Today I am working on the initiation of a deal proposal. Under what conditions would an actor desire to make a deal with another? There are many factors to consider:
First is the consideration of what I want to get from the deal. Currently the proposer of the deal decides what he wants after he proposes the deal; I am tempted to alter this so that the proposer begins by declaring what he wants. The verb structure would then take three forms, one for each of the possible desires:
Subject; wants a deal; for Actor B’s; aura to count.
Subject; wants a deal; for a bead
Subject; wants a deal; for promise no attack
This is troublesome because the verb icon for deal must now be mixed with the sub-icon for the desired goal. I suppose that this isn’t so bad; I already have the sub-icon mixed with the goal icon, which itself isn’t very suggestive.
DirObject could respond to this proposal with his own version of the same three verbs above. But this has to fit with the threaten verb, which I suppose would be just like the three verbs above, except that “wants a deal” is replaced with “demands”. Yes, that can work.
How to decide my desires
The next task is for the actor in question to decide what he desires. How do we do that? I can imagine that an actor needs to develop a strategy at the beginning of each day. Most of that strategy will be devoted toward figuring out whom to attack and whom to avoid attacking. I suppose that this requires a priority list for each actor, a kind of PValue. There would be two such lists: one for priority of attacking, and another a priority for avoiding attack.
So, how do we calculate these priorioty values?
[ Work outside repairing a wagon ]
You know, I really don’t need an array of priority values; I can instead use a PickBestActor. But what is the criterion for picking? It should probably be based on these factors:
How low is my P3Good of me for him for me? In other words, how much do I think he dislikes me?
How much of an advantage do I think I have over him? How likely do I think I would be to beat him?
How much do I dislike him?
Once again, an old enemy arises from the grave: uncertainty. I have put it in, taken it out, and now I think that I have to put it back in again. The reason for this arises from the need to utilize uncertainty in making decisions about what information to pursue. An actor can have three numbers, representing another actor’s shial, katsin, and tanaga counts, but those values could be wildly different in uncertainty. Thus, the actor would have a strong incentive to reduce his uncertainty regarding the aura counts.
But isn’t the pursuit of certainty meaningless, given the rapidly changing nature of aura counts? I don’t think so; a lot can happen in one day. Of course, if you’re getting ancient information from another, that just makes matters worse. Of course, they should automatically be increasing the uncertainty of their values each morning.
[ back outside to prepare a block of oak to repair a wheel ]
Well, I did get one idea that solves another problem: how does the player know his actual values of various values before stating them? After all, if I’ve gathered enough information to know that Skordokott has 1.5 ± 0.6 tanagas, then I should be able to state that to somebody else — but how do I recall that information?
The solution is highlight the best answers in the menu system. Thus, at every turn, the player knows what the ‘honest’ answer is, but is free to choose a different answer.
I think that I’m going to have to put uncertainty back into the mix, but I want to sleep on it.