[design] suggesting actions

4 views
Skip to first unread message

Steve Breslin

unread,
Sep 24, 2004, 12:14:56 AM9/24/04
to
At times, I'm inclined to tell the player what actions are supported
in a given circumstance. I think this would produce (for better or
worse) a sort of "choose your own adventure" aesthetic. For example:

====

You notice someone picking your pocket.
[You can "pretend not to notice," "attack or capture the pickpocket,"
or "call for help."]

>_

====

The purpose of this "action inventory" is not to limit the player's
range of action -- the player has a normal command prompt, and can do
whatever he likes, as usual. The purpose is rather to alert the player
to possible actions specially supported by the game, and to alleviate
the problem of "guess the verb" (or more generally, "guess the
command").

The player might incorrectly assume that the game (or its parser) is
"too dumb" to support certain complex complex. -- And in most cases,
it *is* too dumb, so somehow alerting the player specially when
there's an exceptional situation would be useful, and could be used to
really good effect.

Of course there are obvious drawbacks to my proposed solution: it
seems bossy and blunts the sensation of free-interactivity; it may
disclose more than it should; the intrusion of the UI is (arguably)
slightly ugly insofar as it breaks the flow of the game. Surely there
are other problems as well.

What do you think of this?

Branko Collin

unread,
Sep 24, 2004, 6:50:22 AM9/24/04
to

ver...@hotmail.com (Steve Breslin), you wrote on 23 Sep 2004 21:14:56
-0700:

>You notice someone picking your pocket.
>[You can "pretend not to notice," "attack or capture the pickpocket,"
>or "call for help."]

[...]

>What do you think of this?

There are already games that use this as a system for conversation.
However, when talking to an NPC you tend to try and drain all
information from that NPC, so whether or not you provide a menu system
only matters for playability; the desired result is the same.

Generally, when the author wishes to move the game along, s/he just
does so: "A pickpocket tries to snatch your purse, but you notice and
strike him across the fingers with a cane. The pickpocket yelps with
pain, drawing the attention from your arch nemesis across the square."
(You fault, then, was to walk onto the market, but perhaps that was
the solution to the puzzle of move-the-game-along.)

And when the author tries to attract your attention to the options you
have, s/he provides unobtrusive hints: "You try and make your way
across the bustling market unnoticed, when a pickpocket decides it is
time to relieve you from the few coins you need at the Purple Dragon."

Do you have any specific problem you are trying to solve with this? Or
are you philosophizing about how and if such a system would work?


--
Wel zie ik het betere en prijs het;
toch jaag ik het slechtere na

Jess Knoch

unread,
Sep 24, 2004, 12:01:39 PM9/24/04
to
Steve Breslin wrote:
>
> You notice someone picking your pocket.
> [You can "pretend not to notice," "attack or capture the pickpocket,"
> or "call for help."]
>
>> _
>
> ====
>
> The purpose of this "action inventory" is not to limit the player's
> range of action -- the player has a normal command prompt, and can do
> whatever he likes, as usual. The purpose is rather to alert the player
> to possible actions specially supported by the game, and to alleviate
> the problem of "guess the verb" (or more generally, "guess the
> command").

I prefer a game to have a list of the supported-but-unusual verbs and
commands, such as in the ABOUT text, so that it is very easy to get to when
I arrive at a sticky point and say, hey, what weird verbs were implemented?
Then I can go back and check the list. This works for unusual verbs and
complex phrasings both, and is very unobtrusive since the player must ask
for it. If you go this way, I highly recommend (in fact, I would expect)
that you also include a line like "First-time players should type ABOUT" at
the start of the game to tip us off to your added material.

p.s. In your example I would expect a response from
>ATTACK OR CAPTURE THE PICKPOCKET

since it's in quotes. :-)

--Jess

Kevin Venzke

unread,
Sep 24, 2004, 11:14:11 PM9/24/04
to

"Steve Breslin" <ver...@hotmail.com> wrote in message

> You notice someone picking your pocket.
> [You can "pretend not to notice," "attack or capture the pickpocket,"
> or "call for help."]
>

I think I would be put off by this text. I should be able to get the first
effect by typing WAIT, or doing something unrelated. ATTACK HIM
is easy to guess. Not sure about the last one, other than maybe
YELL (AT HIM?). Trying a movement verb also occurs to me.

I have a hard time imagining that after I have tried all of these
listed options, a clear winner wouldn't stand out. And I will be
annoyed if the game has suggested options that turn out to
be clearly bad ("ATTACK HIM. He pulls a knife and kills
you!").

I think that's my main objection, the more I think about it. If
you're going to list my options, then they should be good options.

Kevin Venzke


M.D. Dollahite

unread,
Sep 25, 2004, 5:48:46 AM9/25/04
to
>At times, I'm inclined to tell the player what actions are supported
>in a given circumstance. I think this would produce (for better or
>worse) a sort of "choose your own adventure" aesthetic. For example:
>
>====
>
>You notice someone picking your pocket.
>[You can "pretend not to notice," "attack or capture the pickpocket,"
>or "call for help."]

This seems like a natural parallel to TADS 3's conversation suggestions (and as
such, I would suggest using much the same syntax -- change "can" to "could",
and make it an in-character parenthetical comment from the parser). It could
prove very useful to provide situtational actions as well as situational
conversation options, for the same reason: to allow a wider range of actions at
certain junctures without creating an exploding number of commands that are
only used once in the whole story.

In this particular example, however, I'm not sure it would be necessary.
ATTACK and YELL are standard verbs, and one can ignore the thief simply by
doing something else. On the other hand, when a situtation comes up where I
appear to have only one try to guess the right action, I've often wished the
game would narrow it down for me a little. There's no worse time to be hunting
around in the dark for the magic verb than when you only have a turn or two to
find it before something bad happens. I don't know if this is really the right
way to approach that problem, though. A more subtle approach, working clues
into the description of the event somehow, would probably be preferable when
all the options are standard general-purpose actions.

As with the Topic Inventory, and for that matter any attempt to introduce a new
convention into an established medium, people will really need to see it in
action a few times before they can really judge whether they like or dislike
it. (And even then, they can be put off by poor usage, or made overly
optimistic by a really excellent treatment, until it's been used enough for a
good statistical sampling.) Even in mediums that encourage experimentation,
new conventions tend to be resisted until they prove themselves.

Victor Gijsbers

unread,
Sep 26, 2004, 8:18:55 AM9/26/04
to
> What do you think of this?

If you want to deliver a story with considerable emotional tension that
is resolved in a dramatic climax, making it fast is essential. No matter
how carefully you stage your final dramatic resolution, if the player
has to search for it to get there for half an hour, the tension you
built up will have evaporated. For some pieces of IF, then, keeping the
story going forward is crucial. I can see at least three ways to do this
- there are certainly more.

A) Restrict the possible actions to such a small set that one is bound
to stumble upon a right one in a very short time.
B) Suggest courses of action the player can take - this is your proposal.
C) Ensure that the narrative implicitly suggests certain actions so
strongly that they are imbued with a 'narrative necessity' which the
player perceives. (I say that a piece of fiction shows much narrative
necessity if its events seem to be necessitated by the preceding
situations, and little narrative necessity if the events are not at all
determined by the previous events.)

It is interesting to note that almost every piece of IF uses each of
these possibilities at least to some extent. (Look at the postscript for
my thoughts on this.) Anyway, it is considered good practice not to rely
on A too much (even though one always relies on A through the supposed
familiarity with the conventions of the parser): one should let the
player try all likely and even some unlikely actions and reward him or
her with a customised message. That leaves B and C as possibilities for
keeping a story tight and focused. I think both can work, but they are
not both suited to the same kinds of IF. So in the rest of this post I
will try to indicate what I think are the strengths and limitations of
your proposal, and what kind of context it needs to function.


Suggesting actions takes something away from the player (reader): a
large part of creativity, and the exhilarating feeling you and has the
effect you hoped for. It also gives something to the player: a clear
overview of his choices, and freedom from the frustration felt when what
we try does not work. This leads me to formulate the following principles:

1) Suggesting actions is a powerful technique when the piece is about
the choices made. A story in which moral choices play an important role
might benefit from a mechanic such as this which ensures that the player
actually perceives the choices and the possibilities offered to him. For
instance, if the story is about the question "What price are you willing
to pay for the power needed to accomplish your goals?", the player
should understand that not accumulating power is a viable in-game option.
2) Suggesting actions is an inappropriate technique when the piece is
about the creative faculty of the player. In a puzzle-solving game, the
fun consists in having the flash of insight needed to solve the puzzles;
suggesting actions would destroy this fun. For a piece of IF which is
about the willingness of the player to manipulate NPCs, or about the
ease with which the player thinks up violent/cruel actions ("O my god,
did I think that up?!"), not suggesting actions is also of prime importance.

These two points might be summarised as follows: if the game is about
choosing, suggesting actions will work. If the game is about thinking up
the possible choices, it will not.

3) If your piece of IF is imbued with a heavy dose of narrative
necessity - either because it is a genre piece, it is built up out of
well-known 'scripts', or because you are a master at creating a powerful
internal logic - suggestin actions may distract from the fun rather than
adding to it. It might even seem patronising. But if your piece of IF
lacks this narrative necessity - and this might be completely
appropriate, some stories will benefit from this - suggesting actions
will probably add to the player's fun, because it will keep him from
getting stuck, and will help him along. To use a metaphor: if the roads
the story can take are brightly lit, signposts are not needed; fi they
are covered with mist and darkness, signposts certainly help.


There are certainly more considerations that can affect your choice
whether or not to use the technique of suggesting actions. But I hope
the above has partly sketched which kinds of IF will benefit from it,
and which kinds won't. Whether you do or do not want to use it therefore
hangs together crucially with the piece you are writing.

Greetings,
Victor Gijsbers

PS.

To come back to my claim that all of the methods A through C are used by
almost every piece of IF:

The parser's limited vocabulary and grammar and the limited world model
of IF is in fact a manifestation of A. Placed in a room, we might try to
start the story in any way imaginable: "Think about your little baby
brother.", "Scrath your nose.", "Try to solve Fermat's last theorem".
But anyone with even a slight familiarity with IF will instead try such
things as "Look", "Examine door", "Go south", "Take raincoat". With
regard to method B, actions are suggested to the extent that conventions
such as the following have been established within IF: If an object is
described in some detail in a room description, examine it. If an object
is described as being in a room, and is not part of the room
description, try to take it. If you meet a character, talk to him or
her. If you find an entrance to some other place, enter it. These
suggestions are, however, not peculiar to the piece, and they are
implicit. Your proposal to explicitly suggest actions is therefore
rather different. As to method C, the construction of narrative
necessity is what makes a player try only a very small subset of the
possible actions he can take. When you are chased by an angry mob and
suddenly find yourself by the side of a river, you're not going to try
and drink the water, you're not going to try and wash your hands, you're
not going to try and build a boat, but you are going to try and swim
across. Every story that can be _understood_ displays some measure of
narrative necessity.

Reply all
Reply to author
Forward
0 new messages