Interactive Fiction Survey results
David Fisher
(If you missed it, it's still here:
http://www.kwiksurveys.com/online-survey.php?surveyID=LIMN_5b7ed309).
Here are the results ... some of them are slightly different to the ones
displayed at
http://www.kwiksurveys.com/results-overview.php?surveyID=6213&mode=4, since
these percentages are relative to the number of people, not the number of
answers (eg. that page says 11% for "Adventure" in question 1, but 57% of
people actually ticked it).
Total number of responses: 94
1. "Which of the following genres of IF do you most enjoy playing?"
50-60% Adventure, Science Fiction, Fantasy
40-50% Humorous, Mystery
30-40% Drama, Historical, Horror, Serious, Surreal
20-30% Espionage
10-20% Romance, Superhero, Wordplay
5-10% Western, AIF
30% of people ticked all three of "Adventure", "Science Fiction" and
"Fantasy".
19% ticked none of these three.
Other genres included Slice of Life, Experimental, Science-Fantasy, Satire
and Folkloric/Fairytale.
2. "What do you most enjoy about IF? Rate the following in order of
importance (1-7, 1 being least important and 7 being most)"
Average ranking:
5.9 Seeing the plot unfold
5.2 Exploring the environment
4.5 Overcoming challenges
4.4 Interacting with other characters
4.2 Experimenting with a system and discovering its rules
3.5 Being in someone else's shoes for a while
3.2 Learning about a historical setting
All seven were given all possible ratings, except that "Learning about a
historical setting" was never ranked first and "Seeing the plot unfold" was
never last.
3. "Do you prefer "puzzle" or "story" IF?"
63% Lean more towards "story"
21% Lean more towards "puzzle"
7% No preference
6% Prefer puzzleless IF
2% Prefer pure puzzle games
4. "What is your ideal game length?"
33% Finishable in an afternoon
31% Finishable in a few days
19% Finishable in a week or so
9% Epic
5% Finishable in a month
3% Very short
5. "What is your most memorable moment from a work of IF?"
Lots of interesting, spoilery answers ...
General themes:
* Initial encounters with IF.
* Plot twists.
* Realising what is really going on in a story.
* Solving a difficult puzzle.
* Emotionally impacting moments, particularly endings.
* Establishment of mood / connecting with the PC.
* Vivid descriptions.
* Humorous moments.
* Memorable characters.
* Thorough implementation / responses to unusual actions.
* "Cool" things (robots!)
* Setting the monitor on fire while playing IF.
6. "Have you written any IF?"
40% Yes, and released it publicly
38% I'm learning, but haven't finished anything yet
14% Yes, but haven't released it publicly
4% No, prefer just playing IF
1% No, the programming seems too tricky
2% No, put off by the writing aspect
(Not sure if this means it was mostly RAIF who answered this survey, or just
that most people who play IF like the idea of writing it as well).
Of those who answered "learning, but haven't finished anything yet":
54% ticked Inform 7 in the next question
19% ticked TADS 3
17% ticked Inform 6
10% ticked Adrift, Hugo or TADS 2.
7. "Which authoring system(s) do you use?"
46% Inform 7
25% Inform 6
12% TADS 3
8% TADS 2
5% Adrift
3% Hugo
1% Alan
1% Quest
1% RAIF pool (I know who you are!)
26% ticked both Inform 7 and Inform 6
10% ticked both Inform 7 and TADS 3
5% ticked both Inform 6 and TADS 3.
45% ticked a single language
10% did not tick any.
8. "What are you most seeking to produce in people who play your works of
IF?"
57% Pure enjoyment / having fun
44% "Aha!" moments when they discover something
38% Emotional impact
27% A sense of wonder
21% Reflection on the issues of life
Other answers included immersion, originality, good writing/story, good
gameplay, exploring and understanding a system, and "Woah, that's cool".
David Fisher
Conrad
David, that was a great survey.
I've made a follow-up survey specifically on IF storytelling:
Conrad.
S. John Ross
Only comment that comes to mind:
> 3. "Do you prefer "puzzle" or "story" IF?"
This is the question that gave me the most trouble for lack of an
"other."
I answered "no preference," not because I lack a preference, but
because I above all seek games where _neither_ "story" nor
"puzzle" (nor some idealized combination of the two) are permitted to
wag the dog.
David Fisher
<1973bb44-2d12-4c62...@l33g2000pri.googlegroups.com>, S.
Do you have an example? (What wags the dog in your ideal game?)
David Fisher
Emily Short
<davidfis...@australiaonline.net.au> wrote:
> 1. "Which of the following genres of IF do you most enjoy playing?"
>
> 50-60% Adventure, Science Fiction, Fantasy
> 40-50% Humorous, Mystery
> 30-40% Drama, Historical, Horror, Serious, Surreal
> 20-30% Espionage
> 10-20% Romance, Superhero, Wordplay
> 5-10% Western, AIF
This question befuddled me a little on two points.
First, insofar as IF divides well into these (mostly book-based)
categories, the results conform pretty closely to distribution that
currently exists: there's lots of adventure, SF, and fantasy written;
relatively little romance; almost no westerns. (AIF is an obvious
exception, but on the other hand I think a lot of the AIF players have
their own subcommunity and might be selected against by the posting of
the survey on R*IF -- and the fact that AIF wasn't an official option
on the survey.)
So I wonder whether the low selection rates for things like westerns
mean a) hardly any of us here like westerns [as a naive reading of the
survey would suggest] or b) very few of us have played many westerns,
because there aren't many, and therefore they don't form a significant
portion of our experience of IF and don't tend to qualify as
favorites. At any rate, the last westernish IF I played ("Gun Mute")
is one of my favorite games of the last year or two -- but that wasn't
enough to make me tick western either.
Which brings us to the second point: lots of the best IF games don't
really categorize easily in these terms at all -- "Gun Mute" is sort
of a western but also sort of a romance, an SF game, a shooter, or a
combat-puzzle game, depending on how you're categorizing. I often
wonder whether conventional genre divisions serve IF all that well;
I'd suggest that IF genres exist, but they're not perfectly coincident
with literary genres.
> (Not sure if this means it was mostly RAIF who answered this survey, or just
> that most people who play IF like the idea of writing it as well).
This is another thing I was suspicious about. My sense (based on
really limited information) is that RGIF has primarily the old-school
IF players -- people who have been around for some time and who are
comfortable with usenet -- who also tend to be the ones who have
sometime or other thought about writing a game; and that casual IF
players (to the extent they exist, and I really do not know what the
numbers are) are elsewhere: Jay Is Games, the Something Awful forums,
etc. Anyway, I'm pretty sure that group is bigger than the survey
suggests. I just don't think they're mostly *here*.
That all may sound like I'm griping, and I'm not -- this was really
interesting to see. There's just a bunch of other stuff I'd also like
to know, especially if I could survey exactly the sorts of folks whom
RAIF doesn't reach, or who aren't active-enough participants to bother
to follow through and fill out a survey...
David Fisher
<b3cec102-1d29-499d...@m45g2000hsb.googlegroups.com>, Emily
Short<ems...@mindspring.com> wrote:
> On Sep 10, 7:07 am, "David Fisher"
> <davidfis...@australiaonline.net.au> wrote:
>>
>> (Not sure if this means it was mostly RAIF who answered this survey,
>> or just that most people who play IF like the idea of writing it as
>> well).
>
> This is another thing I was suspicious about.
I'm all for suspicion when it comes to statistics ...
Just to encourage all you Western Wordplay Romance writers out there,
I hope you're not put off by these results! Think of it this way;
whatever kind of game you write, it will be someone's favourite; and as
Emily pointed out, exceptional games (like Gun Mute) can do really well
whatever the genre. (If there is such thing as genre).
> That all may sound like I'm griping, and I'm not -- this was really
> interesting to see. There's just a bunch of other stuff I'd also like
> to know, especially if I could survey exactly the sorts of folks whom
> RAIF doesn't reach, or who aren't active-enough participants to bother
> to follow through and fill out a survey...
Are there some other questions you would like to have seen included?
(Not that I'm suggesting too many more surveys at the moment ...)
David Fisher
Nikos Chantziaras
> Thank you to everyone who filled out the IF survey!
>
> (If you missed it, it's still here:
> http://www.kwiksurveys.com/online-survey.php?surveyID=LIMN_5b7ed309).
Error Saving Answers (Error #00201) SESS_SID= DB_SID=
Jeff Nyman
On Sep 10, 11:30 pm, "David Fisher" <David.Fis...@efs.mq.edu.au>
wrote:
Here I'll be playing the part of the Devil's Advocate, so none of this
should be taken as negative of the attempt to do the survey in itself.
Here's my overall thought in case you just want to stop reading now: I
think a survey is potentially not the most effective way to figure out
this kind of broad stroke information about behavioral data. (This
doesn't apply to all the questions in the survey, of course, such as
those dealing with which textual IF authoring tool is used.)
> Are there some other questions you would like to have seen included?
> (Not that I'm suggesting too many more surveys at the moment ...)
I think the main thing I notice is that there is a built-in selection
effect here in that the survey is largely going to a group pre-
disposed to answer in certain ways. Rather than other questions (in
the sense of more), it might be worthwhile breaking down the questions
in a more operational or behavioral sense. Don't lead the questioner
with a category-type, for example, but rather give situational
elements that allow for responses that don't simply resort to type.
(Granted, this makes it harder to data mine but it also means the data
is probably a bit more valid. I once did a behavioral questionnaire on
horror movies that attempted to determine what elements people liked
and disliked. It was interesting to see what I got back -- but I would
have got back much less interesting results I think had I just done a
survey.)
By way of example, in case I'm not being clear, consider this question
from the survey: "Which of the following genres of IF do you most
enjoy playing?"
If I've never played, say, "Serious" or "Surreal" (assuming I have the
same definitions as everyone else), then I'm not going to check that.
But you have no way of telling if that's because (a) I just don't like
them or (b) there aren't enough of them for me to play or (c) there
are enough but I just wasn't aware of them.
A single survey question like this can potentially establish whether
or not a relationship exists between two variables (such as enjoyment
and "genre") but it's not sufficient to determine the direction of
causality necessarily. Also this kind of question is non-specific
among the determinant (genre). I would be more interested in what
elements of a work of textual IF people enjoy, and worry about whether
that fits into genre later. (This is similar to what I did with my
horror survey. The overall basis was horror and that was a term that
was defined in the context of the survey, so "Alien" was considered
horror but "Predator" was not. I didn't worry about category types
like "thematic" or "misdirection" or "savage cinema" or "evil kids."
Rather I got operational responses and then determined, based on that,
if those fit into what some would consider a category or perhaps a sub-
genre.)
Another example from the survey: "What do you most enjoy about IF?"
Some of the responses could be considered co-incident with each other:
"seeing the plot unfold" may be part of "overcoming challenges" in
people's minds. You may be able to cull that based on how
*individuals* answered, but collectively it leaves open the notion of
what the data really means. Further, I'm worried this kind of question
may be non-determinative in terms of probative value. "Learning about
a historical setting" may be incidental to a given work of textual IF,
although it's one possible element of it. So, for example, would
people dislike "1893" because of this? (Although this goes back to the
concern of what they have been exposed to, in terms of awareness, and
what they have actually tried.)
One more example: "Do you prefer 'puzzle' or 'story' IF?"
This assumes we all agree what the demarcation point is between the
two "forms." Do we? Also, in terms of probative value (beyond not
telling you what the terms mean), a more important question is how
nuanced people are to this distinction. I know there are multiple
options but here's how I read those:
Prefer puzzle textual IF
Lean more towards "puzzle" (really the same thing as above,
operationally)
Lean more towards "story" (really the same thing as below,
operationally)
Prefer puzzle-less IF
So for me I see this as essentially a binary choice, which the survey
somewhat shows, with 84% going for two of the categories. (Caveat: On
the matter of defintion, I realize with the selection effect inherent
in the group this was mainly directed at, the responses probably do
have some probative value. Remember that devil suit I'm wearing.)
Also, as a possible matter of interest, what I think is sometimes
revealing with these types of surveys is this: *before* releasing the
results, ask people what they predict the survey will show. That will
often give you an idea of bias effect inherent in a group by looking
at the selection effect beforehand.
As another possible matter of interest on methodology, I've found it
effective to have people play certain specific works of textual IF
(that presumably showcase different elements) and then, after that has
been done (even if not to completion), give data-gathering
questionnaires. Another effective technique I've found with textual IF
is to set up "workgroups" wherein the goal is similar to reading clubs
that read the same book and then discuss what they like and don't like
and certain points. In the case of textual IF, this is just letting
people talk about what they like and dislike about the experience of
the work of textual IF, without pre-disposing it to game or story or
genre. (Again, not all of this has to be where the works are played to
completion necessarily since you're not so much putting the emphasis
on *that* work as you are putting emphasis on what *elements* people
like and dislike in *that* work compared to others.)
It's a longer road to getting operational data, granted, but I've
found (in my limited context) that it's a potentially useful one.
- Jeff
David Fisher
news:51a2901d-6b41-4cf0...@z72g2000hsb.googlegroups.com...
>
> I think the main thing I notice is that there is a built-in selection
> effect here in that the survey is largely going to a group pre-
> disposed to answer in certain ways. Rather than other questions (in
> the sense of more), it might be worthwhile breaking down the questions
> in a more operational or behavioral sense.
Is Conrad's follow up survey the kind of thing you are thinking?
> By way of example, in case I'm not being clear, consider this question
> from the survey: "Which of the following genres of IF do you most
> enjoy playing?"
>
> If I've never played, say, "Serious" or "Surreal" (assuming I have the
> same definitions as everyone else), then I'm not going to check that.
> But you have no way of telling if that's because (a) I just don't like
> them or (b) there aren't enough of them for me to play or (c) there
> are enough but I just wasn't aware of them.
The question was really "Given the following list of genres, in your own
view of what each represents, which ones do you think you would most
enjoy playing?"
If someone didn't tick "Serious", I wouldn't interpret that as meaning that
they would never play something with that label, just that their picture
of what "Serious" means isn't in their list of most enjoyable games.
I think I might have been asking a different question to the one you
are interested in finding out about ...
(On the subject of genres -- have you ever searched Baf's Guide or
the IFDB by genre? It may be a very approximate way to label things,
but I still think it is useful).
> One more example: "Do you prefer 'puzzle' or 'story' IF?"
...
> I know there are multiple options but here's how I read those:
>
> Prefer puzzle textual IF
> Lean more towards "puzzle" (really the same thing as above,
> operationally)
> Lean more towards "story" (really the same thing as below,
> operationally)
> Prefer puzzle-less IF
>
> So for me I see this as essentially a binary choice
I totally disagree, but that's OK ...
If I reword the categories, maybe the distinctions are clearer:
* No story, just a collection of puzzles.
* "Puzzle heavy", but they are not arbitrary or random.
* Puzzles are present but are not the focus of the game.
* There are no puzzles at all.
* (As well as "I would enjoy any of these", ie. no preference).
I can't reduce these to a binary choice at all ...
Thanks for your thoughts,
David Fisher
S. John Ross
Not sure I've met my ideal game yet, although the general idea is that
nothing wags the dog :)
Unfortunately, I can only hope to express my design ideals through my
own designs, and my current designs aren't games I can point to as
very lucid examples of anything, though ToaSK, in particular, is (I
think) a work that can be fairly/accurately described as not
particularly puzzle-oriented (though it has puzzles, of a sort) and
not particularly story oriented (though it has story, of a sort) and
not any kind of real fusion of those two things, either. The puzzles
(such as they are) and the story (such as it is) are just part of an
elaborate garnish (there to inform and highlight the structure of the
gameplay). I wish I had a clearer example, but WIPs being what they
are, the rest are all still very "IP." I also wish I could pick just
one metaphor and stick to it.
In ToaSK, the "dog" is a kind of enthused and unabashed celebration of
a certain pile of geek-culture rubbish, and neither the story bits nor
the puzzle bits are permitted to do anything but _be_ wagged by that.
In the kinds of games I seek as a player, the "dog" is similarly
something _other_ than story or puzzle, though I have no ideal
"dog" ... "enthused celebration" is just one breed among many.
Conrad
<davidfis...@australiaonline.net.au> wrote:
> "Jeff Nyman" <jeffny...@gmail.com> wrote in message
>
> news:51a2901d-6b41-4cf0...@z72g2000hsb.googlegroups.com...
>
>
>
> > I think the main thing I notice is that there is a built-in selection
> > effect here in that the survey is largely going to a group pre-
> > disposed to answer in certain ways. Rather than other questions (in
> > the sense of more), it might be worthwhile breaking down the questions
> > in a more operational or behavioral sense.
>
> Is Conrad's follow up survey the kind of thing you are thinking?
BTW, a public response to an email comment:
With Surveymonkey's free account, I'm not able to customize the "thank
you" screen. But rest assured, I will make public the data collected.
I'm already starting to see some trends: but also, interestingly, I'm
seeing some patterned differences. It seems that people have playing
styles, which influences (for example) what kinds of game logic they
like and what kind of PC they lean toward.
The image is still fuzzy: with more responses, it'll become sharper.
Conrad.
ps - And yes, there are of course limitations to the surveying method;
I encourage anyone with the inspiration to post questions which would
get more to the heart of good game design.
C.
Conrad
> A single survey question like this can potentially establish whether
> or not a relationship exists between two variables (such as enjoyment
> and "genre") but it's not sufficient to determine the direction of
> causality necessarily. Also this kind of question is non-specific
> among the determinant (genre). I would be more interested in what
> elements of a work of textual IF people enjoy, and worry about whether
> that fits into genre later. (This is similar to what I did with my
> horror survey. The overall basis was horror and that was a term that
> was defined in the context of the survey, so "Alien" was considered
> horror but "Predator" was not. I didn't worry about category types
> like "thematic" or "misdirection" or "savage cinema" or "evil kids."
> Rather I got operational responses and then determined, based on that,
> if those fit into what some would consider a category or perhaps a sub-
> genre.)
An interesting take, Jeff, and should you choose to design such a
survey, I'd be sure to take it and glad to read the results.
Were I able to put real resources into how people play IF, I'd
probably try to identify equivalency sets for the players: 'If you
liked Predator, you're sure to like Alien...'
Once you have a few dimensions of categorization, (Alien + Predator
predicts Alien 2; Alien + disliked Predator predicts Nightmare on Elm
Street -- or however the data works out), then I'd open up
conversations with people in the various groups regarding what they
liked about the highly predictive movies.
Conrad.