I haven’t made much progress in the last few days; I wanted to step back from the design and think deeply about a number of issues. Ciril Rus has made some programming suggestions that would require serious changes in the code, but seem to offer excellent solutions to some of my problems. They look good, but I want to take some time to contemplate all the implications of the change.
The major issue that has occupied my mind, however, concerns a simple question: how replayable should the design be? I have always assumed that a perfect design is infinitely replayable, and that greater replayability is always desirable.
Yet, I am struggling with a difficult design here, and greater replayability will require more hours of work. The other day I had a conversation with another game designer, a man whose judgement I esteem. He praised a game because it provided a minimum of 60 hours just to complete. This struck me as odd. My wife purchased the entire DVD set for Game of Thrones, and we’re now ploughing through this monster. It comprises about 60 hours of video. So I wonder, is the 60 hours of this game equal in entertainment value to the 60 hours of Game of Thrones? I doubt it. So why would I want to devote 60 hours to playing a game that is inferior in overall entertainment value to Game of Thrones?
I don’t want or need Le Morte D’Arthur to be replayable for 60 hours. It’s a proof of concept; it need only be played once to demonstrate its merit. Yet, designing it for single play (or a few playings) raises new issues that require careful consideration.
Architectures
The original architecture I had in mind for Le Morte D’Arthur looked like this:
In this diagram, time is the horizontal axis. Encounters are the little black circles. The black arrows show how a single playing might jump from one encounter to the next; by skipping some encounters, we obtain replayability. The red and blue lines show the progress of the two state variables, pBad_Good and pTimid_Dominant, that the other actors hold towards Arthur. There will actually be eight of each color, one for each actor. The end values of those state variables determine how actors will choose sides in the final confrontation with Mordred, which in turn will lead to either victory (white circle) or defeat (black circle).
I now realize that this architecture has some problems. The most important flaw is that the player doesn’t get any feedback on his performance until the end. I had intended to provide feedback in the form of facial displays, and those will be important, but I fear that this might not be enough. I need BEHAVIORAL feedback. Aristotle says, “Character is revealed through action.” I need some actions for the characters to display their attitudes towards Arthur.
Perhaps it will be sufficient to indicate character attitudes through the reaction text, but that applies only to the Singular Encounters. In fact, the Singular Encounters are the only means by which individual character feedback can be provided. Which means that I’ll need a lot more Singular Encounters.
The temptation I’m struggling with is to have more extensive tree structures. Currently I have a few small tree structures. For example, the Showdown Battle is a tree with about fifteen or twenty separate encounters in about four layers. There’s a raid with a tree with maybe a dozen encounters. But how how much of the experience should be in trees?
I’ve thought it over at some length and I have decided that the feedback must come in the reaction comments. I’ll use the Tinkertoy Text intensifier system.
