No announcement yet.

Seed Variance in Races

  • Filter
  • Time
  • Show
Clear All
new posts

  • Seed Variance in Races

    During recent CoNDOR races there’s been significant seed variance. Specifically in instrument dungeon generation. Meaning that two racers with the same seed can easily have different layouts, minibosses, item drops, and shop items. I don't want that to happen and would like it reduced or removed.

    To give some context, CoNDOR League Season 9 is currently in its fourth and final week. There's a Cadence of Hyrule Story Mode side event. A handful of people are participating, myself included.

    For those who don’t know about CoNDOR races, they're done on the same seed for all players. But it's with blind seeding - the players won't know the seed until the match starts. Blind Seeding is done to keep the things we love about racing roguelites (on the fly routing, variety in challenge, variety in builds), while reducing the luck factor that’s inherent in the genre. Also CoNDOR didn't start off using blind seeds and there's some history/discussion there. I'm not going to cover that. I just want you to know I'm coming from the POV of a long time CoNDOR/CaHOOT racer.

    So we've started to notice significant seed variance. A clear example of it is in game 2 of d_tea vs Spootybiscuit. (VOD is over at kailaria's channel:
    Temple of Storms, Gerudo Ruins, and Frozen Grotto were completely different between the two racers.

    The seed variance seems to be caused by changes in dungeon order. Here’s the dungeon sequence (and timestamps) for both players.
    # d_tea Spootybiscuit
    1 Temple of Storms 29m 52s Frozen Grotto 29m 41s
    2 Gerudo Ruins 31m 12s Temple of Storms 32m 22s
    3 Frozen Grotto 32m 52s Gerudo Ruins 34m 07s
    4 Lost Swamp 34m 57s Lost Swamp 36m 56s
    They completed the first three dungeons in a different order than their opponent.
    Those three dungeons are the ones with the variation in layouts, items, and enemies.
    Lost Swamp was identical for both racers. They both completed it fourth.

    I left out Kakariko Crypt because the dungeon order does not seem to affect it.
    d_tea enters @ 36m 51s; after completing all 4 instrument dungeons.
    spooty enters @ 16m 46s; before completing any instrument dungeons.
    Both players had the same layouts. The enemies were stronger for d_tea.

    I’ve done some additional testing with the seed. I suspect that the dungeon variance is related to the current ‘difficulty level’. I put the results of that testing in this google doc.

    I would like seed variance to decrease. I would like items, dungeon layouts, and big key rooms to be the same for both racers. Regardless of the order that they complete dungeons.

    One my of my concerns is that players/racers (specifically CoNDOR types) will be turned off by seed variance. Potentially dismissing CoH as a racing game. I think it's an awesome game and awesome to race. I want people to race it, and worried that this kind of variance will get in the way.

    There's more to dicuss in regards to variance. How you think it fits into games or into competitive environments. How it affects the viewing experience. If you enjoy racing it or not. What steps you can take as a racer to reduce/increase your variance. When you'd want to do that and why. That sort of stuff. I'm happy to have that discussed here as well.

    tldr; seed variance bad ...mkay? Pls fix

  • #2
    Entering and exiting a dungeon will also cause it to be re-randomized in an unseeded manner when you later return (likely only encountered if someone dies).