Determining PageRank-like values in IF?

Greetings fellow Quest users!

First, i know this is horribly specific and/or completely unnecessary, but hear me out:

The game i'm currently working on, just as some other IF games, uses an "Open Map Style"-progression as detailed here:

Now, whilst a normal parser-based text adventure has a rather simple map (atleast when you create it with Quest) that makes it comparatively easy to get an overview, both as developer and player, this is not necessarily (e.g. almost never) the case for CYOA games.

This is where my unorthodox approach kicks in: If all the player does is clicking links anyway, isn't it possible to apply normal PageRank logic (or a simplified version thereof) to determine the importance of certain regions? A Hub Area would have a higher Rank than an option side branch, and if you have multiple endings, this allows to estimate the likeliness of each, etc.

I just thought this could come in handy for IF developers, and it's interesting nonetheless. So does anybody with more knowlegde on this knows if this could become feasible?

Just a suggestion. Also, in case somebody has no clue what i'm talking about, the free encyclopedia to the rescue!

I am not to clear on why you would want to do. Is this for your own game? I would have thought you would already know what the hubs are. If this is to help players navigate a game, I would suggest a map, even for a CYOA, in part because players would understand it.

As for how to do it, it should be pretty easy to have a program that first extracts each page and then counts the links to it. In theory it should be possible to build it into the editor.

Why would anyone want to do this? Well, if you ask like this, i might not be able to satisfy you.

The main reason is curiosity, analysis (there's a reason why providers of analysis tools make loads of money). It's not about knowing what leads where, but how important the connections are. (Like, tweaking shortcuts and streamlining the game.) Also, to be frank, in my own games, i know the general layout between hubs, but less important side stuff is something i can't keep track of (mostly because my games sprawls with it).

In more linear games, it also allows you to have a traceable progression tree with realtive likeliness of outcomes easily comparable.

Yeah, it's gimmicky, but it should be really easy to pull of. If i find software that can properly deal with those numbers (you could even use CTRL+F in the game's source code to find out), i'll link it here.

But yeah, having an overview about backlinks leading to a given command/page/room/thingamajig would be very handy as editor tool too.

I tend to visualise my games using Graphviz. It isn't too hard (if you've got any scripying skill) to convert a game to a big list of page1 -> page2 lines, and then graphviz provides several layout algorithms (dot, twopi, neato, sfdp…) to draw a map of them. Makes it easy to see at a glance any paths that might feel too linear to a player, or similar.

My first thought is... Why???
Why would want a program to predict how the game would be played...
Real life program would (or could) predict what stuff people would buy so that a company can make the stuff that will sell, and avoid what won't sell...
(If I follow what you are saying)
But I don't think that could apply to games.
If you could predict HOW the player will react to every choice, then you could just program the story to remove all the 'unfavorable" choices... (We call those things "books")
"Why can't I just pick up the hammer to drive the nail down to stop the squeaky floorboard?"
(Because the programmer want's you to use the crescent wrench in the other room to do it...)
It may be interesting to have a sub that tracked every choice a player makes in a game...
Collect that for everyone that plays the game..
And see how different people play the game, BUT, what use is that answer???
A game company like Infocom may use that to tailor a game to a target set, but other players may like a different type of game.
Also... I think your results could only be applied to that one game. You may be able to use that to tweak that game to be the best game ever... after about 75 revisions, but it will not help for your next project.

That's my $0.02 worth... (But just like free advice... it is rarely over priced...)

That was an interesting read... Quite informative...
But I'm not sure where page-rank could be used in a game...
I agree, Open Map Style may be the easiest to create and use...
But, there may be 2 layers to a game...
The basic map, and the over-layer of the story you are creating...
I have one game, Dark Halloween, that uses one map, but things in the rooms change over time that tell the story...
A open map, with a liner story. (I may revisit that one with a few new tweaks…)

For CYOA, I use Twine, and you can just look at the links (it looks like the images in your linked blog page). What system are you using?

Uh, quite a lot to reply.

First of all: thanks mrangel for mentioning Graphiz. It looks like someone you need to get worked in first, but i think it's just what i need.

Now, before i looked at the time of posts of you, DarkLizerd, i was somehow worried you have a split personality or such. Phew, you only read the article and overthough the matter, i see.

First of all, i'm a freak about analysis, and curiousness is already reason enough for me to want such information. However, as i said, it would help "streamlining" the game. I see now how that could be missunderstood.

What i basically mean is the reduce the amount of time a player is railroaded. If the crescent wrench is in the shack in the backyard, do you really have to cross the garden first when coming from the terrace?

By detemining the PageRank and drawing a map such segments (called in german "Schlauchlevel"; Hosepipe-level) become obvious and can be removed. (They're really not fun, eh?)

Also, it allows me to see if one ending is unussualy hard or easy to achieve. Yes, probably i want a secret ending unlocked through a certain set of actions... but if it is ridiculously hard to get the neutral ending because you always end up as hero or villian, it might need some tweaking.

Also, i've heard other's praise Dark Halloween, i think i need to play it.

Oh, and i'm using the Quest gamebook mode, Pixie. In retrospective, a incredibly stupid decision.

I came across an article a while back which showed graphs of a large number of the first generation of "Choose Your Own Adventure" and "Fighting Fantasy" gamebooks; and it was interesting to see that several prolific authors were recognisable from the shape of their graphs.

This is what one of my first attempts at a CYOA looked like:

(though it looks like some links are missing from the graph)

"Now, before I looked at the time of posts of you, DarkLizerd, i was somehow worried you have a split personality or such. Phew, you only read the article and overthough the matter, I see."

Didn't you play Pixie's Quest???
I cloned myself... I'm not sure who the original is any more...

I cloned myself... I'm not sure who the original is any more...

I needed this laugh today.

Thank you, DL!

One of us thanks you...

Log in to post a reply.