[Morgado]

Originally Posted by MalcPow

We're scouting and evaluating more because we have more little nooks and crannies to go digging into. And it's a good thing. It's working. We're all wary and wondering, comforted when a guy moves in a certain way but not wholly convicted that we know who they are and just how much they're contributing.

I think this may be a case where it's impossible for a developer to turn off what they know and see game information with uncertain eyes. This just isn't the reveal for us that it is for you. We're still the right mix of unsure but not completely lost. So I wouldn't worry too much about thinking through ways to reveal less, just let us keep feasting on all our quirky theories and half-truths and not so penetrating insights. It's great.

I thought MacPOW's post was excellent, and is related to a more general thing about not just gaming but uncertainty in statistical and economic modeling. Consider this passage from Vorderer and Bryant's Playing Video Games: Motives, Responses, and Consequences (Klug and Schell, 105-106):

What are players trying to do as they make decisions during gameplay? They are trying to gain some amount of "control" over their gaming world, resulting in an increased sense of agency.

The "control" the players seek is a curious phenomenon. Control is not what they really want, although it is the word they will use if asked. They want the illusion of control, the feeling that their actions make a difference. Now in most games the player is not in total control - after all, there are opponents (both human and artificial intelligence, or AI) and obstacles whose purpose is to defeat the player's plans. But, the player does want influence over the game system in a predictable way, mixed in with some degree of unpredictability (randomness). An example would be the chess player's desire to have, for example, a rook always move in a straight line parallel with the grain of the chessboard and not, say, move diagonally at random times. This allows both the player and his opponent not only to observe the current position of every rook on the board and predict all potential rook moves both on their part as well as the part of the opposing player, but to plan future moves armed with that knowledge. The player feels in control because he can predict what his opponent is going to do on the next turn, insofar as the moves of the rook are concerned. However, as he projects rook moves further and further into the future, and while all moves are predictable, the human players ability to store and analyze all the potential combinations of those moves grows progressively more difficult. Compounding that, we have the fact the player cannot predict which of the potential moves available to the opposing player, will, in fact be chosen. Thus, we have the delicious combination of predictability, control, and randomness, that is the essence of exciting strategy games.

The rook's behavior is predictable and therefore somewhat "controllable." Players would become very frustrated indeed if they had planned a certain move for that rook but when they took their hand away from the piece after finishing their move, it shifted one square at random from the player's chosen destination. The player would feel as if his strategic planning no longer meant anything. The player's frustration would vary by how often it performed this random move, how predictable this random pattern would be, and whether the piece could accidentally capture this way (because then the player could capture an enemy piece otherwise he could not have reached - if, say, the rook could randomly move diagonally). But, if the player could use the random move to his advantage, because it was semipredictable, then some "controllable" excitement might be brought into play. It would become a different game, for sure, but perhaps experimenting in that way might lead to an interesting new type of game.

Like in economic forecasting models, the requirement is not that all inputs have to be perfect without noise or that the model results have to exactly predict what we think we want to know. What we care is that the signal (the useful information) outweighs the noise (the masking randomness), because then we can say we learned something that's useful; we feel we are getting somewhere.

In the game setting, the players don't want to be able to perfectly control everything - because then you get a reductionist game theoretic outcome that's perfectly predictable from the start... why even play? There have to be some things out of the control of the players - but not too much. It's the hard balance to strike. The players don't want so much noise and randomness that they feel like their planning and strategies are getting swamped by luck and dice rolls. I get the feeling many players have simply given up on trying to figure out things like defensive gameplanning ("Fuck it, i'll just Rex it" etc.) because the things they try all seem to make little difference. At the same time they don't want dominant strategies or dominant solutions (e.g. BPR > ALL) that make everything else "stupid." Yes, that's a tall order, but that's what makes the fantastic games... and not every game will live up to that. That's okay. *shurg*

A sense of helplessness and lack of any meaningful "control" and that there's no way to figure out the puzzle is what frustrates players. The semipredictability Klug and Schell are talking about is analogous to the FOF multiplayer team owners sifting through all the data looking for patterns they can use to figure out booms, busts, and developmental curves. They are trying to use noisy data (ratings, but also other things as MacPOW points out) to try to bend the "randomness" of player quality revelations and ability changes to their advantage. The notion that they can "figure some of it out" is exciting and is what drives them to analyze and scout the information. This may not be the intended game that was set at the beginning, but it is an evolution of the game into something that they appear to enjoy very much.