'Deterministic' RPGs
Martin Julian DeMello
statistics-dependent outcomes, embellished with suitable messages. This
would have to be made reasonably explicit, of course, so that the player
knows that, say, the reason he couldn't pick the lock was that he hadn't a
high enough dexterity to do so, and not that the lock was intrinsically
unpickable[1].
However, as was pointed out elsethread, the problem with randomising is that
you can simply keep trying again and again (saving and restoring if
necessary) until the roll of the dice favours you. A better solution would
be to make the process completely determinstic - maintain a set of statistcs
for the player, and allocate a 'threshold vector' to certain actions, such
that the player succeeds in carrying through the action iff all his
statistics exceed the corresponding thresholds in the action's vector.
Then (and this is the time-consuming part[2]) provide several routes to
solve the problem, each with a different TV, so that depending on his stats,
a player can work through at least one of them (or build up his stats until
he can). One could even extend the system to other physical characteristics,
so that, for example, being male could let you disguise yourself as a guard,
or being thin could let you slip through a window.
This has the RPG advantage of being able to play the game with varying sets
of abilities, while retaining the IF expectation that the outcome of an
action depends solely on the action and the game's state. (Perhaps
'expectation' is too strong a word - I know there have been uses of random
numbers in IF, but IMHO having to deal with them is tedious).
[1] or maybe not, if you take the viewpoint that, for that particular run,
the lock *is* unpickable, and that if you start with different abilities
you're essentially playing a different game.
[2] though having several winning paths is far from a new idea
--
Martin DeMello
Knight37
Dec 2000 18:59:38 GMT:
>One of the RPG/IF crossover ideas I would enjoy seeing is the personal
>statistics-dependent outcomes, embellished with suitable messages. This
>would have to be made reasonably explicit, of course, so that the player
>knows that, say, the reason he couldn't pick the lock was that he hadn't
>a high enough dexterity to do so, and not that the lock was
>intrinsically unpickable[1].
[snip]
>This has the RPG advantage of being able to play the game with varying
>sets of abilities, while retaining the IF expectation that the outcome
>of an action depends solely on the action and the game's state.
Yes, I think someone should do this. You are elected. ;P
--
Knight37
"Origin could be a great company if they would learn to stop talking to
thier customers with "My loyal customers, signed your humble servent" yadda
yadda after slipping people Mr. Winky." -- Falsadoom on csipg.rpg
Sean T Barrett
>A better solution would
>be to make the process completely determinstic - maintain a set of statistcs
>for the player, and allocate a 'threshold vector' to certain actions, such
>that the player succeeds in carrying through the action iff all his
>statistics exceed the corresponding thresholds in the action's vector.
This is an important point. Consider that in most commercial
fighting games and shooters, there's no "to hit" die roll that
determines whether you hit the other guy or not--it's determined
by the simulation, which is sufficiently deeply simulated that
there's no real place for randomness. Some games may use randomness
to determine the amount of damage done, however; but hit points
are *not* a deep simulation of a creature's health and stamina.
Let me see if I can come up with a semi-complete list of reasons
why you might want to use randomness:
1. to determine the outcome of a chaotic process that is unpredictable
at the level of simulation of the game--for the sake of realism
In this case randomness represents our incomplete knowledge
(or simulation) of a process, or our incomplete knowledge or
ability to represent the skills. Two football teams might have
quantifiable abilities, but it's still unpredictable except
statistically which one will win. Unless you simulate the game
play-by-play, randomness for determining the outcome makes sense.
By analogy, two people swordfighting is also an unpredictable
thing, unless you simulate every swing, and random outcomes
make sense, at least if your goal is realism. (If your goal isn't
realism, you can have the same person win every time; in many
games, you never have the same battle more than once anyway, so
the realism isn't actually improved by randomness.)
2. to add depth to gameplay by making player choices more meaningful
Without random outcomes, given actions either always work or
always don't work, which can result in an optimal path through
the game. By making outcomes random, players have to face
trade-offs; e.g. taking a lower payoff action with lower risk
in a certain situation where failure is more painful. For example,
if the player always has two ways of crossing a gap--jumping or
using a grappling hook--one of which takes more time but is
guaranteed, and one which takes less time but might fail, the
player will stick to the safer one when jumping across a pit
with spikes at the bottom, but take the riskier quick one when
jumping across the stream--although this could change if the
player finds herself under time pressure.
3. replayability
Some non-IF games use randomness to increase replayability;
of course you can get replayability by having player-directed
branching instead of randomness-directed branching, but randomness
can be better since it will branch without the player intending
to, guaranteeing the player gets a different experience even if
the player tends to automatically do the same thing as last time.
Reasons not to use random outcomes:
1. can interfere with predictability and hence intentionality
For those of us players who like intentionality (I find it
deeply immersive, but not everyone cares about being immersed),
randomness interferes; the satisfaction of intentionality is
forming a plan and having it work; making the outcomes random
makes it luck whether the plan works or not.
2. try until you don't die
Some players can and will save/undo until they succeed. Some
subset of those players are ruining their own game experience
but can't help it. You can combat this by generating your
random numbers early; for example, Baldur's Gate seems to
generate your new HP for leveling up either at the time you
get the experience needed to level up, or when you gain the
previous level; thus you have to restore pretty far back to
change the outcome.
3. branching interferes with storytelling
In a p&p RPG, the DM knows exactly what series of events
has transpired in the past, and can alter forthcoming events
to improve the game balance and the story. An IF author has
to anticipate every possible path through the game and make
sure they all are well-balanced and tell good stories--well,
if the author wants all players to have an equally good
experience.
SeanB
Paul O'Brian
> Let me see if I can come up with a semi-complete list of reasons
> why you might want to use randomness:
>
> 1. to determine the outcome of a chaotic process that is unpredictable
> at the level of simulation of the game--for the sake of realism
> 2. to add depth to gameplay by making player choices more meaningful
> 3. replayability
Let me add another:
4. to deepen atmosphere
"Enchanter" is an example of a game whose atmosphere is improved by
the various random messages it displays from time to time -- it's
been quite a while since I played, but I seem to recall they were
things like "A leaf skitters across the path and over your feet."
I admired this very much when I played it, and have incorporated
the same sort of thing into both of my games. Of course, you have
to be careful -- a LASH tester, for example, noticed that you
can get the "My foot slips on a patch of ice" message even when
the PC is standing in a pit it just dug.
--
Paul O'Brian obr...@colorado.edu http://ucsu.colorado.edu/~obrian
SPAG #23 will be devoted to the 2000 IF competition, and is actively
seeking reviews! Submit your comp reviews to me by December 10. Thanks!