Welcome to the official Pokémon Forums!

Click here to review our official Rules & Guidelines.

Troubles with the "Test Deck" AI

Options

hello! i have some feedback on the AI used for testing your deck builds. the main issue ive been running into is the lack of actual battling the AI does. i've used the "test deck" feature about two dozen times and nearly every battle ended because the AI would just keep drawing cards until it ran out. the first time or two it was mildly annoying, but it's continued to the point where it's genuinely very frustrating. there's no way to actually use this feature to decide if your deck setup is viable if the bot just draws cards until it runs out. battles continuously end without either player getting the chance to hit more than once or twice. you end up having to battle against a real person, which i think kind of blows when there's a feature that's meant to test your deck without throwing yourself into the pool yet.

i think the AI either needs to be adjusted to actually focus on battling rather than stalling itself into losing in 8-10 turns, or it just needs to be given a different selection of decks. i've specifically noticed it uses the Zacian & Zamazenta deck and stalls itself out with oranguru's ability, professor's research and birdkeeper, and this other one w/ hisuian zoarark vstar where it'll stall itself out with dark asset, professor's research, and various pokeball cards.

Comments

  • skaerkrow
    skaerkrow Member Posts: 20 ✭✭
    5 Agrees 5 Likes First Comment Name Dropper
    Options

    It seems like it just has a priority loop it checks and just does every single possible action in can

    ithink the order is like

    >drop a pokemon >play energy >tool >stadium >item >supporter >optional ability >attack >end turn

    and it'll go back to the top of the list every time it completes an action.

    it's not actually making any gameplay decisions it's just completing every possible action it can every turn. Because it's following that priority order for actions it'll often discard supporters to ultra ball without having another supporter for the turn etc.


    While it's no good for testing viability of decks, i think it is fine for testing if your decks' combos work or how long it takes to set up your board etc. It's a reasonable way to goldfish a new deck but outside of that it's kinda useless.