I almost feel like after playing a game long enough, and grokking a system, I can "see" a possibility space in the form of a kind of cloud of fog, either bleeding out gradually at the edges, or being contained within strict glass walls. This shape, of course, depends entirely on the game. But in a way, I see game design as a kind of sculpture.
I love the concept of "possibility space", and the idea that designers create and tweak rules which "sculpt" and "erode" it into different forms. I love the idea that systems can be represented in some form, other than themselves, and that this form can make systemic elegance or ugliness more obvious to the casual viewer.
I’ve had stabs at explaining it before (there’s more than a few drafted in this blog, but which I haven’t dared publish).
JP just linked me to this site, and I think it might help me start to make a point. Take a look. I'll wait.
The image of a ball's path during a match illustrates the beginnings of how I see possibility space in my head.
Notice in the pictures how one corner of the pitch is fairly empty. What could that mean? Could it be that their forward attacker has a bad left footed cross? Or could it be that the right-back defender is especially good? But this is not a possibility space. This is One Path Through Possibility.
You can imagine what it might look like if millions of matches between the same two teams were over layed on top of each other. The ball paths themselves start to become unreadable scribbles, but all of a sudden, a meta-visualization emerges: a cloud. It would show how those two teams' strategies and abilities bounce off one another. It would show up their most well worn approaches, and show where one team is generally weak, and where the other team is generally strong.
If we're just using two specific teams, we're only concentrating on those teams' approaches to matches. The foggy diagram has more to say about two teams' subjective experience of the game qua system than it does about the game system overall.
Now imagine that we get every football match in history, and lay them on top of each other. All we've done is widen our sample range, but I'm sure you'll see a more generalized cloud. Now we get a better picture of how human beings have adopted the game, and how they end up using it.
Still, though, there are definite areas where the ball is not seen as consistently, even though none of the game's rules explicitly suggest this should be the case. Perhaps you'll see most of the ball's time spent in midfield. That's easily understandable, as a ball must cross the pitch to get from one goal to another, and also since penalty areas are generally well defended to stop opposing players meeting their incentive - the Goal.
Perhaps we'll still see that attacks up the left flank are less likely. But why? There are no rules in the game to suggest this... no slope at that corner of the pitch, rolling the ball back toward the center.
So consider who is playing. Maybe we see this artifact because of our species' predisposition toward right-handed/footed people. This means that there are less wingers who can use the left flank well, and so the tactic is not pursued as heavily. And this idea of exploiting the most effective strategy leads us to another idea: the goals of the game (or even minor incentives along the way) warp possibility space.
If there were no explicit goals (Agon), and no aim to football other than to explore the implicit joy of playing with the ball, there would be no choice but to explore the possibilities that exist (self created/implicit goals; Paedia). Perhaps you'd see a much more even distribution of the ball's possibility across the football field, bleeding at the edges (since people will want to avoid the disincentives of throw-ins). This "pure play" might give us a better idea of how the system distributes the ball's possibility naturally. When we take that image and compare it to the incentivized version of football, we see how the Agon changes possibility.
I guess I'm just trying to say that there's a big difference between "what is likely" and "what is possible", though the former always fits snug inside the latter. The difference between these two is purely human psychology - the drive that Agon and Paedia create for us not only to explore possibility, but to feel we achieve something during our travels.
Next post will be on methods used to sculpt possibility intentionally. But I'll leave you with a something to discuss - if an AI's possibility space for a game does not match a players' has it failed, or are we simply playing a different game, or an alien mentality?