Re: distance-delayed communication (was Re: Ongoing Strategy Games.)

From: cedman@golem.ps.uci.edu (Carl Edman)

Date: Wed, 14 Jul 1993 04:50:31 +0000

In article <troyer.742521923@cgl.ucsf.edu> troyer@cgl.ucsf.edu (John M. Troyer) writes:
   bruno@cerberus.csd.uwm.edu (Bruno Wolff III) writes: I have toyed
   >  with rules implementing, something like the above, where turn
   >  reports and/or orders are delayed based on how far away the
   >  units are from the capitol. I kind of came to the conclusion
   >  that this might not be a very playable system. To make it work
   >  you need a way to have workable contingency orders in outlying
   >  areas. This is not going to be easy to do well.

cedman@golem.ps.uci.edu (Carl Edman) writes:
   >You are right.  Contingency orders are absolutely essential under such
   >a system.

The problem with delayed orders is that your contingency orders
   quickly reach the point of AI.

Either the player can 'program' his units in some sort of language,
   in which case what you really are playing is some sort of robot
   gladiator program,

or the AI is in the program, and you have to make the tactical
   decisions for the human player,

or the AI has millions of user-settable parameters, and the
   interface becomes unmanageable.

Again you are right.  (Or rather, this time you are right.  Last time
Bruno was right).  Making remote units just as clever as local units
which are commanded directly is indeed a formidable task which can
probably not be even approximately done without requiring quite
programming on the part of every player.  And while that would make a
game I (and a couple of other similiarly perversely inclined players)
would enjoy, massive participation which I think is essential for a
great PBeM is unlikely.  However, I do think that a compromise can be
found in which makes remote units react in a reasonable manner to most
predictable mishaps.  And maybe that is all that is really needed.

In particular, Epos uses ON statements.  To understand them you have
to know that everything which happens in Epos generates a descriptive
string, a so-called 'event'.  Every unit which could reasonably
observe whatever happens adds this string to its history.  (It is
these strings which make up a turn report when they manage to
propagate back to the players main unit).  But in addition these
events also double as triggers for canned actions by units.  For
example, the unit Welcoming_Comittee (located in the city of Chardia)
might have received the following ON statement at some time.

ON "?guest enters Chardia."
DO
   TELL ?guest "Welcome to the beautiful city of Chardia !"
   GIVE ?guest 1 gold
DONE

Now assume that the unit Ganor enters Chardia.  The generated event would be:

Ganor enters Chardia.

If Welcoming_Comittee sees Ganor enter, the ON statement will be
triggered (with the substitution ?guest=Ganor).  As a result the unit
Welcoming_Comittee will execute the following commands.

TELL Ganor "Welcome to the beautiful city of Chardia !"
        GIVE Ganor 1 gold

Such ON statements are not turing equivalent, but I've found them to
be surprisingly flexible, easy to understand and adaptable to
virtually every game situation.

You send off a scout in an empire-type game. Should he go till he
   hits a coastline? a river? an enemy? any other unit? Should he
   explore some rectangle methodically? Should he follow the coastline
   or return right away? What should he do if he meets someone else?
   Attack, hide, run away, offer diplomatic relations?

Games with a complex and detailed geography (like eg. Empire) make
programming intelligent units hard.  Fortunately games like Olympia or
Epos have topological graphs as world maps.  That makes most of the
navigation problems easy to handle.  What to do in case of contact,
can be solved by rather straightforward ON statements e.g.

ON "?evil_attacker attacks Harmless_Scout_From_Chardia"
DO
   MOVE Chardia
DONE

You send off an army to a far-away target. Are you attacking a
   city? A hex?  A particular unit? What if one of your allies has
   taken the city by the time you get there? What if your opponents
   pass you on their way back to attack your own cities? Do you stay
   there after you're done? Come back?  When do you send a courier for
   reinforcements? Should your forces split or stay together?

Such complicated decisions can of course not be handled by a dumb
remote unit.  No, if you want to go conquering, you better go yourself
with your army (or find another human player you can trust and have
him command the army).

You get the idea. Or am I way off base here?

No, not at all.

One other thing: If you are using some travel-limited command
   system: if it takes N turns to get the battlefield, orders should
   take 2N turns (there & back) to propagate. But keep in mind that
   someone on the scene will probably email the general who is N turns
   away, thus lowering propagation time to N turns (the time it takes
   to send new units there) and semi-defeating your 'realistic'
   communication system.

Well, if you have an ally on the scene, you probably want to give him
command of the local army anyway, thus completely "defeating" the
system.

Do you then ban email communication between players and force them
   to use in-game couriers to communicate? (Possible, but difficult,
   yes?)

That would be nice.  Unfortunately it is completely unenforcible.  So
even if most players played by the rules, they would only benefit a
number of almost undetectable cheaters.  I'm afraid that instant
communication via email between players is just one unrealistic point
we'll have to live with.

wrt Bruno's comment above. Another possibility is to have a movable
   'central command center' instead of command emanating from a fixed
   capitol, so that distant wars can be executed, at a cost of leaving
   the home front unresponsive.

Absolutely.  The 'command center' in Epos is just simply the player
character himself who certainly can command his armies in the field
himself if he wants to make sure that things get done right.

Carl Edman

References

Re: Ongoing Strategy Games.
cedman@golem.ps.uci.edu (Carl Edman)
Thu, 08 Jul 1993 04:39:58 +0000
Re: Ongoing Strategy Games.
bruno@cerberus.csd.uwm.edu (Bruno Wolff III)
Thu, 08 Jul 1993 05:23:36 +0000
Re: Ongoing Strategy Games.
cedman@golem.ps.uci.edu (Carl Edman)
Thu, 08 Jul 1993 17:38:21 +0000
distance-delayed communication (was Re: Ongoing Strategy Games.)
troyer@cgl.ucsf.edu (John M. Troyer)
Tue, 13 Jul 1993 00:05:23 +0000