News:

All future registrations will need to be approved before activation
due to the increase in spammers registering.

Main Menu
avatar_Ted

Four Against Darkness Conversion

Started by Ted, October 28, 2026 @ 11:28 PM

0 Members and 0 Guests are viewing this topic.

Ted

Four Against Darkness is a pen and paper solo-rpg game that's been around since about 2018.  The link shows the expanded edition, but this digital conversion is based upon the old basic game's 4.0 rules revision so doesn't contain all of the added features of the expanded edition.

The basic premise of the game is you take a party of 4 characters into a randomly generated dungeon and search and hunt, room by room, for the final boss of the dungeon and then try to make it out alive. 

It took about 2.5 days to code with ChatGPT just using prompts and refining results.  It's coded in html and runs in your browser (only tested in Chrome).  This is the beta version.  It's feature complete except I'll probably add the ability to export the dungeon map to a pdf as well as add some more tool tips to help players. After that if I feel like revisiting it I may try to dress it up so it doesn't look like such a generic AI game.  However, my primary goal was to complete a project and even if I don't touch it again I consider this my first one across the finish line. 

I'm sure this is copyright infringement so please don't distribute, it's for personal use only.  I just thought you might enjoy a little browser based dungeon crawler.  That being said if you don't want something like that uploaded here feel free to remove and delete this post.  I'm headed out on a road trip vacation in the morning and I wanted to get it posted before we left.

Feedback is welcome!

Jason

This sounds amazing! Looking forward to trying it out today.

Ted

Looking forward to your feedback!  It might be tough to understand what you're supposed to do since you don't have the rulebook and I haven't implemented a tutorial, but maybe not.  You're a smart guy.  If you have any questions about navigating the interface or want help just post.  I'll be on the road but will prob be checking the forum regularly when done driving for the  day.  You can always text me too.

Jason

You'll have to bare with me since I'm new to these types of games but I did complete a short expedition without any treasures.

While I'm not totally familiar with dice-based RPGs I can appreciate the depth of the game. The amount of options, characters and inventory management is really impressive. This far surpasses any of the text adventures I've made in the past. The user interface is very impressive too. If you told me this was a purchased game from a professional developer I would not question it for a second.

The one thing that dawned on me while playing was how this game can technically run on practically any system since it's HTML and runs in a browser. There's no reason I couldn't play this on my phone. I won't do because I know you're worried about copyright infringement, but this could easily be uploaded to any website (the entire game is less than 2.5MB) and could be played from anywhere on anything. 

I hope you continue developing it. If there's no digital version of this game, couldn't your game become the definitive version with the approval from the copyright owners?

Ted

#4
It probably could if I had the guts to approach them and ask.

I know it needs a tutorial or something helpful to get the player up and running.  Besides that was there anything else that you thought needed addressing.  Any quality of life stuff.  Anything at all?

When you developed your text based adventures did you write out a lot of narrative?  What was your process?

Jason

#5
Let me learn the game a little more before giving any quality of life opinions. Having never played these types of games, I don't really have anything to compare it to. The game currently does of nice job of detailing everything that is happening and I found the "Explain this Encounter" button to be extremely helpful for new users. I like the small touches like the animated dice too.

My text based adventures did not have a lot of narrative. I was finally able to convert the text based adventure game I made in Python back in 2019 into an executable file. I included simple PDF instructions (see below) with the layout of the map - it's a simple 5 x 5 grid with 25 rooms.

It's more of a foundation to get the basic functions working and is coded to be easily expandable, but I never got around to it.

Link to download game: Castle of Darkness


Jason

I've included the source code if anyone is interested in seeing how it works.

Ted

I won Castle of Darkness on easy!

Jason

That was the game I told you about a while ago where I was thinking of adding a rock/paper/scissor aspect to the game instead of just a random number generator for battles.

Ted

what would that mechanic look like in a fight?  You pick one and the computer picks a random one and you see the result?

On easy I never lost a battle so I never lost any life points.  What happens in CoD if you lose to a guard. The guard is defeated but you lose a life point?


Jason

You have 5 lives and keep fighting the guard until he's defeated or you lose all your lives.

I found my old thread explaining my rock/paper/scissors idea: https://forum.digitalgamingarena.com/index.php?topic=435.msg10026#msg10026

Jason

#11
One issue I had with Four Against Darkness is that the game resets if you switch to a different tab. I had a good expedition going and was debating whether or not to bribe or fight a monster (loving all the choices), but the game was reset when I switched back to my tab. In case it matters, I'm using Firefox.

Ted

#12
I'll see if I can reproduce that when we get back from our trip. In the meantime you might be able to save and load to prevent that in the future?  If I remember correctly it says something like save locally?

Thanks for the feedback.

Jason

Ah, that's my bad. I should have thought of that.

How does updating the game work? Do you have it uploaded to ChatGPT and just say things like "adjust X and Y"?

Also, how feature rich is this game compared the the board game or other digital versions? I know you said it doesn't have all the features of the expanded version, but does it have all the features of the standard version?

Ted

There's no other digital version as far as I'm aware.  It's not really a board game.  It's a pen & paper game you play by rolling dice and drawing the dungeons on graph paper.  It's pretty fun to play that way and I think most people enjoy it for the artistic part.

It has all the features of the original version with the 4.0 revision.

I did this one different from all the others I did and am doing.  I used the desktop app of ChatGPT instead of the browser version. I wanted to see how it behaved differently.

Since I had the full game already designed (the published book) I didn't need to feed it new ideas and features.  I stated I wanted a 1:1 conversion of the book so it had me upload every page in chunks of about 6 to 10 pages.  It then built out the game section by section stopping at points to internally test (it did all the testing) and when it was finally done it handed me the finished version.  I started playing it and realized pretty quickly that it had omitted or misinterpreted several rules and features so over the course of another day I told it the missing feature or rule and it would correct.  I want to say it needed about 10 passes and revisions before I released it as beta here.  How the process looked while it was building was I would check on it every so often and if it had stopped working and gotten to the end of a testing phase I would ask it what it needed.  On this program every time it said it had what it needed and was ready to implement the next part so I would just type in "proceed."  After that happened a couple dozen times I just typed in unless you need something from me proceed and off it would go.  It would usually churn for about 20 or 30 minutes at a time between implementations.  One of the interesting results of the process was that it also taught me that I was omitting some rules in my pen & paper play that I never realized I was supposed to be doing.

For every other module I've made and my current alpha of my big game I use the browser version. Because I've been making it "on the fly" I really have just been playing with the alpha and when I think of an idea I want to try I describe it as detailed as I can tell the AI to make a new branch and implement the idea.  There is a lot of back and forth in this process because I learned that the has a tendency to want to take over as the designer or architect if you don't explicitly tell it what you're trying to accomplish. It is also terrible at solving problems. It comes up with solutions that often either are very convoluted and unnecessary or it breaks other things that are already working fine.  Or it just "band-aids" the problem instead of preventing it from happening again somewhere else.  As far as I can tell it's great at coding.  I just have to tell it exactly the method to fix the problems that crop up.  I don't know any code, but I can critically think how the problem is happening and design the solution then the AI speaks the language the machine needs to implement.  I think spending 25 years as a mechanic/maintenance technician has helped me greatly here in being able to communicate with the AI in a way that effectively troubleshoots and solves problems.  So in the stuff I create from scratch I'm the designer and the play tester. 

From those learning pains I've established a few processes that have sped up the development and greatly eased frustration with the AI. Every new idea gets its own separate conversation until it's implemented into the stable alpha.  I have a set instruction list of my role and the AI role that the new conversation gets so it knows exactly what it's supposed to do and not supposed to do.  Importantly one of those instructions is that it needs to ask as many clarifying questions as needed so it doesn't guess and do it's own thing.  It's also instructed before each prompt within the conversation, after I tell it what I want or what is not working in what it produced, that it needs to explain to me technically what it's going to implement and also explain back to me exactly what I want from it so I know it understand the request.  Then if that looks good and I don't need to correct its explanations I tell it unless it needs something else proceed.  Sometimes it asks for an upload of a previous version of the alpha or a save json file, but most of the time it just goes and a few minutes later the branch is ready for testing.  Once in awhile we get stuck for a whole day on one issue (water, and roads and paths) have been challenging), but most days I can implement 3 ideas or even more and push a new stable alpha.

My next goal is to personally learn how to use git and become better organized and learn how to properly push repos (if that's even the right phraseology) because I was so happy with the outcome of Four Against Darkness that I want to convert all of my dozen or so solo board games and even a few other non solo board games.