Friday, 7 October 2011

The GDI (Gamebook Difficulty Index)

One of the most intriging parts of a gamebook is that you are attempting to achieve a goal by successfully completing the story.  Some stories have multiple successful endings, some have just the one.  In each of these stories, though, depending on the author, the difficulty level involved in attempting to complete the adventure varies from book-to-book.

For the past many weeks I have been attempting to construct a system that can be used to measure the difficulty of any gamebook written by giving it a difficulty rating.  Hopefully it can be something that people can use to rate different gamebooks when wanting to choose one to read.  I expect it will also be something people enjoy using to settle debates on which gamebooks are tougher (or easier) to complete.

I don't expect this system to be perfect, though I have tried to make it as unbiased and fair as I could think of.  I welcome any feedback on the system and any potential improvements people might think of.

I have called the system the GDI (Gamebook Difficulty Index).  The main principle I followed when constructing the GDI was that the skill of the reader was taken away when determining the final score (or index).  In the years of reading gamebooks, I have noticed that some readers have had a knack of finding the successful ending(s) more quickly than others, and I wanted that element to be totally removed.

The rules in determining the final index are basic in general terms but are also quite detailed in others.  I will explain how the index works in its most basic terms and then elaborate on some of the complexities (but not all at this stage - I will probably do this in later blog entries or this particular blog post will be too large).

I have applied the index, at this stage, to only one gamebook - My first gamebook, "Invitation to a Feast", a book that I have always considered to be a reasonably easy gamebook to complete.  At this end of this post, I will state what GDI I gave my gamebook, and then explain why the rating is not a final rating but is an interim rating.

Anyway, enough intro....let's begin!!!

The GDI consists of two components:

a) An exploration score; and
b) A gameplay score.

The GDI equals the exploration score plus the gameplay score.

In plain English, the exploration score represents the amount of times one has to read the book to find the successful ending in the gamebook.  If there is more than one successful ending to find, the score represents the amount of times one has to read the book to find any of the successful endings in the book.

At first glance, this score appears to cover all bases, but this is not the case.  When determining the exploration score, all gameplay mechanics are excluded from the book, which generally means all the die rolling mechanics are removed from play when reading it.

Using Fighting Fantasy gamebooks as an example, this means that:
- All combat is removed (you assume that you always win)
- All other die rolling is removed (E.g. Test Your Luck, Skill, Stamina etc.)

For the purposes of working out the exploration score, the story is (almost) reduced to a straightforward Choose Your Own Adventure story.  What is retained, however, is:
- Collecting and using equipment
- The use of Codewords or Events
- Other such similar devices (i.e. Non die-rolling mechanics)

So, just repeating, the score itself is the number of times one reads the book until a successful ending is found.  If you remember my earlier comment, however, my intention is that the skill of the reader is taken out of consideration.  The way this is dealt with is by:


In other words, if you read a page of a gamebook and are presented with three choices to select from at the end of the page, rather than make the choice yourself, you allow the roll of a die to determine what choice is selected.  This general principle is followed all the way through the book until a successful (or unsuccessful) ending is found.  If your adventure ends, you go back to the beginning of the story and start over.


The above statement is quite valid.  To prevent this from happening, however, there are a number of specific rules to follow when reading, so the exploration score that is calculated is a reasonable and realistic one.  These rules are where the complexity of determining the GDI is at its highest.

The rules/principles are as follows:

RULE 1: Each time the gamebook is read, it must be possible to complete the book EVERY time it is read.

This overriding principle is to prevent the reader from wasting many reads wandering along incorrect paths, where there is no chance of final success.  This principle also makes it obvious that when determining the GDI of a gamebook, one must first already know the true path or paths that must be taken to succeed. 

Originally, when designing the GDI, I wanted to give the reader the opportunity to calculate the GDI for a book that he/she had never read before.  This, for me, resulted in many problems (mainly revolving around wandering along incorrect paths for many reads on end and inflating the GDI to what amounted to ridiculous levels).

This rule overrides all other rules listed below.  The way it overrides the below rules is that it changes any potential compulsory selection of an incorrect choice to a random die roll between the correct and incorrect choice(s), where it is known that if the selection of a wrong choice is made, it will be impossible to successfully complete the book.  (spoiler follows)

Example - In the Fighting Fantasy gamebook, "The Warlock of Firetop Mountain", one of the keys needed to open the chest at the end of the game is found early on, in a room, inside a box that contains a small snake.  If any of the below rules force you to bypass the room that contains this key, you are to ignore the rule, but instead, roll a die to decide whether you enter the room or not.

RULE 2: When you are asked to make a choice between two or more pages to turn to, you must always turn to a page that you have not previously read.

Example.  You walk along a road and arrive at a junction where you can travel to the east or west.  If you travel to the west on your first read, you MUST travel to the east on your second read.

This rule is to encourage exploration of all possible page numbers in the book, thus logically ending in the eventual discovery of the page number where you successfully complete the book.

I use an excel spreadsheet to create a gamebook flowchart to track which pages I read on each occasion and highlight the page numbers that have already been read.

The above rule has one critical exception, however (not including rule 1 above, of course).

RULE 2 (exception): If, on a previous read, you have discovered that possessing an item of equipment or event (codeword) is needed to turn to a previously unread page, and you know both:

a) How to get to the page where you gather the item of equipment or codeword; AND
b) You have previously read the page where you cannot proceed to the unread page without having the item or codeword,

You MUST select the page number that will get you first to the item of equipment or codeword.

This is a complex exception and is best illustrated by using an example (spoiler follows):

In the Fighting Fantasy gamebook, Citadel of Chaos, the villain, Balthus Dire, has a combination lock to get into his war room for the confrontation at the end of the game.  Let us say, for example, when reading the book, you discover the location where the combination is found but do not get to the combination lock at the end.

Continuing on, in a later read, let us assume you discover the combination lock for the first time, but this time, do not have the combination.

Because you have discovered both parts of the "solution" to this particular puzzle, for ALL subsequent reads of the book, you must ALWAYS seek out the combination first so that it can be used to unlock the room at the end, because there is an unread page that needs to be read.  As long as there are still unread pages to be read after unlocking the door to the war room, this general path (seek the combination first) must be followed on every subsequent read.

If you have multiple examples of the above event, (i.e. You have discovered more than one item/codeword and where it can be used later on in the story), you must roll a die to decide which item/codeword you focus on (unless, of course it is possible for you to follow both).

RULE 3: When you are asked to make a choice between two or more pages to turn to and there are more than one unread pages to choose from, you must roll a die to choose between them.

This rule was alluded to earlier.  Again, it encourages exploration of the entire book.  Rule 1 and Rule 2 (exception) can override this rule, however.

RULE 4: When you are asked to make a choice between two or more pages AND all of these pages have been previously read:

a) Select a page that will eventually lead you to an unexplored area
b) If all pages (or none) eventually lead you to an unexplored area, then roll a die to choose the next page

Where possible, all of the above rules can override this rule.

RULE 5: Making choices that do not involve turning to a page.

A typical example of this is choosing a list of items to buy from a store, merchant etc. where all the items are listed on the one page.  In this scenario, once again, you roll a die to choose whether to buy an item or not.  If you have limited money, then once again, you roll a die to reduce your selection of items to buy.  Again, this rule can be overriden, primarily by rule 1 or the rule 2 exception.

I realise this is a complex system, and I would not be surprised if some of you that read this end up totally confused by it.  For those of you who can follow it, I welcome any feedback on it.  This system certainly does not have to be the final system and if somebody can find a way to genuinely improve it, I welcome any suggestions.

So, in summary, after following the above rules, you will eventually find a successful ending to the gamebook.  Once this is done, apart from recording how many times the story had to be read in order to win, you must move along to the next element of the GDI....

Thankfully, the calculation of this score is very simple.  All you have to do is the following:

1) Read the gamebook again as you would normally read the gamebook (i.e.  All gamebook mechanics are reinstated).
2) Using the path to victory you discovered when calculating the exploration score, as much as is possible, read the story through to its conclusion.

Similarly, the gameplay score is the number of times one reads the book until the same successful ending, discovered while determining the exploration score, is found.

That's it!  The total GDI, I think, is a reasonable score that should represent the average number of times you would expect an average gamebook player to read the book through to a successful completion.

As I mentioned earlier, I have used this system to rate my first gamebook, "Invitation to a Feast", and this is the score I came up with:


GDI = 4

Finally, to wrap things up, I made a comment earlier that this score is what I classify as an interim score.  The reason being is that I could just as easily test the book again and come up with a different score.  What I plan to do is to indeed, test the book again, calculate the GDI and then calculate the average GDI based on the two tests.  I will continue to do this until the average GDI varies by less than 1 point between tests.  I will then conclude the test and consider the GDI as the final GDI.

A fair bit involved, yes, but I think, a worthwhile exercise, particularly for those who enjoy statistics (like I do!).  What I am most pleased with is that this system is unbaised and impartial and will really come into its own, once some final GDIs for different gamebooks are calculated.  We should then be able to compare gamebooks and truly classify them from easiest to complete to hardest to complete.

The next book I plan on rating is one of the most famous and difficult Fighting Fantasy gamebooks written, Creature of Havoc.  I am certain the GDI will not be a low as 4 for this one!

Thanks for reading,




Thanks to all those who made comments about this system on this site (and other sites that I frequent).  I plan on responding when I get a chance but the feedback, so far, has been appreciated.

In the ensuing analysis by myself and others since I released this article, some weaknesses in the system have come to light.  In particular, I gave the system a real workout by testing it with the fighting fantasy gamebook, "Creature of Havoc".  I figured that if any gamebook would find any faults with my system, Creature of Havoc would deliver the goods, and it did.

For those of you who are familiar with the story, my system broke down when it became apparent that it was very, very difficult to find the vapour of tongues (language) when using my original system, and did not represent a realistic scenario at all.

I have since made an appropriate adjustment to rule 1 (which was the main cause of the problem).

I also realised that I forgot to include a couple of rules in my original post.

Rather than edit the original entry (above), I will make a new blog entry with the updated system, so those of you who are interested can follow the natural evolution of the GDI from its beginnings, to where it is now.

With the adjustments made, I have found that I can now progress through Creature of Havoc in a gradual, realistic way each time it is read.  I haven't yet completed the story but I don't envisage any great problems with the GDI now.  Any further changes from now on, should be minor (if any, I hope!)


P.S.  The updated GDI system will follow in a new post once I have written it.  If I get a chance, I may explain in more depth how the original system failed, for those of you who are interested.


Slloyd14 said...

I think I've done something along these lines but in a none systematic way. It sounds like it may take some time, but I might try it with a few books. I might start with a non random one such as Heart of Ice before moving onto a Fighting Fantasy one (but not Crypt of the Sorcerer - the gameplay score would be through the roof).

J P Barnett said...

Ah yes, the Razzak hits you in consective attack rounds and you lose rule. (There may be other issues too but I don't know of them as it is not a book I ever finished).

I am very interested in finding out the score for Creature of Havoc. Hence, I am working on it at the moment.

Ben Britten said...

This is excellent work and very interesting. I am glad there are others out there who are trying various ways of analyzing gamebooks.

I am working on a similar system to rate our gamebooks but it is based more on the 'branchyness' of the narrative. It isnt fully fleshed out but the basic idea is inspired by fractal geometry.

If you imagine a normal linear narrative (ie a non-gamebook) that would get a gamebook dimension of 1 (fully linear). A narrative that was a sandbox world where you could go anywhere and do anything would be a 2 (infinitely non-linear). Most gamebooks would fall somewhere inbetween. The closer to 2 you get, the harder the gamebook (or so the theory goes).

To calculate the gamebook dimension you traverse every single possible path through the book (which can get fairly complicated when shops are involved) and then analyze them for things like shortest path from start to finish and longest path, how much overlap between various paths through the narrative etc. and you then use all of that info to generate a number. This can be quite computationally intensive, but that is what computers are for :-)

One nice thing about going digital is that you can opt to include the dice mechanics as well and run statistical analysis of all of the various paths over and over again thousands of times to figure out if the game is properly balanced or not.

In any case, once I get my system all worked out I will definitely do a blog post on it on the Tin Man Games site. (and maybe even provide some code) But I am glad there are others out there treading in similar waters trying to help classify gamebooks. I would love to hear more about how your system fares under multiple passes, I really like the exploratory nature of the rules you have laid out here. Where my approach is more structural, yours is a bit more experiential which I think is very valuable.

Ben Britten
Tin Man Games

J P Barnett said...

Thanks Ben,

I like the sound of your 'branchyness' system. I agree, it is certainly taking a different approach to my method. I will be very interested in reading your blog entry once it is done.

One old FF book that would be interesting to put through that system would be Scorpion Swamp (where you could revisit old areas).

If you read my update above (in the main post) you will see what has happened to the system when giving it a good workout with one of the most notoriously difficult FF books written (from an exploration point of view), Creature of Havoc.

The next book I am really interested in testing my system with is actually my own gamebook (my second), The Wounded Falcon. Although short, it has a relatively unique "hidden choices" system which really opens up the book for the reader to attempt all sorts of things from an exploration point of view, with a penalty incurred for "poor exploring"

Thanks again for your feedback