Forum adverts like this one are shown to any user who is not logged in. Join us by filling out a tiny 3 field form and you will get your own, free, dakka user account which gives a good range of benefits to you: No adverts like this in the forums anymore. Times and dates in your local timezone. Full tracking of what you have read so you can skip to your first unread post, easily see what has changed since you last logged in, and easily see what is new at a glance. Email notifications for threads you want to watch closely. Being a part of the oldest wargaming community on the net. If you are already a member then feel free to login now.

Hello Everyone,

I am currently in the process of creating a system involving 40k. I have a reddit post with the same name as well, but I have decided to also seek help from these forums, as they have helped me find some answers on some of the questions for 40k while i was building the system (lurking of course). In any case, I will post some science stuff in the next paragraph about the system I am creating. If you're not interested in that just hop it over to the paragraph after that.

I am creating an AI system that should be capable of dividing a limited amount of resources over a complex environment prone to change in the best possible manner. 40k is a great example of a non-costly way to test the hypothesis and the system. The system does the following: whenever it is presented with an enemy army it tries to create and adapt it's knowledge to produce the best possible (to it) answer. This is by no means a commercial system, and will never be such. It is a simple test if the hypothesis and the AI could potentially work in such an environment effectively. Currently the system is aware of Space Marines, but not much else unfortunately. There are plans for the future, but I am on a strict deadline.

There are two surveys that I have created that I would need your help in completing. The first deals with a total of 10 matches. In each match a "problem" army was presented to the system, and in each match a "solution" army was created by the system with a suspected win chance. Your task would be to take a look at the two armies and agree or disagree with the system, providing comments and reasoning. You are of course welcome to actually play test if you have the time, but I know it is time consuming to do so and I will not ask you to do this. I also understand that these armies may be simple and even downright horrible! I have created them specifically to fit into the Combined Arms detachment, and not to fulfill a specific role. However, any and all input is appreciated!

The link to the first survey: https://docs.google.com/forms/d/1jqPFGIBhL0lYeq5ZLloh6Hnb7UwFLcCrkexJMMELlPo/viewform

The second survey deals with explanations. It is shorter as a survey, but with a bit more to read. After creating any army the system uses self generated explanations to try to reason around what it did. I need to see if people find these explanations useful or not. Again, all input is appreciated!

The link to the second survey: https://docs.google.com/forms/d/1Fgu0K3wj03vK0tgf23mlOsvvLp7qLOuiGYsP7G-_BVI/viewform

Either survey is not particularly long, and feel free to quit the survey at any point (none of the questions, except the very first are mandatory). Be sure to submit the survey before you quit if you have made any answers! The surveys are 100% anonymous, and only timestamp your submission.

If you have any more questions, I will gladly answer them. If you need more information, I will gladly provide as much as I can. If you wish to be credited for your help, private message me the date/time of submitting the survey, and what credentials you would wish to be addressed by.

Thank you for the help and for taking the time to read this!

P.S This AI is weak AI, and thus it is a Machine Spirit esque system, not an actual AI in the 40k sense of an AI system. Trust me, as It's creator I am well aware of it's stupidity. Chance of heresy is 0%.

I like the idea, but evaluating this much lists this big feels like work for me.

I like the idea, but evaluating this much lists this big feels like work for me.

I understand that it is, it's also why I'm asking all kinds of input, even doing one of the questions on the survey helps.

The reason I can't do this on my own (and unfortunately have to take someones time without any compensation, which sucks), is because I can't. If I did the tests on my own, my hypothesis holds no valid results, since I could be accused of bias (for example, I could make all or the majority of the armies the system generates win, just by playing poorly with the other armies). All i can offer is credit for you taking the time to do this.

I would work with smaller lists. Trying to evaluate a 2000 pt game is pretty hard.

Also, I would clarify what types of missions are being played and how much terrain there is. The type of mission can play a huge factor in who will win. Kill Point missions like Purge the Alien are going to be hard on the MSU armies, whereas objective-based maelstrom missions will favor them.

I would work with smaller lists. Trying to evaluate a 2000 pt game is pretty hard.

Noted! I hadn't really thought about it, so I tried to follow the general path the local gaming club (which is also lending me a hand with this project) is most used to (which is the 1850 point values). For some reason i figured it would be one of those things where the higher points were "seen" more than the lower points, and thus were weirdly easier to judge. Nevertheless, what do you think would be a good starting point? 1000? 500?

As for the terrain, I have a random terrain function that generates terrain (since different people like different terrain setups). In this case, the setup that was chosen was: 15% of the board is difficult terrain and 10% buildings.

For the mission, this is intended to be a objective-based missions, no mysterious objectives.

Thanks for the input though, I will make a note of the mission type as well (and do some research about it).