Self aware replay game theory

Self aware replay game theory

Imagine a game with an usually broad series of outcomes with a huge number of in game characters.

Perhaps at the end of the first playthrough you reflect on the outcomes you experienced. What if you were to simulate the game being played by an AI that perfectly emulated the decision making process you yourself employed when you played the game.

Let’s call that emulation of you, your personal emergent behavior.

What if the game included an end game set of god-mode tools to understand and edit that personal emergent behavior.

Decisions like dialog options, who to fight, who to team up with, who to pursue romantically were all predictable and predestined according to the AI behavior that emulated your playthrough.

Now in editing the core structure of your personal emergent behavior it resimulates the entire game.

Again let’s imagine a game with an incredible amount of disparate outcomes such as some characters die or live, love or hate the player character, etc.

Imagine in editing yourself being simulated in the playing of the game to make decisions according to different rules or priorities.

Imagine if you could then preview the outcomes of that alternate version of yourself by any criteria you could think of.

For example you change your personal emergent behavior ai to prioritize saving lives and the outcome does indeed assure none of the characters die.

What if you could then replay the game, BUT the simulated decision making version of yourself that prioritized other characters surviving would tell you what to do in order to assure that result.

Now in your guided playthrough there is no certainty that every single suggestion is required to achieve that outcome, but any decision could be.

I think that would be a fascinating way to replay a game.

Comparing all decision making priorities with all desirable or undesireable outcomes and experiencing the game pursuing or avoiding those paths.


Leave a Reply

Your email address will not be published. Required fields are marked *