Jump to content

Gijs-Jan

Development Team
  • Content count

    296
  • Joined

  • Last visited

  • Days Won

    3

Everything posted by Gijs-Jan

  1. If you guys manage to get a save of the behavior then I can probably get it fixed quickly. Otherwise I have to spend my time playing the game to encounter it, as it is a bug that doesn't seem to happen often (reliably that is). And with the deadline approaching I am currently focusing my efforts on the Harridans taking flight.
  2. Good to hear :-); Except of course the CTD!
  3. Just to be clear, this "feature" was only to speed up movement, and should not affect gameplay in anyway. On a lighter note, I submitted a fix today so it should be fixed in the next build.
  4. Well, that's a bug then, and I'll fix it :-) If you have a save game that showcases the behavior, that would save me quite some time.
  5. Gijs-Jan

    V22 Experimental 7 Released!

    Funnily enough, the value calculations weren't the big problem. The primary problem is the path calculation. It is the major bottleneck as the pathing system wasn't designed with heavy AI use in mind. Especially pathing to a location which was ultimately unreachable had a high cost; which the AI now has added checks for to exclude.
  6. Gijs-Jan

    V22 Experimental 7 Released!

    The AI turn times were improved by several fixed/enhancements: - Tiles which are considered unreachable are no longer taken into account during the value calculation. This solved turn times for maps which have a lot of "black areas"; bases defense, final mission. - During path calculation; the AI is now a lot more selective in which locations it may target for pathing. - A small caching layer was written around the pathing algo to ensure that if the state of the game doesn't change, it just returns the same path for a given source & target. - Some animations are skipped when we are sure the player cannot observe them.
  7. Gijs-Jan

    Secret LOS modes for testing!

    No; not directly as the AI uses a lot of inferencing on where you are. It never used the "global LOS" for this, only the actual LOS of the squad. So, unless this got bugged in the process of implementation, there should be no consequences.
  8. This was indeed the case; and at some point the decision was made that some aliens should stay inside the UFO at all times. Anyway, this should be fixed in the next build, along with a far more diverse number of AI settings. For the modders this will mean that the AI weights can be set per Race, per Rank, per Behavior Type.
  9. Gijs-Jan

    Last aliens in UFO- AI behaviour

    This should be solved in the next build
  10. WARNING: Peek behind the curtain! Ok, so to provide some insight into this: The aliens within a UFO have a slight tendency to stick to upper floors, with higher ranks having a higher tendency. (FloorPreference aiprops.xml) This should be (somewhat) negated by their incentive to search unexplored tiles for Xenonauts (ProbabilisticSearch; aiprops.xml), and even further negated when they can reach "known" targets (Squad LOS), or receive requests for backup (AI attacking Xenonauts). The behavior (sticking to upper floors) is amplified by the need to stay out of unexplored danger zones (Teleporter with no cover). So by tweaking the weight of the Probabilistic Search, we might give the AI more incentive to come downstairs (more risky behavior). Or we could give the AI knowledge about the UFO being breached, which would make it want to move downstairs, and then trigger Squad LOS / backup requests.
  11. If an alien is surrounded by fire on all tiles, then yes it might still move through fire. However, there should be some pretty heavy "incentive" for it to "want" to move through fire.
  12. I'll take a look at this. There should be some easy implementations that should fix the worst of it.
  13. This should be fixed in the new release. This isn't an AI bug, so you'll probably want to post it in the ground combat thread.
  14. Thanks for reporting it, but it was a game mechanic problem due to some debug code being pushed in. :-)
  15. Gijs-Jan

    OP Reapers

    My roommate decided to print them out. Btw, Reapers shouldn't approach unless they are certain of an attack/kill. The cover/hiding behavior should be easy to implement by adjusting the pathing weights. *Edit: wrong thread*
  16. *Wrong thread * Btw, Reapers shouldn't approach unless they are certain of an attack/kill. The cover/hiding behavior should be easy to implement by adjusting the pathing weights.
  17. I've been reading through the posts, and I actually agree with both opinions. Evidently, the more aggressive AI is being received well, it's the side effect (empty map) that's unwelcome. I agree with kabill in the sense that the best way to solve it is to either add, or force some stragglers to remain behind. I'll take a look at it and try to present an alternative to Chris.
  18. Gijs-Jan

    V22 Experimental Build 1 Released!

    This is on the latest experimental branch? And it is primarily passivity?
  19. Ah, fair enough! Forget I said anything :-)
  20. Which is what I tried to express with the transforms remark. :-) This is due to how some RNGs manipulate the bit representation of the number; which gets distorted if you discard sections of it in the transform. Anyway, if I have some spare time somewhere I'll take a look. Other than that, our first hunch is still (if it exists) some problem in implementation. The way to showcase this is to present some save where we from our end can see the problem happen. :-) Btw: If you encountered this during game development, you should've probably switched your RNG implementation. (Previous iterations of Java's Random were notorious with skewed random results.) IMHO; I always use MersenneTwister; making all points moot. (I need "better" RNG for most of my AI implementations, as they are academic)
  21. This is true for "heavy" (Cryptography, some AI techniques) usecases of psuedo randomness. I myself always roll my own RNG based on Mersenne Twister for any and all non-trivial use, however this really shouldn't be a problem for the accuracy calculation in Xenonauts. If something is wrong, it will most likely be through the use or transformation of the output of the RNG.
  22. This really shouldn't be happening, the only way the AI can use psionics on you is if you are visible by some alien, or if you don't move at all.
  23. Gijs-Jan

    whats goldhawk's next project?

    I'd almost work for free to make that game.
×