AI logic: over-reaching its cards / and failing



  • TLDR: why does the AI over reach its resources rather than build where it is, losing a key position to an opponent who clearly will take the spot they are competing for on their next turn !?


    OK, So the AI builds a road towards a a key spot, in order to block another player also trying to build there.

    So, it could build a settlement right where it is, not the ideal spot, but effectively blocking the other player from building the next spot over, or it could build one road over and build exactly where the other player wants.

    The AI as expected attempts/makes a few trades, has enough to build right there... but nope, it builds one more road over, then tries to make a few more trades, but either doesn't have enough to build a settlement, or can't get the trade it needs to build.

    Why is the AI so greedy/stupid. It should have tried all the trades before building that road, if it had enough then build road and take the spot, if it doesn't, build where it is. Rather, it moves prematurely, over-reaches it's resources or ability to trade, and then loses the opportunity to build / and loses the opportunity to block.

    This was extra frustrating for me, because I needlessly traded resources I needed, expected the AI would block the other player (who was way to strong)... but nope, it failed, because it is stupid and for some reason isn't programmed in simple logical decisions (do what you can / don't try to build to a spot you don't have enough for, and will clearly be beaten to by your opponent)

    Does this need a screenshot !? Please fix the AI. It can't be counted on like humans can with strategic trades/blocking players, and is one of too many reasons the AI unbalances the game. [It's not like I'm even asking for the AI to predict future outcomes, just see that what it's trying to do obviously wont work in the immediate next round]



  • This will not be fixed any time soon. The developers are probably working on another expansion that will also have the same issues. I guess they try to get GoT Catan out by the time the last season begins...



  • @Stroom Well, I figure if I can just roll out problematic scenarios, they can ate least build test cases. Although it's not really clear if agile approaches are in use :-)

    I didn't know there was a GoT expansion/version... interesting! Ideally they'd roll in critical and/or a block or related bugs/enhancements with each expansion release, but I don't know how much capacity they have.



  • katan.io is moving a lot faster, at least based on the bi-weekly bug fix reports.



  • @Stroom Even a bit of time fixing minor bugs and/or small features would certainly go a long way to increase confidence (although I know many bugs have deep implications, and many platforms makes bug fixes a challenge) If Katan.io is web based, that's a more focused platform (I'll check it out!)


Log in to reply