Home Creators Posts Import Register Favorites Logout

Content

i hope those of you who celebrate christmas managed to have a lovely one! and if not... well, fuck it, new years is around the corner and if you ask me that's the better holiday anyway. we don't need christmas. 

my christmas somehow lasted three days, which some people might think was awesome, but it was actually a small fiesta for each one of them so no thank you. those of you on the discord server may have seen some of my complaints about part of the matter. if you didn't see it you missed nothing, since i have a whole lot of words about it that i could say and haven't. but, they say a picture is worth a thousand words. so, as an augment to what i said on the server, and for those who did not see that, allow me to show you the following cursed image as a brief but effective summary of my holiday:


at first glance, this may look something like a normal christmas tree. unfortunately, if you look closely or at all, you will notice that it is not.
what you see above is nothing more two blankets tied around a hoola-hoop suspended from the ceiling.
this was my mother's christmas tree.

i will leave it at that.

i did manage to get some work done over the holiday, mostly in continuing making a second combat model for starlight. i had it to a point that it could be structured in the animation editor despite not yet being finished. so, i just loaded that up to take a screen for you. unfortunately since i last saved and closed it, the animation editor has seen fit to replace most of her hair with various pieces of her skirt.
so that will need some fixing.
i'll just wait until it's done to show.

for reborn, it may not look like it on the development blog since i'm mostly updating that at the bigger piece milestones, but story is inching forward. i wasn't able to work on it over christmas, but we are getting there. i have one more large scene to do immediately, and then everything is bite-sized. hallelujah.

we are also commencing our apparently-annual tussle with the AI. unlike the past two episodes, there's no need for huge overhauls or tons of work writing code for moves anymore* so our problems are fairly localized, dedicated to resolving specific issues. some things we've hit recently:

  • the AI now understands prankster moves are priority moves ... somehow it didn't before.
  • the AI can now prioritize moves that are likely to crit vs players stacking defenses
  • AI is less likely to prefer moves that never miss unless there are accuracy drops
  • the AI will not switch due to not being able to hit an opponent if they have a move like Forest's Curse, etc.

* I actually do have some more code to structure for some moves, but that is solely for cases where the move could be used to target an ally. when we were going through moves before i initially tried to account for that, but we weren't sure exactly how it would look inside our structure to calculate VS an allied AI target, so we couldn't properly score and consider some things for these moves. thus, we decided to put it off until the other stuff was done. originally this would have been done during e18 but since it is basically so minor, we didn't see it worth holding the episode up for.
unfortunately, there is one battle in particular in E19 that is definitely going to require some allied targetting.
and i'm looking quite forward to seeing how that one goes for you all.

There are two other larger AI things we consider taking on this time. Currently we're unsure if these are actually worth pursuing for the amount of time it'll take, so consider these 'maybe's:

  • The first thing is a method to back out of wanting to switch Pokemon. Currently if the AI decides to switch, the AI is going to switch even if it has nothing good to switch to. This sometimes leads to really awkward cases particularly vs underlevelled trainers who begin spam switching. Personally, I consider fixing that a high priority since i think it's one of the few remaining cases where the AI really just bleeds out. Marcello considers it less important because as he says it is unlikely to change the outcome of those battles where all those underlevelled Pokemon are getting swept anyway.
    So as much as I call this a high priority, the problem is that in our current structure that we just created and have running fairly smoothly now, backing out of a switch is not really easy, and it would probably require a lot of work and rewriting to get working. It's sort of a yikes, looking at it.
  • In contrast to the above, the second thing is much easier to implement, but I'm not sure how worth it it would actually be. Right now our weather changing moves like Sunny Day, etc, have the same style of defined and conditioned code as most other moves. This is probably mostly fine, but we're considering updating that again to use a separate method like our Field changing moves do. This additional method would have the AI consider more thoroughly how much it likes the current weather, and decide better based on that whether or not it really wants to change it. In short, rather than just deciding if it likes the weather it can change it to, it also considers if the existing weather is disadvantageous for it or advantageous for the player.
    The thing is, given how infrequently Pokemon actually have weather changing moves, and given that they are already coded and probably fine for most cases, I'm not sure if this is even worth our time. So although it's practical to implement, I'm unsure if anyone would notice the improvement.

Ergo, random survey question: Have you noticed any pain points with weather changing moves in the AI, and if so when?
Based on y'alls feedback I'll determine if we really need to do that.
And for the switching thing... I may just look for alternative solutions ot the underlevelling problem in particular. Then again, it already strongly discourages underlevelled switching so zzzzz.
We'll get there.


One last announcement-- I've mentioned it in a few places, but Cass and I will be duo-streaming Rejuvenation v1`1 soon! It'll be through twitch rather than Picarto for you all, so anyone is invited. But I have yet to play Rejuvenation, so I'm looking forward to it... I guess it'll be like Reborn, but if I didn't have to have been the one to make it? Cool!

So look out for that one. Just in case anyone hasn't seen it, my twitch is on https://twitch.tv/amethystblack

And speaking of Rejuvenation our good friend Jannifer Janson has just launched his very own Patreon as of his new release. I'm really proud of him and he absolutely deserves the support, so if you also happen to be a fan of Rejuvenation, please consider pledging to him as well! 

Until next time. Hopefully soon!

Comments

Leone

That tree is an engineering M A R V E L 😮