The website still thinks Symmetra is a Support for gods sake. It still thinks Defense is a role in the game. On top of this, more profiles are privated than not which naturally messes with the data that the website automatically collects.
Stop using it as a source for info on heroes. It is objectively inaccurate and you shouldn’t trust it.
Its okay, it just shows a rough lay out of what is really happening.
No it’s actually pretty accurate.
Sorry.
its accurate. they just didnt update the roles. why do you think it isnt accurate just because of the roles not being changed? it still shows their pickrate/winrate
Give us your alternative source for stats for ladder play then.
That only matters if people who hide their profiles do things drastically different than those who don’t. Every stat is from a sample, not the entire population. Also overbuff while not perfectly accurate, is the best we have. Or we can just go by what people say, and every hero is the best or worst depending on who is talking.
IF ONLY overbuff could do this…
Imagine if Blizzard WOULD do something like this.
If it’s not a credible source, then what’s yours?
Since you clearly know what’s credible and what isn’t
Private profiles aren’t that big a deal for it.
Imagine this. You have 10000 players. Over the first 9000 players you find a theme “All the top scores are by this one character.” then I say “But that 1000 you don’t see means it’s invalid”
Once you have a large sample size, you can reasonably assume private profiles follow the same patterns fairly closely.
So yes, while the roles aren’t correct, the pick and win rates tend to be pretty accurate. The real issue is that people want to use weekly and GM, which DOES have a small playerbase/sample size. Using monthly time for the most recent patches AND a long enough time for the players to have a good sample size. That’s the way to go.
What should we use for stats then?
the true issue with overbuff isnt data reliability (which doesnt mean that isnt the case)
its that whats actually shown on there is pretty useless as far as the game
global pick and win rates are terrible metrics…we make mountains out of mole hills with the win rates…theyre usually not more than like 5% apart…and the pick rates mean next to nothing other than people think theyre good
now if Overbuff could somehow do the stuff in the link i put up earlier…THEN WE’RE TALKING!!!
blizz get on that plz
I wouldn’t say it’s horrible but I also wouldn’t say it’s pretty accurate either.
Private profiles will heavily skew the results.
Also the fact that the Overbuff maintainers can’t even correct hero roles throws off results as well. The site gets massive traffic and the maintainers are just leaching off the ad revenue while not fixing the site.
I’d say T500 leaderboard is more accurate for GM; overbuff is better for all ranks
No, we shouldn’t balance around GM-Top 500 tho
Well I’m not saying you have to balance around T500, but looking at leaderboards is just in general way better than looking at overbuff gm specifically
I think we should use both to get stats, better than either or
But like… why haven’t they updated them? Imagine trying to show someone a scholarly research paper but the format is off and everything is mislabeled. Your data and numbers can be correct but that paper is going straight into the trash because consistent unfixed errors puts all the content into question.
“This is why X hero is OP/Weak and here are my sources. They are 100% accurate trust them”
“Why is Sym a support on here?”
“Oh yeah just ignore that blatantly inaccurate part, it’s irrelevant. No sketchy implications there.”
Its more people have to make smart judgements on what makes sense. Yes, the site is old and arguably outdated and being poorly maintained for new changes without much explanation for how stuff works, but its still a valuable source of information that is not as broken as one might think as many stats still reflect what you can find form your own profiles and top 500 rankings and other players.
We simply have to understand where its broken and where its not. The all heroes picked is most reasonably not broken as it has accurately reflected top 500 character selections compared to GM for a super long time. It is not know though if that is Asia, Americas, Europe, etc. only regions or all. Season 27 is odd in that Asia has a HUGE variation from Americas/Europe in tank picks.
It is however broken in using the hero pick rate for doing tabs like offense, defense, and support as it inaccurately still clumps the wrong units together. You can still compared stats form top 500 from recent picks to say the 90-percentile and higher for characters in Comp/QP and get a feel for it representing well.
The only reason to continue using Overbuff is because its established with a large amount of continual usage and its generally still better than people using anecdotal evidence.
This is ideally what you should be using overbuff for. Not as your entire argument, but just a backup.
I have to admit, I’ve been worrying a bit too much about what overbuff says. It’s very tempting to only look at numbers and not WHY those numbers are a thing.