[MUD-Dev] RE: Understanding Simulation

[shren]

I'm trying to engage my brain well enough to respond to all of this.
We'll see how I do.

On Fri, 11 Oct 2002, it was written:
> [Shren]

>> In my ecology simulation thought experiments, a creature becomes
>> tougher as it becomes more rare.
 
> I think this is a great kind of solution to these problems (I
> would compare it to the suggestion made earlier in the thread, to
> increase the cost of finding prey - a different way of increasing
> the cost of kills).
 
> I have some half-baked comments on the implementation & results.  In
> the below I'm assuming something like logistic growth.
 
>   -- OK, so as population decreases, the cost of making a kill
>   increases, and so we assume the number of kills 'bought' by
>   players (e.g., the kill rate) decreases. Let's assume that the
>   increase in difficulty and/or difficulty when there are no kills
>   are set sufficiently high to prevent extinction entirely.  

I have no problem with making the last monster immortal.  If you can
get a good migration algorithm, I've always thought it would be a
good idea to have new monsters of a type literally spawn out of old
monsters of a type, so the monsters will tend to occupy regions,
even if those regions shift about.

>   In this case, it seems that things will reach an equilibrium
>   when players hit a difficulty that is just tolerable. I wonder
>   whether this is acceptable (e.g., a system to ensure that
>   everyone is fighting just barely tolerable monsters).

Well, say we have a statistic called race karma.  The more of the
karma you have, the more powerful you are.  Whenever a member of
your race dies, it's karma is distributed, with some degree of
randomness, about the other members of the race.  Whenever you
reproduce, you split your karma with your offspring.  Compare to
algorithms that determine termination of distributed algorithms.

So, the race as a whole over time would get tougher over time given
a decreasing population, but the exact power of an individual member
of the race is something of an unknown.  Also, if monsters wander,
then tactical situations are always different.

>     Objection 1. I neglected population size and per-capita birth
>     rate as parameters. As these increase (e.g., as the supply
>     increases), the prices can get lower without hurting the
>     populations' ability to regenerate at all. Thus the
>     tolerability of the fighting is independent of Shren's idea,
>     but birth rate is very important to this independence.

And birth rate is very tricky.  Birth rate and migration rate should
vary from monster race to monster race.

>     Objection 2. Local fluctuations in supply (artificial or
>     natural) are surely going to create fluctuations in the
>     difficulty of the critters.

Something else to consider is giving the monsters a 'triviality'
level.  The 'trivialness' of a monster ranges from 1 (for named
demons and other ultrabaddies) to 10000 (for, say, kobolds and other
cannon-fodder monsters).

Keep a decaying (say, loses 2% every hour) count of the number of
times a race has been killed (Decaying kill count, or DKC).  Use
DKC/trivialness as an additional boost to monster difficulty.  Thus,
you can kill lots of kobolds, for days, and kobolds will never get
much harder (except as population dictates), but killing but one
named demon is going to boost the difficulty of the race mightily,
making the slaying of a named demon epic.

Say I kill one thousand kobolds.  1000 kills over 10000 triviality
is .1, a rather trivial boost to toughness.  Say I kill 10 named
demons.  10 kills over 1 triviality is 10, a very significant boost
to toughness.

>     Objection 3. Just barely worthwhile fighting isn't different
>     from the status quo!!
 
>     Objection 4. In some games it won't matter whether fighting is
>     worthwhile, because there are other things to do.

Migratory monster populations, varying birth rates, and dynamic
'mosnter territory' actually creates work for explorers.

>   -- Also, it seems that those who kill the most will set the
>   difficulty of the NPCs. That's not good if it is important that
>   newbies get to kill things. Even if everyone had the same
>   ability to kill, the people who are willing to pay more are
>   going to contribute substantially to setting the price
>   (difficulty). So we might worry about casual players getting
>   lost.

A very valid note.  Usually in my games I consider reducing the
magnitude of power difference between newbie and veteren character
power, making group tactics more important than raw power.
Triviality, above, helps this a bit, as the newbies can chase
kobolds while the masters hunt the named demons.

It's still not a very newbie friendly game, however, as the monster
populations move about, and you've no guarantee of finding kobolds
instead of orcs, trolls, ogres, or some other henious nasty that can
pulp you like an orange.

Another note to consider here is that I usually try to incorporate
permadeath into my games, and reduce the 'mean time to ultimate
power' (like the 7 gmed skills in UO) to perhaps 12 hours of
playtime.  Permadeath has all sorts of interesting effects, like
making it riskier to do things for the sake of discovering what
happens.

To lower the advantage of attacking first in permadeath games, I
usually incorporate 'stance', fast regenerating hitpoints.  (See the
very recent discussion on Halo and the effects of Halo-style
regeneration on combat.)

>     Objection 1. If the people increasing the difficulty of the
>     NPCs were all hunters of roughly the same ability to kill, it
>     would be fair for them to set the difficulty. It's a solution
>     because all the people who need to be able to hunt the
>     monsters regularly are given the tools to do so, while
>     bureaucrats and thieves hunt at their own risk. That's fine if
>     you're willing to embrace this lack of substantial differences
>     between players in kill capacity. On the other hand, it's a
>     big problem if you have powerful old-timers and weak newbies
>     beating on the same populations. Very level-based games will
>     probably need to keep the partition of prey populations by
>     level (e.g., smurf village vs. Olympus).

Yes.  I usually think about interesting ecologies over player
survivability.  On one hand, this sucks for the players.  On the
other hand, I can picture a type of hardcore player who would enjoy
trying to survive in someone's ecology simulation - 'alone and
afraid in a world that isn't made for you' is very much the feel.

>     Objection 2. As casual players stop killing, the supply
>     increases slightly and the price decreases. Where this price
>     is absolutely seems, again, to depend on birth rate. Thus this
>     concern doesn't really seem to hold water - you solve the
>     problem for casual players by increasing the supply (e.g.,
>     cranking birth rates per capita or overall).

Consider this.  Usually the comparison is Risk vs Reward - a dragon
is of a fixed difficulty, thus fixed risk.  In this case the reward
determines the risk - monsters with better rewards will be higher
risks, as they are hunted to low populations and have a consistantly
high DKC/triviality.

>   -- I think a lot of the interest in the system lies in the
>   fluctuations of supply and demand. But this isn't compatible
>   with McDonald's style consistency in product. Some designers
>   want this, although I think it is not hard to make a case that
>   variable payoffs can be very compelling, particularly if the
>   fluctuations are mostly up above the limits of player toleration
>   rather than below them.
 
>   -- I wonder about the function of player kill rates
>   vs. difficulty.  It would be nice if it was a clean line, but it
>   probably isn't.  In some cases kill rates aren't going to be
>   sensitive to difficulty until you run up against kill
>   *possibility*. If you let players do better in groups then you
>   are working against the maximum effective group size that can be
>   mustered. It probably also depends on what kind of difficulty it
>   is (I won't even pretend to know how).

I haven't done enough actual code and testing to get into the bits
and bolts of scaling difficulty.

> In any case, this concern alone could completely destroy
> everything I wrote above while assuming a simple supply-demand
> kind of situation. Oh well ;)
 
>> Extinction becomes extremely unlikely, however, a creature can
>> easily be nudged into limited terrain.
 
> The use of space here is very cool - in effect you are rewarding
> players for spreading out. Lots of interesting side effects of
> this - for example, if things are stable then people might develop
> little hunting territories, but as populations decreased in some
> places people would move on, possibly into others' territories...

Perhaps the most interesting bit from my point of view is that the
monster difficulty rises as you compress the monsters into smaller
spaces.  If there are, say, 50 viable city sites and the players are
using one, then the monsters spread out over the territory of the
other 49.  As more cities get founded, the monsters get nastier and
encroach on the towns more often.  Thus, you have motivations both
to defend a city (you don't want monsters spreading through your
streets) and destroy one (each other city makes it harder for your
own city)

> This and PVP makes for some interesting fights over resources ;)
 
>> Then again, however, I strive for dynamic ecology instead of
>> realistic ecology.
 
> The demand for "realistic" ecologies is what I perceive to be the
> biggest reason for the difficulty in using ecologies in MUDs.
 
> As if that weren't enough, I don't think that the typical idea of
> "realistic" is really particularly realistic (as I argued earlier,
> extinction is a great example - extinctions do occur in the real
> world, but the real world has more than four species as well, so
> each extinction, while sad, is for most purposes a drop in the
> bucket).

I'd be curious to see the cubic feet per player for each of the
different games and compare them to the real world, as well as
travel rates.  Extinction is highly likely because no place is
truely remote - unless you stop it.

Ecology is almost the wrong word.


_______________________________________________
MUD-Dev mailing list
MUD-Dev at kanga.nu
https://www.kanga.nu/lists/listinfo/mud-dev