Re: Olympia: Stacking and Visibility

From: dweinste@hal.gnu.ai.mit.edu (David Weinstein)

Date: Thu, 11 Jun 1992 21:47:22 +0000

Carl Edman writes:
#David Weinstein writes
#> The discussion has concerned the degree to which people
#> (mostly) would be  visible when either in a large stack, or
#> a crowded location. Proposals have ranged from entirely
#> invisible to everyone can be seen. I would propose the
#> following:
#>
#> The base chance of being seen in any location is the unit's
#> percentage of the location population multiplied be
#> some number to make it reasonable (perhaps 5?).
#>
#> Example: A Unit of 20 men in a crowded bazzar (1000 people)
#> has only a (2% x 5) 10% chance of being noticed.
#
#Theoretically a good idea except that it wouldn't work.
#All you would have are people LOOKing some ten times every
#day to make sure they see erybody. Or why not a hundred
#times ? Write a little script to strip out the redundant
#information out of the LOOKs and you have circumvented
#your proposal completely for free. (That is free to
#everybody except Rich who has to use up huge numbers of
#cycles).

First, do LOOKs take time? If not, should there be a limit on how many
could be done in a single day?

If I had someone whose sole job was to stand in the bazaar and keep an
eye out, I'd hope he'd spot a unit of 20 men. But he might not spot 1
(and by the numbers above a single unit would have a %.5 of being
spotted if the stealth and observation modifiers evened out).

If looks were limited to (say) 10/day, that would mean that the odds
of even a single dedicated person spotting someone moving through
the bazaar are 1 in 20. Not unreasonable. Add in the modifiers based on
Stealth and Observation skills (which brings these back into play), and
you have a reasonable model of reality. And it preserves the usefullness
of hiding in crowds, and the relative ease of spotting even the most
stealthy in an almost empty tavern.

#But then you say, why not just have a flag which says if one
#unit missed another ? First of all, because a simple
#minded implementation of such a flag would cost many
#megabytes of storage. Second of all, because it is not
#clear when you should clear this flag again.

No, I don't think you should have a flag. I think instead that
frequent looks *should* increase the chances of spotting someone;
but that they should either take time or have the time involved
by simulated by a limit on looks/day.

#Having things fail or work based on random chance always
#has the possibility of giving you trouble in such a game.
#In the end I don't doubt that you could come up with
#reasonable answers to these questions, but it is also
#very likely that such reasonable answers would require
#more effort than it is worth, particularily compared
#with such relatively simple solutions as have been
#proposed by others here.

You need a degree of random chance to make it worthwhile. Otherwise
everything is deterministic. Your observer will always (or never)
spot someone. And that is inaccurate. The problem with the other
proposals is that they remove the usefullness of stealth if you
can hide someone in a stack, or hide your entire force simply by
splitting them up into smaller groups. And that simply is unrealistic
and a gaping loophole. (Aha, there are 400 peasants, and 600 soldiers,
but you'll never notice the soldiers because they are cleverly hidden
in groups of only 10) or (Aha, I spot you lowly thief despite the
10,000 people at the Coliseum and your sneakiness).

This model is relatively straightforwards, and has the advantage of
preserving the usefullness of both stealth and observation, while
also keeping the ability to hide in numbers.

Were it to be made *more* complex, each look would decrease your effective
ability to hide (if you are poking around, others can see you more easily)
and possibly even result in diminishing returns on your looking (someone
who is obviously looking around will be avoided), but these are not
necessary.

I should note that this also would provide yet another niche for player
enterprise; high stealth high observation spies or "hidden guards", high
stealth couriers, and so on. Moreover, the city guard (or whoever has
decided to protect the city), may recruit both high observation high
combat types (to serve as obvious deterents) and high observation
high stealth types to keep a more careful look around. A player
with a network of observers can become quite wealthy selling information
as to various whereabouts, and so on.

--Dave

--
Dave Weinstein
Internet: dweinste@gnu.ai.mit.edu

References

Olympia: Stacking and Visibility
dweinste@hal.gnu.ai.mit.edu (David Weinstein)
Thu, 11 Jun 1992 16:24:05 +0000
Re: Olympia: Stacking and Visibility
cedman@714-725-3177.nts.uci.edu (Carl Edman)
Thu, 11 Jun 1992 20:29:14 +0000