The free AI-service by ChatGPT have created a lot of headlines in the news recently, as students use it for 'cheating' and a debate if AI it self is 'dangerous' for us and can be trusted. I got a little curious about GhatGPT, but I don't have any homework anymore to ask for helping. -So I asked the ChatGPT the following question to see if this AI-service could make any sensible free wargaming rules for army men:
"Can you write a free fast playing easy wargame rules for army men. with movement in inches and reduction for obstacles. Can you make it with a firing chart for pistol, smg, rifle, siper rifle, lmg, hmg, bazooka, flamethrower and mortar with different ranges and d6 hit probability. It also would need some hitting chart to calculate effects on fire with a d6. Can you also make a d6 based morale chart. I want the rules to contain rules for hand to hand fighting. I also want to have rules for vehicles like soft skins, armored and tanks regarding range and firing and how to calculate hits."
The ChatGPT instantly replied:
"That's it". -If you're not happy with them, you can use the chat-function to ask for further details and give specifications for the rules and get even more details for your rules. -But these rules first suggested by my request seems surprisingly pretty playable to me, so I'm happy with the ChatGPT's inital suggestion. I wonder how well the rules work, so I'll have a testplay of these rules soon. A battle report will therefore shortly be posted on this blog so stay tuned. Please subscribe and follow this blog for updates.
Go to ChatGPT if you would like to have a go for your self. If you manage it to suggest some sensible rules, I also hope you share them somewhere (on your blog, in a Facebook group or in the comment-section here). A 'warning' though; Using ChatGPT can be very addictive. Enjoy!
I love ChatGPT! I use it for DnD, for tightening up my website copy (not for my blog), even for writing lyircs.
ReplyDeleteThis is an interesting little set of rules. Two things leapt out at me.
The notion that LMGs can fire twice per turn, but no other weapon can was the first that gave me pause. Does that mean it rolls two dice instead of one? Or does it mean that it can fire and then fire again later in the same turn? Since there's no turn sequence, I'm inclined to interpret it as the former. Which then brings up the question, why can't HMGs or SMGs fire more than one die?
However, I realized, that difference may be baked into the ranges and "to hit" values. It's very clever really. The HMG hits on a 3+ out to 24", that's 6 more inches and a greater chance to hit on that one die, at the loss of the ability to move and fire in the same turn. Now, statistically, I have no clue as to the probability of rolling 4+ on two dice vs rolling 3+ on one die. Nonetheless, I think it's an intriguing way to handle the differences between weapons.
The mention that the sniper rifle can only shoot once per turn is perplexing. Again, as there's no turn sequence, I would assume that unless otherwise stated, a weapon could only fire once anyway. So, why mention it? Now, I think talking about the intentions of the AI is a bit of a stretch, at least with ChatGPT, but, if we attribute the rules to an anonymous human author, I think we'd certainly be inclined to consider what they may have been aiming for (pun intended).
I haven't come up with an explanation that I find satisfying and that doesn't bump up against the LMG rule, but the process, and the fact that it comes from AI generated ambiguity is itself exciting.
Thanks for sharing these rules!
Thanks John! I noticed some of the same things as you mentioned, and as a 'first edition' the rules are not perfect. I considered to ask for further details and more specific rules about turn sequence and when a 'morale test' was needed, but on the other hand I wanted to just try out the ChatGPT's capacity to write a ruleset without too many corrections and further specifications. -So that's why I accepted the rules initially suggested.
DeleteDespite its shortcomings I'm about to play a game using these rules just to see how playable they actually are, even if they lack some basic rules about turn sequence etc. I'll guess I just have to improvise while playing. For example roll for initiative every time and for LMGs I'll just roll 2 dice once (so it shoots only 1 time but at a higher rate of fire and therefore more likely to hit something). For morale tests I'll just follow the example of other rules by testing it when te units are halved.
I plan to use your car search-scenario again for such a small game (I really liked that one).
I got curious about your question John, but couldn't figure it ot my self, so I asked the ChatGPT. I have no clue if it's correct, but this is how it answered:
Delete"To determine which outcome is more likely, let's analyze the probabilities of rolling a 4+ on both six-sided dice (d6) and rolling a 3+ on one d6.
The probability of rolling a 4 or higher on a single d6 is 3 out of 6, since there are three favorable outcomes (4, 5, and 6) out of six possible outcomes (1, 2, 3, 4, 5, and 6). Therefore, the probability of rolling a 4+ on one d6 is 3/6, which can be simplified to 1/2 or 50%.
To calculate the probability of rolling a 4+ on both d6, we need to multiply the probabilities. The probability of rolling a 4+ on the first d6 is 1/2, and the same goes for the second d6 since they are identical and independent. Multiplying these probabilities gives us (1/2) * (1/2) = 1/4, which is equivalent to 0.25 or 25%.
Comparing the probabilities, we find that rolling a 4+ on both d6 (25%) is less likely than rolling a 3+ on one d6 (50%). Therefore, it is more likely to roll a 3+ on one d6 than a 4+ on both d6."
Hi...One suggestion is you might want to add a Sequence of Play to these rules. Did you have to create those nice neat boxes for the charts? Any results I've received from ChatGPT have been just straight text.
ReplyDeleteI noticed that the ChatGPTs initial suggestion for the rules lacked several things, as the turn sequence. On the other hand I wanted to just try out the ChatGPT's capacity to write a ruleset without too many corrections and further specifications, so that's why I accepted the rules initially suggested. I also plan to play a game using these rules just to se how playable they actually are. Then I'll just have to improvise, perhaps rolling for initiative every time/turn.
DeleteThe charts did not come with the 'boxes', but in the ChatGPT's suggestion the text had such a layout that it was almost just like drawing a line between them to get the boxed charts.
I asked the ChatGPT the same question again, and then it suggest some slightly different rules. -But with the chartbox.
Delete