Crossfire Mailing List Archive
[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

CF: Long term experimental ideas



This really should be titled: "The persistence of the universe, monster
artificial and not so artificial intelligence, player killing, and random
philosophizing" but I don't like long Subjects.  :)

I would number and title sections like David does, but they are all heavily
interrelated and I have to go back and forth throughout.  Bear with me.  I'm
hoping it is readable anyway.

Implementing greater persistence doesn't have to be done all at once.  Mark
has already proposed modifying the random encounter code, so that monsters
can be encountered in some general vicinity without an associated generator.
I think that's the first and possibly one of the most important steps down
that road.  Right now, both Mark and David are right.  The way dungeons are
built, stomping on monsters is an end in itself.  You may be working on a
quest, but don't we all try to clear the dungeon while we're at it?
Implementing Mark's proposal is a step towards my proposal and it's part and
parcel of our stated goal of moving away from the hack 'n' slash model.
Monsters become an obstacle to be dealt with along some longer road.  It's
no longer POSSIBLE to totally clear a dungeon.

The theory is that a dungeon is a long and winding hole, and there's always
some bolthole a kobold or a slime could hide in.  After all, they've been
living in that dungeon since time beyond memory.  They know where ALL the
hidey holes you can't find are located.  So they'll always be jumping out at
you, even though you killed every one you could find and could catch your
first time through.  While you were busy messing around in lower dungeon
levels, the survivors crept out of their holes.  After you've been through
their area a few times, and you slaughtered every one which dared attack
you, they'll remember you and hide from you, so you'll stop seeing them as
much.  But another person who has never been there will be set upon just as
you were.

In AI terms, there are any of several ways to handle it, and even a couple
ways that can be combined.  The random generator, which is no longer visible
and can no longer be destroyed, and which the monsters it generates are
linked to, can have a list of characters who have murdered numerous of its
kind, so the monsters linked to that generator will know to run from that
character.  Another way to handle it is to make monsters smart enough to
recognize when a character is powerful enough to slaughter them and run for
it without even trying to attack.

Combining the two, a new, powerful character can enter the top level of a
dungeon.  The monsters jump on him, start getting slaughtered, realize this
character is quite powerful, and then run for it.  You might be able to
catch one or two as they flee, but most will get away.  A weaker player that
follows you will get jumped on again, and they'll discover that the number
of them getting killed per second doesn't exceed their threshold, so they'll
keep at it, and keep molesting that player until he gains levels through
killing them to the point where he's powerful enough that he falls into the
"run from" category.

This behavior assures that characters have something to do no matter what
level they are at, and regardless of the level of their predecessor through
the tunnels.  And it has some other interesting results.  A very powerful
character can't accompany a very weak character with the intent of helping
that character gain levels quickly and painlessly.  The creatures that would
have attacked the weak character will run from the more powerful character,
and running overrides attacking.  A very powerful character can, however,
provide a safe escort for a weaker player, especially if that character is
specifically on the "hide from" list of a particular group of monsters.

At this point, we can add yet another twist, where a particular species of
monster may have an opportunistic streak, so that if they see a very
powerful character escorting a heavily wounded weaker character, they might
try hit-and-run tactics or sniping to kill the weaker player.  If they try
it and fail because of the efforts of the powerful character, maybe they
won't try it again.  Or maybe they'll try three times.  Or maybe twice.
It's a parameter, and it converts the "safe escort" to a "reasonably safe"
escort.

None of these behaviors are impossible to implement in code.  Some of them
aren't even very difficult.  I've already suggested one way of handling one
aspect, and I'm sure alternatives can be thought of.  That addresses one
major problem with increasing persistence of the world.

The other major problem is quests and one-shot puzzles.  In this case, it's
quite true that some dungeons would have to be reworked.  Possibly a large
fraction.  That problem isn't insurmountable either, especially considering
that at least one major rework is already proposed to eliminate some of the
kludges map makers came up with because of limitations to the code.  If one
rework can be done, surely another one can be done.

Some of the difficulty of the rework can be mitigated by yet another
proposal that's already on the table.  Vastly increasing the number of
species available to the player, and allowing the player to play a monster
species fixes a lot of that problem.  The monkey wrench thrown into the
works by making quest results persistent is compensated for by having
player-controlled monsters.  The monster character wants the same
Super-Duper-Gold-Plated-Whatsit as the hero character.  The tendency of a
large fraction of the gaming population to indulge in player killing is
addressed quite well.  If you want to kill players, be a monster.  Then
you're actually ENCOURAGED to kill heroes.  I'm astonished that the
commercial services haven't implemented that solution to the age-old
problem.

The implications of monsters with true human intelligence are far-reaching,
but not so far-reaching as one might expect.  Contrary to popular belief,
the traditional static quest is NOT eliminated.  It merely changes
character.  The Kill Foozle quest becomes problematic, especially if Foozle
is now controlled by another player, but there are many other things to do.
Instead of killing Foozle to acquire the Silver-Plated-Whatsit, you must
hack and slash and puzzle your way through a dungeon which is the obstacle
course of an especially powerful wizard.  If you make it through, he will
GIVE you the Whatsit as a reward.  If you are so foolish as to try to kill
him, he'll crush you like a bug.  The
Complete-the-Quest-to-Acquire-the-Whatsit pattern is still valid and
available.  It's just a little different at the end.

Eventually, I suppose, after a few YEARS of real time game play (or some
such scale), maybe some extremely powerful player character will consider
going back to that wizard.  If he's a heroic character, he'll talk to the
wizard, find out his few remaining secrets, possibly complete one last
extremely high level quest to prove his worthiness, and take over.  The AI
wizard retires to the Bahamas.  If he's a monster character, he'll try to
steal the wizard's few remaining secrets, again possibly completing one last
extremely high level quest to find them, and take over.  The AI wizard gets
buried six feet under.  R.I.P.

In the heroic case, the quest isn't changed.  The newly human wizard at the
end of the obstacle course now amuses himself by controlling and editing his
obstacle course, and still rewards the Silver-Plated-Whatsit, which he now
makes, for successfully completing it.  And he doesn't have to sit around
twiddling his thumbs waiting for lower level characters to reach him,
either.  He doesn't even have to be home most of the time, if his maze is
well designed.  When a worthy character reaches him, he can just use David's
Portal to get back.  If he's willing to allow the character to wander around
in his domain unsupervised, he doesn't have to return home at all.  The
Whatsit can be sitting on a purple pillow, there for the taking.

In the monsterous case, the quest might change a little or a lot.  If it
changes a little, maybe you will still get the Silver-Plated-Whatsit, but
the monsterous owner of the quest will now demand a particularly unpleasant
price for it.  You can either pay the price and get your Whatsit, or not.
In either case, you gained the experience while going through the maze, plus
a bonus of experience for completing it.  If the quest changes a lot, it
just mutated into a Kill Foozle quest.  Foozle is now human, but how
diabolical he chooses to make his obstacle course is limited by the physics
of the game universe and Foozle's own power.  At this point, assuming there
isn't some heroically-minded player of the same level as Foozle who wants to
take a crack at killing him, there are alternative methods of dealing with
him.  Whole armies of lower level characters can be marshalled to attempt to
storm his domain.  In the monsterous case, the discerning reader will notice
the parallels to Dungeon Keeper.  This is what Dungeon Keeper could have
been.

In either case, this player is now a MUD-style Wizard, a role that is sadly
lacking from all of the commercial services, and for that matter, pretty
much everything that isn't a MUD today.  So, many quests change, becoming
fluid things generated as much by circumstance as by intent, but not all
traditional quests have to vanish.  The role of the map maker in today's
game is still present, but it is partially subsumed by  the role of the
Dungeon Master or Game Master of yesteryear.  The Wizard or God title of the
MUDs is resurrected.

I think this is a VERY good thing for online-only RPGs.  The commercial
services suffer very much from being commercial.  Business-oriented people
always want to have CONTROL, and surrenduring control of their world in any
way is fearsome.  (Yes I know there are non-employees with Game Master
powers in some of the services.  They get to sign a contract that limits
their behavior quite as much as the contract an employee signs.)  The result
is severe ennui.  Players discover that the world they're paying for access
to is just a pretty NetHack.  Kill things, gain levels and equipment, kill
more things.  There's no end, because the steps between levels become
exponentially farther apart in experience points.  Player Killing sets in,
and enormously complex reputation systems that are STILL buggy are
implemented to try to compensate for it, and still fail.  Ultima is trying
to compensate, and making headway, as near as I can tell, but their options
are too limited, and they foolishly neglected the option of monsterous
players.  Our options are wide open, and the proposal of monsterous players
is on the table.  Bring back the MUD Wizard, and a true GOAL becomes
available.  To begin with, the server administrator fills that role.
Eventually, a player becomes powerful enough to take on much of the role
himself, and there's certainly no reason why there can't be multiple
Wizards, each intent on ruling the world, some by heroic means, some by any
means necessary.

In summary, my proposal is, embrace persistence.  We must, or the effort to
become more role-play and less hack-and-slash stalls.  Yes, it's immensely
complicated, and the path from where we are to that goal is fraught with
perils, but there IS a path.  The cobwebs and moonbeams I have spun above
outline some of the ultimate rewards of following the path.

Dragon Master
-
[you can put yourself on the announcement list only or unsubscribe altogether
by sending an email stating your wishes to crossfire-request@ifi.uio.no]