I’m talking about how the devs decide which heroes are balanced, which are over performing, and which are underperforming, so they can distribute buffs and nerfs accordingly.
If I’m not mistaken, they only use a hero’s unmirrored winrate in masters+, and the method they use to calculate winrates is flawed as well?
I think the focus on high ranks balance is the right choice, but what about the unmirrored winrates and winrate calculations?
They haven’t really gone into exactly how they are looking at it. Here’s the thing.
Does it truly matter? As long as they are consistent in their approach, that is going to create a “balance.”
People would still be disappointed if the method was different. People would still be disappointed if they did not even use winrate at all.
It does not substantially change anything. There will still be well designed characters such as Juno that are hated if people picked them often. Because nobody actually cares about balance. They just pretend to. That’s reality.
It is what it is tbh. People hate seeing some heroes way more than others and, unfortunately for people who like those heroes, community sentiment against them is really negative.
Mauga is pretty meh right now and I like playing him but he’s not really allowed to be good. Fair enough I suppose there’s a lot less people that like him compared to those that don’t. They don’t even micro buff him to crit through hazard block lol
On the other hand Rein consistently over performs statistically but he’s kept good because he’s popular
If it’s what the people think they want then it is what it
What about how winrates are calculated? I remember someone that posted on here used to talk about how they weren’t accurate or something, can’t remember for sure who it was though. Is that actually a thing or not?
Oh, that was GreyFalcon. And no, he was wrong. His whole thing was that win rates should be calculated separately on attack vs defense. And I suppose you could argue that on payload maps. But so many maps are symmetrical now. Most of our game modes don’t even distinguish between attack and defense (koth, flashpoint, push, clash). The only main game modes that do are (escort & hybrid). So, like I got his point but I think it’s mostly moot and even so not that statistically significant.
I think how they calculate them is just fine. The only thing I want is more transparency. I want to know the win rates and pick rates. I want to see the unmirrored win rates of all heroes not just the couple they throw into a dev blog.
And honestly, I want to know if maps have any significant bearing. As a personal anecdote, my win rate with Sym is ridiculous but that’s because I only play her on maps where I think she is good. And I don’t try to play her just any and every situation. I think she’s “niche” and if you are niche then those map win rates all the sudden matter a lot. Even generalist heroes will have stronger or weaker maps but specialist heroes might have huge swings.
That’s the thing. Clearly Kiriko and D.Va are also very popular. I don’t think anyone even doubts this. But they both have people who also hate them vocally. Even though I think those people are a minority they are extremely vocal of their opinions and thus they don’t get to be good either.
And yes, Mauga is hated but he’s also weird because he does do really well in pro play even when he flounders in the ranked meta. And honestly I think there are just some changes they could make to him that would fix what people find frustrating about him. They have made him a lot better over time. I don’t hate him nearly as much as I once did. But I still think they need to push into his gun play on squishies a bit more.
Also with Rein, it’s not just that Rein is popular. That is certainly part of it. And he is. But like I said before so is D.Va. It’s that Rein’s power curve makes him much much better in say Plat than at Masters. So, they look at him in Plat with a 56% win rate and have to decide if they can buff him further to get his Masters+ win rate in line or if that will cause a riot.
No. I did not think that a powerful Reinhardt was what I wanted. I absolutely knew without a shadow of a doubt that is what I wanted. I would take it a step further and argue that Reinhardt being viable/powerful is objectively good for the game.
Then they are wrong.
May I kindly remind, we have no evidence that ranks even carry meaning. There is little to suggest the game assigns skill labels and ranks with fidelity and integrity. More than 1:1 players:accounts immediately cancels standard ELO/Glinko/TrueSkill ladder validity, and this team has written intent and motives (via patented rigging ™) to manipulate players.
Suppose ranks did mean something, then ‘diamond’ would be even more correct. Mainly for the sake of normal pdf inflection point. Around Diamond, you have the most sensitive dSR/d%people, and can infer more surprisal (a stats term) per data point. You can do this at mid-silver as well, but Diamond is better sweetspot, since it gives people something to aspire to.
From there, you discount (maybe sub-exponential kernel) a window above/beyond that locus.
See above. ‘mid-diamond’ (whatever that means) is ‘math better’ if you’re going for rank-based data. But there are better, less subjective methods, such as APM and accuracy % by sequence complexity (game tree parsing difficulty, principal variation ratios, traversal rates) to discern risk/reward values for various interactions (combinatorially up from 1v1 all the way to 5v5).
These are just some of the quantitative tools/techniques we use to mine strategies and ‘balance’ difficulty when trying to code general game playing bots/AI in say, military applications.
Of course, this isn’t used in the games industry because games are for fun and profit, not for winning wars against adversaries.
On a side note, if you wanted to balance (in a math sense, ignoring fun) for the ‘highest ranks’, you wouldn’t even use pro humans as playtesters. You would be using the top AI against top AI. And the results would be quite difficult to understand and probably look like turtleish boring play that cheese a pawn mismatch at depth 72. Or the OW equivalent of some ult economy delta during overtime.
They have consistently hired some of the worst decision makers in the industry.
You really have this level of tin foil hat on? Like seriously? This is absurd my dude. If this is your starting position then why even bother reading the rest of what you have to say.
I just don’t understand how there are so MANY of them. Like, at least half of these accounts have to just be the same person, right???
What I said is true.
We have no evidence that rank labels imply skill. This is literally what the very act of ‘balancing’ is supposed to address → making sure the right people are at the right place on the ladder, for the right reasons, like risk/reward/difficulty/payoff. Meritocracy. But alas this DEI team balances for fun and popularity and profit, over skill expression.
By their very nature, balance adjustments are an admission that ranks are off. Which means the feedback loop is implicit, and very brittle.
The rest of what I said works on the condition that rank is somehow meaningful, and goes into detail about what types of metrics you would use beyond rank-based data. As balance around rank-data, representation and winrates, are antiquated concepts - especially if looking at GTO (game theoretic optimal).
Their balance team doesn’t care because they’re not math people.
It’s demonstrably not. Your ideas are not in any way based in reality. I have no idea how you even got so turned around.
You don’t even grasp what my ideas are. Show your math or show yourself out.
I do. And they have no basis in reality. It’s like looking at a chair and saying, “that’s not a chair. It’s clearly a sofa. Prove to me it’s a chair”.
I’m sorry but if you want to prove that ranks mean nothing when that’s obviously wrong, you have nothing to contribute and I’m wasting my time even with this response.
So heroes that are played a ton or by one-tricking automatically have a more balanced win rate.
Another thing I believe: Some heroes are more likely to be picked in a losing situation, lowering their winrate but not because of the hero power level.
What’s more interesting is how high people climb with the heroes because of the hero. You can partially see that when when you look at the leaderboard distributions. This is what feels intuitive. If someone is higher than they “should be” because of the hero choice then they’re “boosted” so the hero should be adjusted.
I’m not paid to put that idea into formulas but I’m sure there are ways.
they dont balance the game, they just change the meta to make people come back, nobody is that stupid to do such awull balance choices right, RIGHT? balance around what the league did was what killed overwatch