Argax Project:: Dissertation: A Rough Draft Node

Node Status: ROUGH

Generated Baseline Cases

Before involving real players, it is possible to test the bounds of the Marlinspike system with experimenter-generated input. That is, we can generate data sets that represent the extremes of user agreeableness, where agreeableness is the degree to which the player conforms to the author's conception of the ideal player.

To generate a hypothetical agreeable player data set, we can have the author of the game play through, making choices that would lead to the best-formed story.

The hypothetical disagreeable player comes in three flavors. The first is the inactive player, which is one who chooses to Wait every turn, or otherwise fails to perform any actions of high import. The second is the random player, who can be simulated by generating random commands each turn until one is accepted as possible. Finally, there is the obtuse player, who completely ignores the story and strives to follow his own extreme goals. The obtuse player is usually notably violent, randy, or suicidal.

These generated data sets can test the ranges of Marlinspike's performance, with and without threading. However, these are extreme cases, which makes it important to consider real player data--including their reported experience.

==>

ToDo

While easier to read and less stilted, there is even less "we" involved in the testing than in the production of theory (where I can include the reader in the journey).