I was reading the above paper – which has the same title as this blogpost – by Ely, Frankel, and Kamenica, and have found it interested. Although, it doesn’t explicitly mention our industry it seems like it could be as useful for computer games as anything else.
The game I’ve made that would bear the most similarity to a game I’ve made would be SquareNet. However, that was a game made deliberately suspense-free. Since, the game ends whenever the player wants their can be no suspense since everything is in the player’s power. Saying that, there is still suspense in taking the individual shots. This suspense is produced almost-exclusively by the vicious ‘1’.
The part of the paper that struck me most was the notion of suspense working due to different activity phases; there is suspense in flipping the coin because the flip is informative and then comes the reveal, however not in generating a random number since only the reveal exists. In a turn-based RPG, the random number brings suspense since the first phase is determining one’s target and attack. This suggest a strong discontinuity between one attack with one enemy and going for two slightly different attacks with one enemy.
That is the case for the decision contributing to suspense. One thing that this paper makes me confront is that the resolution must contribute to suspense. That’s something that I balk at. SquareNet was completely determinate even with regards to the AI. A little while ago I made a turn-based RPG where everything was determinate. With single-player games, when you can’t rely on hidden information -for example, exploration -, its a case of no randomness no suspense.