Why not let us tell that, in pretty much those terms, to the game?

[sabotai]

The problem is that the AI must be given enough leeway to adjust to game situations. Say you want a 50/50 split in RB carries, but by the middle of the 3rd quarter, one is averageing 6.5 ypc while the other is only getting 1.6 ypc when they both have, say, 13 carries. It'd be kind of silly to stick with the 50/50 split when one is obviously having a much better day.

I don't like the FOF way of gameplanning (FTR, I do think 2k7 is much better than 2k4) having you input screens of percentages. I certainly think there are other ways to do gameplanning, but to force the AI to stick to other, rigid percentages isn't the answer. It's one thing to tell the AI you WANT the RBs to split carries, but it's another to make the AI do it regardless of the game situation. The AI has to be able to adjust the gameplan based on in-game situations and how the players are performing.

One thing that would improve FOF though, IMO, is if it used weighted percentages. It seems like in FOF, it essentially rolls 1-100 every time and matches against the gameplan. I've had a game where 90% of my runs when to the left (and averaged about 2.0 ypc going that way), the rest went to the middle, even though my gameplan called for 30% to go to the right, 30% to the middle (40% to the left). After a few runs to the left, it'd be nice if the game weighted the rolls so that the likelyhood of it coming out "middle" or "right" were much higher, to bring the percentages more in line with the gameplan. And then after it did, "unweight" them. (EDIT: If I remember correctly, my example happened in FOF2k4, so maybe it changed in 2k7)