[MUD-Dev] Re: How to handle log-outs in a totally dynamic world.

John Bertoglio alexb at internetcds.com
Sun May 10 22:48:11 New Zealand Standard Time 1998

From: Ben Greear <greear at cyberhighway.net> Date: Sunday, May 10, 1998 10:01
Subject: [MUD-Dev] How to handle log-outs in a totally dynamic world.

>As I've posted before, I'm working on a space game which will be constanly
>moving foward.  In other words, no static world to reboot to, reboot to
>the last saved image (a whole other can or worms!).
>I am going to try the model where the players are always in the game,
>even when logged out.  There will be safer spots than others to stow
>your craft and belongings, but nothing is guaranteed.

The first thing to address is how this feature will benefit game play or
further the goals of the game world and structure?

What does "always in the game" mean? A lot of design decisions depend on
what scale the players are functioning in. Live on board a space station or
large colony ship would create the need for lots of AI (perhaps user
programmed/selected). The key here would be to make it worthwhile for live
players to interact with logged-off players being "run" by AI routines.

If the game is scaled at a ship to ship/company to company level, the needs
do not change but behavior modeled does. Now you need routines to deal with
exceptions. Example: You log off giving your space trader captain orders to
to from planet a to planet b, buying selling certain commodities in each.
You could come back in a few days of realtime and find quite a nest egg
waiting. (This, BTY, is something we are using in AR. Characters use
off-time to study and inprove existing skills. There is no risk in this
activity except the normal risk inherent in skill improvement.) On the
other hand, if the trader was attacked by pirates...you could log into a
"you are dead" message...tough thing to pull off, I think.

>Another issue is that I want politics and the economy to be a live
>thing (In other words, build an empire, a giant Conglomerate, or
>both, or none...)  I will have NPC empires and corporations, but I hope
>that these will come to play a minor role once the game matures for
>a couple of months.

See Tradewar and similar BBS-type games. It would be cool to build NPC
organizations which can be bought into and eventually controlled by player
run-syndicates. Attempts to model a steady-state economy are very
difficult. However, there a many models of trade simulations out there. I
think you will find the need for NPC run companies will continue as certain
trades are found to be unprofitable. Live players will gravitate away from
those activities unless very sophisticated market balancing is in place.

>So, how to pull this off?  It needs to be fair, and at the same time,
>just because one is logged off shouldn't mean they are invulnerable
>or invisible to the game.

A question here is whether "fair" is a reasonable goal at all for a
competitive game. Was it "fair" to pit NBA All-Stars against the world in
Olympic basketball? Fair is a goal which is incompatible with your stated
goal. Turing a player, a ship and/or a intersteller business over to AI is
not something from which fairness will spring. It might be a interesting
world to play in, but not because it is *fair*. I agree that is difficult
if when modeling a world, certain things happen because a player is not
around. It is not very satisfing to attack a space ship only to be told
"Your laser blasts have no effect because the ship is not currently
controled." A game play result is 12 yearolds who cut school will have huge
advantages in the world because they can learn the AI and plunder offline
characters one-by-one. All in all, a real challenge.

>My current plan is to allow them to set certain flags and attributes
>to determine how they react.  These will be suggestions only, and the
>server will be the final arbiter.  If someone plays an asshole of a
>character, then I'm not going to let them set attributes to be a totally
>nice guy when logged out....

An interesting variant would be to only allow the player to interact
through the game interface. This would make it much more difficult to
determine if a player was live or memorex. The player would record a series
of macros for various situations and play them back. Only in some areas
like a bar or club would free chat be possible. You might allow only subtle
changes in behavior profile once the player had logged enought time in a
particular. Example: Say the character gets certain benefits from and
aggressive (nearly abusive) stance with vendors. However, that behavior is
also more likely to provoke an assualt from an NPC. Trying to reset the
profile to a much nicer stance (to make sure you are alive when you log
back in) might be greeted with "Zug's personality can not change Commercial
Agressiveness that dramatically in such a short amount of time. Perhaps you
would like a PAT (Personality Alteration Training) answer to make this

We have done a similar thing by making a number of areas controled by
various character profiles. This allows characters to build quick responses
to standard situations. NPC's use the same profile engine which lets them
show reasonably well developed personalities and PC-like behavior much of
the time.

>Basically, I'll use the same AI engine for the logged off characters
>as I would for the NPC characters.  Might tweak things a little, but
>hopefully I won't have to.

>I would be interested to hear other's opinions and experiences on this

>Ben Greear (greear at cyberhighway.net)  http://www.primenet.com/~greear
>Author of ScryMUD:  mud.primenet.com 4444
>MUD-Dev: Advancing an unrealised future.

John Bertoglio

MUD-Dev: Advancing an unrealised future.

More information about the MUD-Dev mailing list