Quote Originally Posted by TheAmishPirate View Post
I'm running into holes in the Crisis dialogue design. Specifications are vague on the following:
  • The player alternates between picking red and blue options. Does this just get them nothing?
  • Is there still a red/blue/white option, regardless of which option is picked previously?
  • The player picks primarily white options. Is this is a fixed result, or does this randomize? (My vote is fixed result. Dialogue trees with random results are bad.)
  • Will the red/blue/white option change depending on which Mr. Johnson you're facing?

I'm going to give this some thoughts myself, but what did you have in mind?

EDIT: After pacing about for a bit, here is how I envision it:

Imagine you have two buckets; a red bucket, and a blue bucket. Over the course of the Crisis negotiations, you will be filling up these buckets based on your responses. Blue answers will fill up the blue bucket, red answers will fill up the red bucket, and white answers are wild cards; they can fill up one bucket, both buckets, or neither bucket. At the end of negotiations, we look at the buckets. If the blue bucket is sufficiently filled, you get some advantage in board position. If the red bucket is sufficiently filled, you get information on Mr. Johnson.

I'm thinking that negotiation results may not be a binary "reward/no reward" thing. Perhaps each bucket has different levels of fullness, and the quality of your reward depends on how full the bucket is. For instance, if a player hops back and forth between red and blue answers, they might fill both buckets halfway, and their rewards are halved as a result. Maybe they get only a bit of board position, and only one profile is knocked off the list. This would make branches with differing depths more difficult to balance, as the player would have more opportunities to fill either bucket, but that's an issue of implementation rather than design.

EDIT 2: I'm getting looming dread on the design of the dialogue tree itself. Specifically, how we can format this data in such a way that it's easy for us to manipulate and visualize what it will do. It's too big a thing to worry about for the prototype stage, but it's going to be in the back of my mind as we're designing how this all is supposed to work.
- If you go all blue 100% you get outcome #1.
- If you go all red 100% you get outcome #2.
- If you do any other combination you either get outcome #3 or outcome #4.
- Outcome 1 is better board position. Outcome 2 is Johnson info. Outcome 3 is both. Outcome 4 is neither.

The thought process is, you can play it safe for board, you can play it safe for info, or you can roll up your sleeves and say 'I got dis' to your people.

There are 8 Johnson personalities. The first crisis will reduce that to a potential 4. The second crisis to 2. The third crisis to 1. Each dialogue scene will be written from the perspective of trying to force a wedge between those potential Johnsons - i.e. if he's an AI he'll respond like this and if he's a drunk he'll respond like this. I'll do a few different clusters of these divides - one cluster might be Johnson 1-4 vs 5-8 in round one, the second one might be 1,2,5,6 vs 3,4,7,8 in round one, etc. Keeping it 'paired' somewhat means that I can save a bit on writing - round 1 might have the widest range of Johnsons but if round 3 is always 1v2, 3v4 etc then it'll cut down on the exponential writing curve thing.

Strategically I'm going to try and make sure it's worth going through a round or two of info even if you later take a round of position because it'll still rule out at least a few potential endgames - and in a strategy game knowing which corners you can cut is the key to victory.

Does that sound vaguely sensible?

Good robbits! I like the variation between the outer metal plating and the inner bits.