Chris Crawford on Interactive Storytelling (44 page)

What’s the difference between blocking the option to jump off a cliff and killing the player if he does so? Isn’t it really the same thing?

Not at all. In the first case, you treat him like a child; in the second, like an adult.

But then the player will have an unsatisfying experience with my design, which will reflect badly on me.

No, it doesn’t. Someone who leaps over the barrier rail at the Grand Canyon and jumps off the cliff knows perfectly well whom to blame for the consequences. True, there are immature people who try to blame others for their own errors, but in an interactive experience, the player can always go back, do it differently, and enjoy the improved results—thereby demonstrating beyond any doubt exactly where the original error lay.

If correcting the player is a bad idea, then aren’t all drama managers bad ideas?

Only drama managers that oppose the player’s wishes are bad ideas. The story-builder who’s creating a drama manager is a guide, not a parent. A guide’s role is less obtrusive than a parent’s. The feedback a guide provides isn’t a black-or-white prohibition or command; it is instead communicated in degrees of encouragement or discouragement. These degrees could be quantified, although most human guides don’t express their feedback in numeric form because most people are more literate than numerate. Storybuilders, however, work through the computer as their medium, and the digital medium is unquestionably
numerate. Therefore, the feedback drama managers should provide must take numeric form: a score.

What!? You want to turn stories into games?

No. In the first place, I’m talking about storyworlds, not stories, and storyworlds are larger, more abstract structures than stories. Second, “game” is an overly broad term that can mean many things to many people. I am not proposing to transform
King Lear
into Doom or
The Odyssey
into PacMan. The crucial factor to consider here is the wisdom of the scoring system.

Scoring systems, be they for exams, sports, or games, have always been constrained by the computational weakness of the people doing the scoring. For generations, teachers, umpires, and boardgame players had limited computational facilities; hence, scoring systems were clean and simple. The computer blasted away all those limitations, yet scoring systems remain as mathematically primitive as they were in the days of Monopoly. There’s no reason for basing scoring on body count or the number of jewels a player has collected.

To illustrate, I’ll build an example based on an Arthurian storyworld. The storyworld begins with Arthur married to Guinevere and ruling a land at peace. Arthur’s primary concern is his legacy; he would prefer to have an heir and a stable, happy kingdom on his departure. Unfortunately, Guinevere is barren, and Mordred, his illegitimate son, is completely unacceptable as an heir. Therefore, Arthur’s first priority is to prevent Mordred from gaining control of the kingdom. This priority is easily reduced to mathematical form:

IF (Mordred becomes king)
    THEN Score = -100
    ELSE Score = +100

Arthur’s second concern is for the kingdom’s political stability, which is based on the overall sense of community the other Actors share. A simple way to reduce this idea to numeric form is to calculate the sum total of everybody’s good feelings toward each other:

Here’s an important dramatic point: Death holds no terror for Arthur. He knows he has to die someday, and all that matters is leaving his kingdom in the best possible shape. There’s one other loose end he must take care of: Excalibur. The sword is his personal symbol of royal authority; if he cannot find a proper heir, Excalibur must be returned to the waters from which it came. Here’s how to express that idea:

IF (Arthur dies and Excalibur is returned)
     THEN Score = Score + 20
IF (Arthur dies and Excalibur is not returned)
     THEN Score = Score – 20

You could also throw in additional points for dealing with a possible romance between Lancelot and Guinevere: point losses for a romance, small point gains for a reconciliation with Guinevere after an affair. These considerations can just as readily be addressed in the
GoodFeelingScore
; after all, if the queen cuckolds the king, people are going to get upset with one or all of the three characters and take sides in the dispute. Ultimately, this will have negative effects on the
GoodFeelingScore
.

You might also want to include some consideration for the Grail search, but I shall ignore that factor here, just to keep things simple.

Here’s how this scoring system might evaluate a number of different scenarios:

A:
Arthur maintains a happy kingdom. Lancelot and Guinevere keep their pants on, everybody loves Arthur, Mordred revolts and is killed in a lopsided battle, Arthur dies in bed, and Lancelot throws Excalibur in the lake. Score: high.

B:
Arthur keeps the peace, but Lancelot and Guinevere have their romance. Arthur somehow holds things together, sending Guinevere off to a nunnery and banishing Lancelot. Mordred exploits the ill-feelings these events generate to raise an army and revolt, but Arthur defeats and kills Mordred. Excalibur slips beneath the waves on Arthur’s death. Score: moderate.

C:
Guinevere and Lancelot get it on, Arthur executes one or both of them, much fighting ensues, and Mordred reaps all the benefits, kills Arthur, grabs Excalibur, and seizes the throne. Score: negative.

The point of this example is that the scoring system provides strong motivations for the player to behave in a manner consistent with your aesthetic goals, yet it doesn’t mandate or prohibit any behavior. The player of this Arthurian game must pursue your intended goals to gain points. People can accept the impositions of a scoring system, but arbitrary constraints on behavior don’t go down as well. You can build scoring systems to reflect any dramatic consideration; complexity is not a problem with computers. The only constraint is your discomfort with algorithms.