Don’t forget to check out DictatorOfRome.com for a daily dose of Roman themed logic puzzles – new features are being rolled out frequently so make sure to check back regularly!
Now to figuring out whether the rules worked beyond just my gut instinct honed over years of not designing, but at least playing, games.
Print and play
I didn’t have any cards designed as yet, so printing out a draft deck and playing with it was out of the question (and way ahead of where I was in my thinking anyways – how the hell would I print the cards? I still haven’t got to that). My next thought was to generate a deck online (I did ask ChatGPT for a long shot, not expecting anything, and it delivered on just that – nothing sensible), but I only had a vague sense of how I’d design them and then still no idea how to print them other than on a flimsy hone printer.
So how to go about it?
My only thought was to find some blank cards online and write on them (not fully fledged designs, of course, but at last names and other attributes). It didn’t take long and then I played a few hands against myself to see how it could work. The short of it was that the rules seemed to vaguely hold, but then I did design the game so was perhaps a bit biased. I only used the basic rules, though.
What next?
Simulation
ChatGPT had utterly failed me with design (note, even if I had gotten it to work for personal play testing, I had always planned to commission templates and art from humans), but I thought it might be a bit better suited for simulation. I had no basis for this. Nothing I had ever read suggested this was a good idea. Why not give it a go, though?
My initial prompt is below – feedback on the rules and an ask to playtest. I got done sycophantic response about the rules which I’ve largely ignored.
I did get some initial results as output below. It said there were 100 hands though which didn’t feel quite right.
So then I went down the rabbit hole!
I won’t bother to share the back and forth drivel here, but I sat fir a couple of hours at a time and kept asking it to simulate slightly different things – correcting it when it got rules wrong and adapting certain rule ideas when they appeared to create nonsensical outcomes (in one simulation, player 1 won 90% of quick games), I changed things around.
What it left me with was a set if rules I thought good enough to commit to a deck and that could be used in physical playtesting.
Next steps?
So all there was to it now was to get some cards and start plastering, which takes us to the next post where I’ll talk a bit about the card creation process.