AI Testbed
  • Register









My thesis was completed and handed in last month, and I had my viva last week. Academically it has exceeded every expectation, but no one is interested in that, what's important is the results and analysis from the project. Now, I don't think I can share the full thesis until after my mark is finalised (which is my plan), but I can summarise my findings now. Essentially the result was that the computer tuned AI sucked, but wasn't a complete failure. This first attempt was a shot in the dark though, and honestly things worked better than they had any right to be. Details below.

The final conclusion was that everything worked great, except the GA fitness algorithm. This basically just tracked a few statistics about each game and then turned them into predicted survey responses. The algorithm was based entirely on guesswork, I had no previous studies to work with, and no experience with game analytics; this study has helped with both of those. So to realise the benefits, I think all we need to do is sort out the analytics side and the automated analysis of that. The analytics used in this study were all home made, done on an extremely tight schedule (i.e. a couple of days) and I know for sure there are better ways of doing it, unity has everything you could want here. That just leaves the automated analysis.

On the automated side there's no standard system for doing it, but there were some survey metrics that did have some level of success with prediction here, certainly nothing particularly accurate, but something at least. I'm certain that with decent analytics info, and some calibration tests, we could produce a solid fitness test. The thing with this is that the fitness doesn't have to be extremely accurate or detailed to produce valuable results. If the automated system can get an AI this complicated half tuned, in an afternoon then that's a big improvement over what we have now. Additionally there's not just one "perfect" tuning, there's going to be multiple generally good solutions, and countless more specialised tunings for different tastes; you just can't experiment that much doing it manually.

End result is that I'm excited for this technology, I'll be pushing ahead with it at some point in the future for sure. Do I think a GA tuning is necessary or the right idea for the early Hunter's Moon releases? No. Do I think it'll be invaluable at a later stage, especially for the general simulator project (whatever we end up naming it)? Yes, absolutely.

I'm hoping the thesis will be up in the next month or so, whenever I get my final grade. There's a couple issues, typos, etc. with it but it's presentable enough.

Add comment

Security code

Hunter's Moon

For details and all the latest news on Hunter's Moon development:

See Here.

AI Testbed

For details and all the latest news on Foamy's AI Testbed development:

See Here.

GOM News

For all the latest news on the studio.

See here.

Foamy's Blog

To see recent news from Foamy, the lead developer for Hunter's Moon and the AI Testbed.

See Here.

Latest News

  • New quick status page

    See it here: Status

    I am really bad at keeping the blog up to date, and my plans have to change fairly frequently due to personal, professional, resource, and industry situation changes. This page is an attempt to have something that is quick enough to update that I am able to do it regularly. It also has very little information, so I don't have to worry about my confidence in prediction, or sharing things that are not ready. Every project that has been started will have an entry here that clearly states my current assessment of its status.

    {jcomments on}

  • New plans

    I think it's about time to talk about some longer term plans again. I've posted a bit about the current work on the ranged weapon interaction system, and also a bit on the HoO collaboration, but it's been a while since I said anything relating to our primary game ambitions.

    We have some new plans, but it's nothing drastic. The previous plan was to develop Hunter's Moon (the werewolves) and then using all the tech it used to build a general purpose combat simulator (GPCS, as yet unnamed). That plan heavily depended on having a year or so of solid full time development; probably funded through a kickstarter.

    Read more ...  
  • A sign of things to come

    With the musket and pistol kindly provided by the hearts of oak team all attention now focuses on the ranged weapon interaction system!

    I am super excited about this, and I think everyone will love what I've come up with.

    Time-scales are fuzzy, but hopefully not as long as it took to put the AI in place last year.

  • Colaboration with the Hearts of Oak team

    This is a big announcement that should have been made a while ago, but now that the AI is implemented on the combat testbed it can't wait.

    A few months ago I approached the lead developers for the Hearts of Oak project with a proposal for collaboration. Basically to provide our combat engine to them in return for publicity, assets, and the joy of contributing to such an important piece of work.

    A summary of the agreement is as follows:


    Read more ...  
  • Combat Testbed Updated to 1.3

    Big update with the addition of an AI bot to play against in single player! This is the first pass on the combat AI, a very important milestone.

    Full changelog:

    1.3 - 2015-01-03
    Added an AI bot for single player
        Added as a variety of the orange arena, in addition to empty, and one with training props as before
    Cosmetic improvements
        Added body tilt/lean during turning
        Made players collide again

    Get it here: Combat Testbed