According to overbuff grandmaster games, Genji is doing just fine, yet he’s being buffed. McCree on the other hand is at the bottom, but he just got nerfed.
Are the overbuff numbers incorrect?
3 Likes
In the fact that they still? havent updated the defense/offense roles & since we have private profiles, the site is even more useless.
Blizzard has the real stats. Overbuff mostly takes minority/random numbers and puts them some place.
I kinda wish that blizzard/ow team had site of their own with real stats of how heroes are doing, etc.
24 Likes
Overbuff statistically is correct. The reason Genji got buffed is because he is deemed “unviable” without his Dragonblade. I’m still surprised Dragonblade didn’t get a compensation nerf however. Regardless, Genji was underwhelming without Nanoblade which is why they buffed his base kit.
2 Likes
Probably good for pickrates.
Probably broken for winrates.
4 Likes
Overbuff isn’t an infallible source of information. There are bugs that lead to number errors, the sample is biased, and it doesn’t include everyone. For some unfrosaken reason Mei tops the damage chart at 40k per game.
With that said it’s still accurate with a very small error of margin. Almost every stat is going to be near identical to what it is in reality. It’s miles better than relying on public perception, personal experience, or analogies. Just remember that numbers don’t always tell the full story.
12 Likes
The fact that they haven’t even removed the defence category makes the website very untrustworthy in my eyes. It’s clearly been neglected for a good while.
When did Blizzard remove the defence role? Back in 2018 I think? It’s such an outdated site.
Nobody should be trusting it.
What’s Overbuff exactly again?
What are the number Mason. I don’t know.
It’s using brand new stats scraped directly from profiles, so the category that heroes are in couldn’t be less relevant. They could be all labeled as turds and it wouldn’t matter.
The biggest problem is its sampling bias. Right now it can only update from non-private profiles. It also doesn’t update absurdly frequently - so a lot of the week-to-week stats are skewed if not many people have bothered to refresh their profile. For example, I just looked up my alt-account and the last time that it was refreshed was 5 days ago - and I play it more than my main account.
Additionally, if someone un-privates their profile then their stats could flood the system all at once and skew it slightly towards whatever heroes that they played. Doesn’t really matter if someone played 1 hour in gold, which has a lot of players and thus data, but it matters a lot if someone played 100 hours in GM, which has less players.
ngl that’s pretty mediocore, particularly for a hero that is so saturated with onetricks. He’s also not viable at all in OWL despite there being many people there who also used to nearly onetrick him, so I suspect that that is more of the reason behind the buff.
2 Likes
See I don’t think that this makes it biased. There’s no inherent reason why someone would or wouldn’t have a public profile. It’s not like there’s an incentive that only some people can get to have their profiles private; it’s just the default. And if we assume 1000 people at each rank have public accounts, at least for pickrate and winrate these numbers are accurate.
3 Likes
I’m sure that being more disposed to play certain heroes have would make you at least slightly more disposed to have a private profile. I know for a fact that community-disliked onetricks (Sym, Zen, Bastion, Sombra) would be much more inclined to have a private profile.
Perhaps in the grand scheme of things it doesn’t have too much of an impact, but the fact that a study is opt-in as opposed to truly random definitely deserves a mention in the context of statistics. It definitely matters more than the name of the category that the hero is in…
3 Likes
I would love to see that transparency too. They might even get some good feedback from the community on how to balance the game?
not accurate at all
people who main X heroes are more likely to have private profiles like moira, sym, hog etc, which deflates the heroes pickrates from where they would be, leading people to believe they’re underpowered
while people who play generalist heroes are more likely to have public profiles like genji, mccree, hanzo, inflating their pickrates on overbuff making people think they’re more powerful than they actually are
1 Like
Its not entirely accurate but its the best we got.
All of the accounts you see with private profiles aren’t part pf their data pool so i would say not super reliable, unless youre just tracking your own progress.
Another source of bias on Overbuff that people love to ignore is that it’s more accurate on higher ranks for the simple fact that it can only collect voluntary data. Many mid and low rank players either don’t know that Overbuff exist, or don’t see any reason to share their data in the pool.
That means, percentage-wise, we have a more accurate vision of what is happening in the high tiers than low tiers. Diamond and up are probably very accurate to the real data, while silver and bronze data are probably ok for most analysis, but have a higher deviation to consider.
Pool size is fine. Overbuff is still very popular. And I think it will remain popular enough to be relevant until Blizzard finally decides to give us a public API.
I mean that’s fair, but I think that just the fact that there’s no real influence one way or another for someone to have a private profile or not is random in and of itself. Remember, the majority of the community isn’t on the Overwatch Forums or on Reddit. They won’t have nearly as much pressure or incentive to hide their profile as say a one-trick Sym would recieve on the forums.
Maybe if you look at solely the population of the forums or Reddit, but we make up maybe 5% of the entire playerbase, and there’s a pretty good reason why this is the case. We’re kind of biased against each other.
There’s no hard evidence of this outside of these two platforms, where people don’t just casually flame other players for having 20 hours more on Symmetra than on Soldier.
2 Likes
An issue with Overbuff is many stats are done per win.
For example, the damage stat you see for yourself is the total damage you have ever done divided by the number of wins you have.
According to overbuff, we’ve 4 roles and symmetra is a support hero
So i don’t know actually
It should be roughly accurate but I would take it all with a giant grain of salt.
There may be bugs in the way they process the data, and since it’s only public data there may be a bias for certain types of people who make their profile public.