Collectible Card Game Headquarters

Hiiiiiiiiiii I'm wondering if theres any way to change the programming of the AI in DotP 2015.

I like the game and most of its aspects but what turns me off is the AI that sometimes does very stupid things. For example it doesn't seem to understand or correctly factor in creature abilities (Elder of Laurels, as a specific example). It'll do game-throwing things like attack with an important creature just to have it die to an obviously superior blocker. When this happens it feels like I'm playing against an opponent who is tossing Lightning Bolts at itself... and that takes all the fun out of playing against AI.

I don't know a whole lot about the AI other than an article I read describing what it does - that is, check every possible way to play and find the one that seems most promising, given how the board looks. I'm not sure if it takes into account what the player COULD have in his/her hand, but it doesn't matter. I just want an AI that can look at the board, look at its hand, and consistently figure out a good way to play, at least over the span of 1 turn.

Oops I think this post belongs in the Programming section of the forum? Can someone move it there if possible?

Also I'm not new here, I just can't remember my old info! I have a little experience in making custom decks/cards in the older DotPs, but I've never tried messing with the AI programming. I really hope it's possible!

I'm a bit surprised to hear that, do you have the AI difficulty set to the highest level? I haven't played much of 2015 yet, but in previous years it didn't make catastrophic immediate decisions like that. Generally speaking the AI would be pretty good at conducting combat in a sensible manner and would be good at understanding how to use instant spells and abilities at the last minute when they can make the difference. To me the main shortcoming of the AI has been that it would overvalue immediate creature advantage and would typically use resources at the first opportunity when more value or advantage would have been achieved by holding on.

I'm playing on the hardest setting. The AI doesn't consistently make these types of mistakes but once in a while it does, and it's incredibly frustrating when it happens. Maybe there's something wrong with the specific programming of the card (Elder of Laurels) that causes the AI to not recognize/consider things about it? I remember having the AI play a custom deck, and it would have no clue what to do with the new cards.

It wouldn't be the first time that a card had issues in it, fortunately we are in good hands here! Should 2015 get cracked open enough to get the modding scene rolling again this year then I imagine some kind soul would make an equivalent to RiiakShiNal's 2014 Core Fixes Mod, and such things in the existing cards will get checked over by the experts along the way.

Absolutely. I'm not sure if a core fix would be the number one priority when we can mod, but it will definitely happen at one point or another. I expect Riiak will be the one to make it again, because he tends to be the most focused on making the game work how it should (whereas the other major players tend to focus on adding new and interesting cards and/or decks). If he doesn't, someone will (myself being a possibility, if no one else does).

It's not necessarily that I'm more focused on making it work the way it should, I just picked did the Core fixes for DotP 2014 because my mods were the most affected by core bugs at the time (due to Annihilator allowing the sacrifice of mana tokens). If you look at DotP 2013 thefiremind was the one in charge of core fixes there. So I imagine the first person who really starts in on the core fixes for one reason or another will end up as the person in charge for DotP 2015 so it could be anyone at this point.

Regardless, who ends up in charge of managing the core fixes we all still help out, just like in previous games. If you look at the DotP 2014 release thread there are lots of contributions from other people (both reports and fixes).

I see. I wasn't very active for 2013 modding, so between the manual mana, objectDC, and core fixes, I assumed that was your norm. But yeah, that makes sense.

Well, I created the Manual Mana because I wanted to be able to use mana sources that couldn't be coded by default, ObjectDC because I needed a way to handle the values for Manual Mana, and Core Fixes because my code was most affected by bugs in the handling of invisible tokens (game classified them as permanents even though according to magic rules they aren't because they never had a type rather than simply losing their types). It just kind of worked out that way.

If the AI doesn't know what to do with a card, it's the card's problem and not the AI's right? I remember back in DotP 2014 that the AI would not know what to do with custom-made dual lands. If a card is well programmed on its own it'll work with the AI? I've always wondered if the AI used any additional info beyond the programming of the card to make decisions (like hard-coded stuff for each and every "official" card that would guide the AI in using it well.)

As a side note, I'm sad they didn't include Restoration Angel in 2015!

Im not the best person to answer this but not always is the card problem, last night i was playing without mods and the AI used Act of treason, and even before the spell resolved he used something else and killed the creature he was going to get ( he wasn't lacking life or mana so if he wanted he could attack me with my creature and the kill it on my turn)

That sounds like an extreme case of the bad AI decision making I'm talking about.