News Realms Remove Ads - Go Premium! Sign In or Register

Bluetracker

Tracks Blizzard employees across various accounts.

As someone who is intimately familiar with the process of balancing large scale games, and how deeply connected perception and performance are, there are a couple radical ideas I've thought about, and I'm very curious how they would play out and whether any game has tried them.

As someone who is intimately familiar with the process of balancing large scale games, and how deeply connected perception and performance are, there are a couple radical ideas I've thought about, and I'm very curious how they would play out and whether any game has tried them.

  • Chadd Nervigg
    Posted 3 years, 8 months ago (Source)
    As someone who is intimately familiar with the process of balancing large scale games, and how deeply connected perception and performance are, there are a couple radical ideas I've thought about, and I'm very curious how they would play out and whether any game has tried them.

    • Chadd Nervigg
      Posted 3 years, 8 months ago (Source)
      Caveat: This is all just personal naval-gazing, no plans to implement anything like this in Hearthstone, though thinking about it is interesting.

      • Chadd Nervigg
        Posted 3 years, 8 months ago (Source)
        Context: Normal balancing is based on a combination of performance metric data, and player perception. Players view tuning as a developer choice, and regularly petition devs to buff/nerf things.

        • Chadd Nervigg
          Posted 3 years, 8 months ago (Source)
          RADICAL IDEA #1: DYNAMIC BALANCE Some automated process is connected to metrics on tunable stats, and updated very frequently (daily?). Example: Starting health for every hero in a MOBA is dynamically tweaked, daily, to maintain precisely 50% winrate of that hero.

          • CrunchTime4L
            Posted 3 years, 8 months ago (Source)
            @Celestalon I think that would be better in BGs. Not sure how you can balance standard cards automatically that aren’t minions.

            • Chadd Nervigg
              Posted 3 years, 8 months ago (Source)
              @CrunchTime4L Yeah, I think this sort of idea wouldn't work at all in collectible Hearthstone; the numbers involved aren't granular enough.

          • Chadd Nervigg
            Posted 3 years, 8 months ago (Source)
            RADICAL IDEA #2: PLAYER-DRIVEN BALANCE Some in-game method for players to vote on what to buff/nerf, on a very frequent (daily?) basis. Example: Every day in WoW you get a token that you can drop in a bunch of categories labeled "Buff Hunter" through "Nerf Mage" = +/- 0.1%.stats.

            • Chadd Nervigg
              Posted 3 years, 8 months ago (Source)
              Interestingly, I think both of these ideas have nearly no impact on actual balance directly, but massive impacts on the social systems surrounding the game. Petitioning the devs turns into petitioning other players. Playerbases turn more segregated and opposed to each other.

              • Chadd Nervigg
                Posted 3 years, 8 months ago (Source)
                It's really hard to predict social reactions to things. It's easy to predict that something like these would make the game more balanced, but hard to see what that social ramifications that would have. Do players just quit more when they know that it's their fault they lost?

                • Chadd Nervigg
                  Posted 3 years, 8 months ago (Source)
                  Or do players hate on each other for playing really well at something because it's "ruining the curve" for the median player? Does balancing for the average mean that high skill players get a very unbalanced experience that they hate?

                  • kilmarnok1285
                    Posted 3 years, 8 months ago (Source)
                    @Celestalon I don’t think the player base will ever be coordinated enough to actually tip the scales one way or another, hubris I know. I would think highly skilled players response would depend on how heavily you pushed on the scale to balance things for the median player.

                    • Chadd Nervigg
                      Posted 3 years, 8 months ago (Source)
                      @kilmarnok1285 There's no doubt in my mind that they would be coordinated enough. The viral nature of social networks means that viewpoints spread and echo chamber.

          • flame_eh
            Posted 3 years, 8 months ago (Source)
            @Celestalon Personally, as a fan of fighting games I hate this idea. It doesn't give the player enough time to evaluate if the issues they're experiencing against other players are caused by them, or the character being balanced. Will be too easy to reflect on every loss as the character.

            • Chadd Nervigg
              Posted 3 years, 8 months ago (Source)
              @flame_eh I think it's actually the opposite of this take: The character is always perfectly balanced, as a known fact. Which means you only have yourself to blame (and is probably a net downside).

          • holinka
            Posted 3 years, 8 months ago (Source)
            @Celestalon Sounds a bit like the dynamic weapon pricing CS did way back in the day. Fwiw, it was eventually removed primarily due to player opposition. https://t.co/v6VVWdqxbJ

            • Chadd Nervigg
              Posted 3 years, 8 months ago (Source)
              @holinka Yep, sounds very similar.

          • joegoroth
            Posted 3 years, 8 months ago (Source)
            @Celestalon Some armchair psychology on do we crave balance or do we crave change, and does the fact that something is hot or not serve as the catalyst for change. IE, there are no assigned seats in college, but everyone sits in the same chair each class anyway.

            • Chadd Nervigg
              Posted 3 years, 8 months ago (Source)
              @joegoroth Yes, exactly! These ideas are absolutely psychology questions, not gameplay questions. I think every hero having a perfect 50% winrate would actually be boring for a lot of players: I'm good at this hero, so have a 52% winrate. Why swap to something else that I'll be 40% on?

              • Chadd Nervigg
                Posted 3 years, 8 months ago (Source)
                @joegoroth Just so that I can get better at it eventually, and get up to... 52% again? No thanks, I'll just keep playing the same hero I'm already good on, until I get bored and quit.

                • joegoroth
                  Posted 3 years, 8 months ago (Source)
                  @Celestalon Also think it makes the barrier to entry as a live service game harder and harder, if you have 100 heroes and each one is S tier, the amount of shit you have to be prepared for is unfathomable. If there's a meta I can focus on preparing myself for a smaller sample size.

                  • Chadd Nervigg
                    Posted 3 years, 8 months ago (Source)
                    @joegoroth Yeah, that's absolutely true. There are hundreds of LoL champs, but you only really see a few dozen of them often enough to learn how to play as/against them. Seeing everything evenly would probably be overwhelming.

                    • Chadd Nervigg
                      Posted 3 years, 8 months ago (Source)
                      @joegoroth (I think you'd still see a meta develop though, but it'd be based on what's flashy/fun/easy.)

          • ChaosOS_59
            Posted 3 years, 8 months ago (Source)
            @Celestalon Bunker Build Time +- 5 seconds! Definitely only viable for games with large numbers that are sensitive to sub-10% adjustments, but I do think this could work. You'd still need developer intervention for desirable results imo. Also, which winrate stats? Global? Top tier?

            • Zinnin
              Posted 3 years, 8 months ago (Source)
              @ChaosOS_59 @Celestalon I think is the most interesting question when considering auto tuning, in many games perception and reality of tuning have variance through different skill ranks. If community is accepting of the auto tune system, may be able to have dynamic tuning for different skill brackets.

              • Chadd Nervigg
                Posted 3 years, 8 months ago (Source)
                @Zinnin @ChaosOS_59 Yeah, thoughtexperimenting out the idea, it does tend to lead me toward "Would we need to have different balance numbers for each ranking tier, game mode, whatever?" Then after that, do players rank themselves within a tier? "I like the Plat meta, I'm a god there. Hate Diamond."

                • Zinnin
                  Posted 3 years, 8 months ago (Source)
                  @Celestalon @ChaosOS_59 Yes that's a good point, tuning based on skill tier would also have weird knock on effects for smurfing, if a high skill design gets tuned up in low mmr because players are bad at it, smurfs may find smurphing behavior even more enticing, which could lead to player social issues.

                  • Chadd Nervigg
                    Posted 3 years, 8 months ago (Source)
                    @Zinnin @ChaosOS_59 Yuuup.

          • Muffinus
            Posted 3 years, 8 months ago (Source)
            @Celestalon Haha this sounds legit cool. Imagine a game that auto balance patched on the first of every month. AI picks from the lowest and highest performing heroes and makes one big change on each. Changes are based on what abilities are outsize but what the number change is, AI decides

            • Chadd Nervigg
              Posted 3 years, 8 months ago (Source)
              @Muffinus Yeah, I think monthly might actually be a better direction for this. It lets things feel seasonal, gives you time to develop skill at a certain meta, but also lets you know when it's worth coming back if you dislike a meta.

            • Muffinus
              Posted 3 years, 8 months ago (Source)
              @Celestalon Maybe players can vote on which AI suggested changes make it in? But the dev's never get to touch it. Enclosed system. Like a Moba terrarium.

              • Chadd Nervigg
                Posted 3 years, 8 months ago (Source)
                @Muffinus Actually, that gives me an idea. Maybe the right use for this IS as a side mode. Like, it's a special mode where the special thing is that it's a player controlled terrarium. "Here's a special WoW server where players control the balance. Have fun, enjoy the chaos."

                • Muffinus
                  Posted 3 years, 8 months ago (Source)
                  @Celestalon Twitch Writes The Patch Notes

                  • Chadd Nervigg
                    Posted 3 years, 8 months ago (Source)
                    @Muffinus Oh god. Yep, exactly.

        • ZoggHS
          Posted 3 years, 8 months ago (Source)
          @Celestalon How do you measure player perception without players telling you what to nerf tho?

          • Chadd Nervigg
            Posted 3 years, 8 months ago (Source)
            @ZoggHS That's a very common way that we do measure it, yes. (Along with reading all player feedback in general)

      • Chadd Nervigg
        Posted 3 years, 8 months ago (Source)
        (In case it's unclear, I think both of these ideas are deeply flawed. I just like thinking about how it would play out, what those flaws would be in practice, etc. From that, maybe there are bits of this that are reasonable for certain games. Maybe.)

        • DevSpacePrez
          Posted 3 years, 8 months ago (Source)
          @Celestalon There's a lot of subtle things kinda like that which I know have been done. Capcom made the difficulty dynamic in RE4 but didn't tell anybody. The zombies become more aggressive if you're landing most of your shots. Players don't notice but everybody has a good time.

          • DevSpacePrez
            Posted 3 years, 8 months ago (Source)
            @Celestalon The fun question is, would it have worked the same if this was advertised to the player?? Psychology is fun.

            • coil780
              Posted 3 years, 8 months ago (Source)
              @DevSpacePrez @Celestalon It's interesting because as these "tricks" become more well known, they also become accepted and even expected. Things like coyote time, enemies' first attacks always missing, magic bullet, etc. But you're right that advertising it would probably still backfire.

              • Chadd Nervigg
                Posted 3 years, 8 months ago (Source)
                @coil780 @DevSpacePrez Yeah, I think it was @32nds that first opened my eyes to this sort of secret magic. I love this stuff.

    • RDF1987
      Posted 3 years, 8 months ago (Source)
      @Celestalon I feel like the constant adjusting concept could work in a game like csgo. Too many people buying ak's? Up the price. Eventually and ideally guns would reach an equilibrium of their true value.

      • Chadd Nervigg
        Posted 3 years, 8 months ago (Source)
        @RDF1987 Funny you should mention that, because that exactly happened in CS, and most it was widely hated.
Tweet