Other D&D Variant Simple 1D6 D&D variant created with ChatGPT

kermit4karate

A strong opinion is still only an opinion.
Good morning. If anyone's interested, here's a simple D6-based D&D variant game (2 pager) created with ChatGPT's help. I've been refereeing iterations of this system for a couple months. Great for beginners, or really anyone who wants to actually roleplay with some margin for error and very few TPKs. It's very learner friendly. Open to suggestions. This plays fast and lean.

POCKET-SIZED D&D PLUS
 

log in or register to remove this ad

This is pretty neat, as a condensation of the rules printable on two sides of a sheet of paper, and playable with a d6 and index card character sheets.

I had to infer that AN = armor number or the equivalent of AC, and DN = difficulty number rather than DC for Difficulty Classbut that was no trouble.

One section I'm having a bit of a challenge with is the magic. I feel like that needs a little more explanation. I appreciate the idea that you improvise the exact spells/interpret them by the guidelines, but I'm not clear, for example, how you differentiate a light, moderate, or heavy damage spell. Is the idea that you need to roll a higher DN, say, 5/6/7 for light, mod, heavy spells? Control spells specify that they should be DN 5-6 for 1rd of disabling, which gives a guidelines to go by, but the other spell types don't. Cantrips are mentioned as being free (not limited in slots), and under Wizard one of their abilities under "spellburst" is "Firebolt(1)", so I'm guessing maybe the intent is that 1 damage cantrips are only DN5 to cast and have no usage limits?
 

Looks ok, and easy to learn which is a plus. I see under Hero Point that you can only use 1 per adventure but Humans get 2. Not sure if you get a HP per day or per adventure and this means that humans can use the 2, but on different days of the same adventure?
 

Looks ok, and easy to learn which is a plus. I see under Hero Point that you can only use 1 per adventure but Humans get 2. Not sure if you get a HP per day or per adventure and this means that humans can use the 2, but on different days of the same adventure?
Up to the DM's discretion, which to me makes as much logical sense as saying "recovers after a long rest" or similar. Who's to say when someone will get a good night's sleep or will toss and turn? Why wouldn't that be just as unpredictable as it is for people in real life? The more specific rules governing when someone -- be they a spellcaster or someone else -- recovers a temporary resource like a hit point, spells or mana, were always unfounded and unnecessarily complex to me anyway.
 

This is pretty neat, as a condensation of the rules printable on two sides of a sheet of paper, and playable with a d6 and index card character sheets.

I had to infer that AN = armor number or the equivalent of AC, and DN = difficulty number rather than DC for Difficulty Classbut that was no trouble.

One section I'm having a bit of a challenge with is the magic. I feel like that needs a little more explanation. I appreciate the idea that you improvise the exact spells/interpret them by the guidelines, but I'm not clear, for example, how you differentiate a light, moderate, or heavy damage spell. Is the idea that you need to roll a higher DN, say, 5/6/7 for light, mod, heavy spells? Control spells specify that they should be DN 5-6 for 1rd of disabling, which gives a guidelines to go by, but the other spell types don't. Cantrips are mentioned as being free (not limited in slots), and under Wizard one of their abilities under "spellburst" is "Firebolt(1)", so I'm guessing maybe the intent is that 1 damage cantrips are only DN5 to cast and have no usage limits?
I appreciate that, especially your comments on magic, but I'm not sure if adding more detail there would result in a better gameplay experience. I almost just want to leave that up to the DM's discretion.

I find it telling and interesting that your inclination with regard to DN, light, mod and heavy spells, and damage and usage limits for cantrips, were exactly what I had in mind. Those are exactly how I've been playing them. So, perhaps the rules on magic are clear enough such that most people would intuit them similarly?
 
Last edited:

This bothers me a lot - "Written with ChatGPT" ? There are a LOT of games designers out there creating games without AI help and making them available free or for a charge within the OSR community.
I do realise that people who are not able writers can still have cool ideas and ChatGPT is a tool to enable them to get their idea out there but I would encourage people to look at the many, human-written OSR version that are out there.
So, I'm not saying what you have produced is not a worthy product, I am just wondering if you could have written it yourself?
 

This bothers me a lot - "Written with ChatGPT" ? There are a LOT of games designers out there creating games without AI help and making them available free or for a charge within the OSR community.
I do realise that people who are not able writers can still have cool ideas and ChatGPT is a tool to enable them to get their idea out there but I would encourage people to look at the many, human-written OSR version that are out there.
So, I'm not saying what you have produced is not a worthy product, I am just wondering if you could have written it yourself?
Let me start by saying that the negative comments with little more argument then "But it's made by humans, thus better!" argument is going to drive more people into the arms of AI/LLM generated content.

AI/LLM are tools, tools used by humans. Those humans decide how to use the tools, and what rolls out depends on those humans. That does NOT mean that what rolls out is (any) good or well thought out. Something like this is probably more of a thought experiment then anything else, as many before you pointed out flaws in the system and things that were missing. This is more of a show of what an AI/LLM can throwup with little input, finessing and/or human editing.

I've been using Deepseek (r1 free) for generating room descriptions for the Undermountain adventure (5e Dungeon of the Mad Mage) from the little context that was given for each room. Just getting an acceptable output took a while to figure out, things from atmosphere, dungeon construction and general features for each level. And even then getting out exactly what I wanted wasn't happening, technical issues, memory issues, things I didn't like. Getting an AI/LLM to spit out exactly what you want in the style you want is like herding cats. You need to realize that good is good enough, and when it isn't, manual editing and/or rewriting entire sections is your friend! I've done the first five levels + Skullport, that's a few hundred descriptions. Could I have written those 100% myself? Absolutely, but how long would that have taken? Would it actually be better? Very long and probably not, unless I spent even more time on it. It still took a LOT of work. I then put it through an AI powered text-to-speech service, also a lot of work (and money to get a good quality output), but still less and better then if I did all the voice acting myself. My players really like this and didn't really notice (yet) that the text that's being read is AI generated.

Last week I started with Midjourney for image generation, getting that to function exactly to my wishes is even worse, it's like herding angry cats! Again, quite a bit of money, quite a bit of time. For the first session I settled on images that illustrated not exactly what was there, but gave an overall atmosphere. Sure, they figured out quickly that they were AI generated images, but they don't care, they like the atmosphere it creates. Midjourney gives you far more tools to edit and specify exactly what you want, but again you need knowledge and skills to make it work better for you, taking up even more time to learn new skills. But that's still less time then drawing/painting it yourself, even if you've been trained for that (like I have been), now imagine people that have not been trained for that. And again, will my human made illustrations be better then what the AI can generate? Probably not. And hiring someone else that does it in the exact style you want is going to be expensive, and will take WAY too much time to complete. My intention is to have atleast one illustration for every room/encounter, that's going to be hundreds of images.

AI/LLM is a tool for me to create more immersion/enjoyment for my players in a more realistic time/money budget. While leaving me time to create more bespoke adventure components for the adventure. In the same way that Waterdeep: Dungeon of the Mad Mage saved me oodles of time/money as a DM. I see no one complaining about published adventures being the ruin of smalltime bespoke illustrators/writers DMs... Not everyone likes published adventures, to be honest the quality/writing is actually lacking a bit with most of the WotC stuff anyway, the Deepseek writing is an actual improvement over that... ;) And in the past I would have written my own adventures, have done many a time. But having a full time job that's brain intensive tends to mess that up, being older I also have more interests that distract me, as well as responsibilities. You don't have to use published adventures, you don't have to use AI/LLMs, you don't have to use D&D(5e).

That said, there are MANY, many RPGs I would use before ever considering using an RPG created by an AI/LLM. Even one page systems by writers I trust WAY more then ChatGPT. I would LOVE to spend $100-$200 to get a digital collection of 500 excellent illustrations made by human illustrators in style X about exactly subject Y, but not many others would, would they be able to sell a 1000+ of those bundles at that price? Probably not, because if they could, they would. In my case the artists that I want the style of are either dead or retired and at an advanced age (or have moved on from the style they did many decades ago). Also keep in mind that not everyone has a well paid job which results in a decent amount of disposable income, and in many different parts of the world, what's considered 'good' makes a magnitude of difference.
 

This bothers me a lot - "Written with ChatGPT" ? There are a LOT of games designers out there creating games without AI help and making them available free or for a charge within the OSR community.
I do realise that people who are not able writers can still have cool ideas and ChatGPT is a tool to enable them to get their idea out there but I would encourage people to look at the many, human-written OSR version that are out there.
So, I'm not saying what you have produced is not a worthy product, I am just wondering if you could have written it yourself?
Sorry it bothers you. I get it, but the world turns whether we're bothered by it or not. Horse and carriage companies were bothered by the invention of the automobile, but that didn't stop their industry from disappearing within a few years.

AI is here to stay. The prospects of it bother me too -- greatly -- but my discomfort with it is irrelevant.

I'm able to do more in seconds now with ChatGPT than I or any other human could do in weeks. I'm still mortal. I don't have the time to ignore a tool that makes me more than 100 times more efficient.
 

So, I'm not saying what you have produced is not a worthy product...
The game I shared isn't "a product." It's a TTRPG created by a machine that analyzed the entire history of gaming, read through hundreds of thousands of pages of game mechanics online and millions of community discussions like ours in 10 seconds and produced a complete game from its analysis. In 10 seconds.

Not to sound melodramatic or scary...but the machine is coming, and it's coming for entire industries, devaluing human creativity and effort because it can do much of what we can better. And make no mistake, it's better. Within 5 years it'll have surpassed the greatest human writers and artists who have ever lived and create work indistinguishable from human output.

Sorry! It's true though.
 

Not to sound melodramatic or scary...but the machine is coming, and it's coming for entire industries, devaluing human creativity and effort because it can do much of what we can better. And make no mistake, it's better. Within 5 years it'll have surpassed the greatest human writers and artists who have ever lived and create work indistinguishable from human output.
And now you're just trolling! 🧌

AI/LLM can do certain things, other things it absolutely sucks at, and there is no way of really improving that due to how the whole LLM thing actually works. 63 years ago the Jetsons predicted we would have flying cars, but they didn't predict the smartphone... And Elon Musk has been announcing full self driving (level 5) in 1-3 years for Tesla since 2013... AI 'Fans' have been 'predicting' that seriously next year AGI will show up! And Bitcoin will go to the moon! ;)

Even IF LLMs could write better then any human past/present/future, it doesn't matter. We have put man on the moon, that doesn't mean we have a space base or human cities up there. Just because we put a couple of billionaires into space, does not mean that it's economical to do or even wanted by the general populace in our lifetime. We can have fully automated production lines that make furniture that will outlast the next 10+ generations, but we either buy relatively cheap folded paper products at Ikea or make our selves or pay people to make hand made furniture for extremely high prices. We have very good movies, but way too many people would prefer to watch Big Brother. We have very good literature, but way too many people prefer dimestore novels or pulpy romance novels that they sell everywhere.

Foxconn stated in 2011 that they would replace their workers with one million robots in three years. 5 years later in one factory they replaced 60k workers with robots, leaving 50k human workers. We're now 14 years down the line instead of 3 and does Foxconn have a million robots by now? Maybe, but they currently still have more workers then they had 11 years ago working for them...

AI/LLM and generative AI is amazing! But it's not "let's replace all humans with AI" amazing, many, many have jumped the gun and are now dealing with the mess or will be dealing with the mess in the future. From chatbots that hallucinate horrible solutions for customers and don't recognize their own products, horribly messy codebases that will have security issues up the wazoo and will eventually implode!

And then there's the cost aspect. ChatGPT you can use a lot for free, or you pay $20/month for 'Plus', $200/month for 'Pro'. But they still lost $5 billion. It took Amazon about 7 years before they made a profit, 10 years before they made a consistent profit. But it took Spotify 17 years, Netflix ~10 years. Models are getting more expensive to make and run, they need more/better hardware, while consuming more and more power. The hardware is not keeping up, nor is the power generation. Eventually this bubble will burst, and I doubt that this will happen before Open AI makes a profit.
 

Remove ads

Top