Post by TomL on Feb 26, 2013 10:16:28 GMT -5
Cory's thread here startradersrpg.proboards.com/index.cgi?action=display&board=rules&thread=3364&page=1#37777 about superlatives for describing wild zone ratings ("safe", "risky", "unbelievable") got me thinking about how does a game engine model a character's subjective feeling about a possible encounter.
TLDR: instead of presenting to the player absolute values for danger level, how about the game engine keeps a running tally of character's encounters and describe a zone as more or less dangerous than other previously encountered?
For the sake of discussion, let's just talk about wild zone danger levels, currently model as from 0 "friendly" to 25 "toxic radioactive magma world". Consider this:
A player's newly created character, steps up to a planet's wild zone, which the game describes to the player as "risky", which we know in the backend is 7 on a scale of 0 to 25. How would a newbie character know that this wild zone is "risky"? As far as they know, this is their first wild zone encountered, ever. Perhaps because they have no frame of reference yet, this wild zone might appear to the character as "unknown danger level, no prior experience". The character jumps in to explore, the game engine weighs all variables (explorer_level=none, exo_scout_vehicle=none, scout_officer=nope, etc) and then informs the player that they just got trounced by the wild zone, badly.
Player limps away from planet, with the impression that this wild zone is "the most dangerous zone, ever".
Now the player's character has a gut feeling for wild zone to gauge other zone encounters. Does this zone I'm considering exploring look more or less dangerous than others I've explored in the past?
With enough history and experience, a character may be able to tell if a zone is slighter more dangerous than those experienced, versus very very much more dangerous than those experienced.
The thorough way of modeling this would be to store a record for every zone encounter. Over time, the engine would have a histogram of zone encounters as a function of zone difficulty. Assuming a gaussian distribution, you'd have a nice bell curve, where the center of the curve would be the player's "comfort region" of zones of average danger level. Any future zone encounters could be compared against this curve: within 1 sigma of average, the new zone would appear "average". 1 sigma above average, "looks dangerous". 2 sigma above average: "quite dangerous". 3 or more sigma above average, "off the scale dangerous, do not attempt!"
A slightly cheaper way to model this would be to store the # of zone encounters and the sum of encountered zone danger levels so as to be able to produce a running average. So the engine can inform the player that a new zone looks more or less dangerous than average.
Or simply storing a player's wild zone difficulty encountered min and max and median. Then the engine can relay perceived difficulty ("more or less dangerous that usual") or whether is zone is "the most dangerous looking zone you've ever seen!"
Tom
TLDR: instead of presenting to the player absolute values for danger level, how about the game engine keeps a running tally of character's encounters and describe a zone as more or less dangerous than other previously encountered?
For the sake of discussion, let's just talk about wild zone danger levels, currently model as from 0 "friendly" to 25 "toxic radioactive magma world". Consider this:
A player's newly created character, steps up to a planet's wild zone, which the game describes to the player as "risky", which we know in the backend is 7 on a scale of 0 to 25. How would a newbie character know that this wild zone is "risky"? As far as they know, this is their first wild zone encountered, ever. Perhaps because they have no frame of reference yet, this wild zone might appear to the character as "unknown danger level, no prior experience". The character jumps in to explore, the game engine weighs all variables (explorer_level=none, exo_scout_vehicle=none, scout_officer=nope, etc) and then informs the player that they just got trounced by the wild zone, badly.
Player limps away from planet, with the impression that this wild zone is "the most dangerous zone, ever".Now the player's character has a gut feeling for wild zone to gauge other zone encounters. Does this zone I'm considering exploring look more or less dangerous than others I've explored in the past?
With enough history and experience, a character may be able to tell if a zone is slighter more dangerous than those experienced, versus very very much more dangerous than those experienced.
The thorough way of modeling this would be to store a record for every zone encounter. Over time, the engine would have a histogram of zone encounters as a function of zone difficulty. Assuming a gaussian distribution, you'd have a nice bell curve, where the center of the curve would be the player's "comfort region" of zones of average danger level. Any future zone encounters could be compared against this curve: within 1 sigma of average, the new zone would appear "average". 1 sigma above average, "looks dangerous". 2 sigma above average: "quite dangerous". 3 or more sigma above average, "off the scale dangerous, do not attempt!"
A slightly cheaper way to model this would be to store the # of zone encounters and the sum of encountered zone danger levels so as to be able to produce a running average. So the engine can inform the player that a new zone looks more or less dangerous than average.
Or simply storing a player's wild zone difficulty encountered min and max and median. Then the engine can relay perceived difficulty ("more or less dangerous that usual") or whether is zone is "the most dangerous looking zone you've ever seen!"
Tom
Perhaps something for ST2? Hopefully possible for ST. 
