user avatar
Jason H
@crasong
Sometimes I do funny voices. Opinions are my own. He/Him Prev: Software Engineer on @LeagueOfLegends Champions team. Bel'Veth, K'Sante, Briar, Ambessa
Planet Earth
Joined July 2010
Posts
  • user avatar
    That's a wrap! My 4th and final champion, and I'm so proud of what the team accomplished with Ambessa. It's bittersweet that my time on LoL was cut short by layoffs, but I have no regrets knowing I gave it my all. GGWP, onto the next adventure!
    Ambessa hits PBE today. Hard. With twin drakehounds.
    00:00
  • user avatar
    As promised, I'll be sharing some info about Bel'Veth's tech - challenges we faced, and the solutions we came up with! First up, let's talk about her voidlings
    @_nyanbun_ did Bel’Veth need any crazy new tech
  • user avatar
    Replying to @crasong
    we began exploring the idea of "Bel'Veth spawning void lane minions" very early on. The goal was to be able to convert a dying lane minion into a voidling, and a voidling would behave exactly like a lane minion.
  • user avatar
    Replying to @crasong
    Another of Bel'Veth's key defining features is her uncapped attack speed. Historically, reworked Kayle has had the most complex auto attack setup, due to her evolving states and similarly high attack-speed identity
  • user avatar
    Replying to @crasong
    A lot of love and care was put into these voidlings, and I feel so grateful to have worked with people who, besides being amazing at what they do, go above and beyond to iterate and get things right!
  • user avatar
    Replying to @crasong
    However, VFX was unable to tie their particles into our animation blending system, which meant that the swipe VFX would continue playing even if her auto attacks were canceled midway. Here's me spamming attack and stop to illustrate this
    GIF
  • user avatar
    Replying to @crasong
    Voidlings would have the same stats as a lane minion. So everything from health, to resistances, attack range, etc. So a melee minion would become a melee voidling, and so on. Simple enough that we had a working prototype at the start.
  • user avatar
    Next up, let's talk about one of her most endearing aspects -- Her SLAPS!
    @_nyanbun_ did Bel’Veth need any crazy new tech
  • user avatar
    Replying to @crasong
    So instead, we tried to spawn regular melee and ranged minions, and simply disguise them as voidlings. This didn't work for two reasons:
  • user avatar
    Replying to @crasong
    1. Spawning and disguising minions was expensive. Spawning a whole wave at once would cause framerate drops that are unacceptable by our standards. 2. Lane minions have bonus stats that grow overtime, and these bonuses were siloed. This meant we were spawning level 1 minions.
  • user avatar
    Replying to @crasong
    Until now, we had been spawning a voidling, and then copied each stat from the dying minion over to the voidling, and that worked fine. However, for cannon and super minions, we can't simply copy their stats over, since they are significantly more powerful.
  • user avatar
    Replying to @crasong
    The solution was to refactor our barracks system to allow designers to spawn custom lane minions easily with very little script. The result is that spawning ~100 voidlings at once would be similar to spawning 15 in the original prototype.
  • user avatar
    Replying to @crasong
    Our first hurdle came from the decision to cut the cannon/siege and super voidlings, and have just the melee and ranged/caster ones, in order to address gameplay and scope feedback from stakeholders.
  • user avatar
    Replying to @crasong
    Another bonus is that the refactor sets us up to create new kinds of minions in the future more easily!