10/01/13 - Fight or Flight Inspired Test

While writing the proposal just before Christmas I referenced a point i'd noticed as part of the GTA 4 analysis back in October about fight or flight reactions. When attacked the background NPCs reacted in different ways. In the original footage i took the "older lady" was the only one who seemed to take offense to being pushed and looked like she was telling me off almost in a "respect your elders" stereotypical manner where as many of the other younger NPCs just acted as if it was nothing and carried on. the stereotype point might be more of an interesting side note than something i should really include as it would be better to point out the behavior of different types of NPCs than different types of one NPC. The NPC i chose to stalk later in the footage demonstrated that despite continuously being bumped into there was still no reaction until being hit lightly by a passing car.. multiple minors equal a major?

Over the next couple of days i'd like to use this as the basis for a test.

Things to consider

• Particular stereotypes of character help dictate reaction - grumpy old lady who will tell someone off but will run past a certain point of danger, gangster type who reacts badly to the smallest of aggression intentional or not. average citizen who will just brush off attacks no matter how continuous. do all members of a stereotype act the same way? film game play.
• Levels of danger taken into account. Small push has the same reaction as a big push, being punched, having a gun aimed at them.

Proposed outcome.

If an NPC had a list of reaction points that ranked the intensity of their response on a scale of 1-5 to commonly occurring interactions from the player. that could be a good final outcome to consider..

This test would look at the effectiveness of including a characters reactions to minor and major physical interactions. like being bumped into and being hit.

It could be done in more detail involving accidental and intentional interactions but that seems like it would be too complicated for a game company to include. 

Plan of attack

Tomorrow's post

• Pick  two of each NPC type, one female one male, minor interaction (push past), multiple minor (repeatedly push),  major interaction (punch) 
• Edit video together.
• Observational analysis
• Identify areas for improvement

Saturdays post

• Live motion reference
• Storyboards
• Begin tests

Monday's Post

• More tests
• Analysis of tests - self critique
• Analysis of tests - peer review
• Outcome, would it be needed for credible performance to include this. 

------------------------------------------------------------------------------------------------------------

Side note: I noted a point where the NPCs in dishonoured didn't react to each others presences resulting in one killing the other and acting as if nothing happened. this video has recently been brought to my attention displaying a similar NPC interaction that seems to have gone wrong.

------------------------------------------------------------------------------------------------------------

Things to do.

• Different sorts of NPC and their implications? will i only focus on one sort of NPC to improve? would limit research (needed). surprisingly don't think I've done a post on the different types and level of detail each would need.
• Finish Likert "initial reaction" test. 
• Explore the idea of a static reaction and one that can change over time. explore the idea of a selection of reaction guidelines in general. has anyone done it previously? How do they judge how an NPC will react already? Have i already looked at this briefly? Who knows... this post will tell me.