Select the hero whose stats you want to view, change to competitive, select “OW2 S09”, select rank you wish to view. Don’t use 3 month old data, and don’t use “All Tiers”. All tiers creates an average from each individual rank.
See this happen too often when people talk about Brig having the highest winrate across all ranks.
You can also click on their role beneath their name to view stats for the entire role for the season. I’m not sure why the main page doesn’t let you sort by seasons.
As I said to someone in a different thread:
What, pray tell, should we use instead?
Overbuff is all we have. We wouldn’t have to rely on it if Blizzard would just give us the data. Can’t go off anecdotes, because this community will happily parrot the same terrible takes from the likes of Flats and Samito as if their word is the gospel.
No matter how “correctly” you use it, it will always be wrong if 90% of the accounts are private profiles (read as smurfs who abuse cheap heroes) or haven’t been searched to even appear in there.
Unfortately although wr only have overbuff we have nothing. Like other posters say withouyt taking into account private profiles any data taken is basically guess work and incomplete
Yet we can see that the heroes who got buffed were, in fact, underperforming. Does that not indicate that it’s actually as close to accurate as we’re going to get?
When it suggests that the characters you like are overly powerful, create a forum post decrying Overbuff as a washed up stats site that’s about as accurate as a horoscope.
When it suggests that the characters you like are weak, create a forum post touting Overbuff as incontrovertible proof that your character needs a buff.
This isn’t accurate either cause it only functions on those who have profiles public, which most ow players have it off since they don’t go into settings EVER to make it public which is probably 99% of the player base.
It doesn’t need to have a vast majority to draw from as long as it’s got enough for a sample size. I imagine that it has.
But Overbuff also doesn’t need to be accurate down to the tenth decimal point. It gives us an idea of how things are going and like Kaitsja has said not far above my post here: certain heroes (Junk, Sym, Orisa) have recieved buffs who are also relatively naff according to Overbuff. It could be coincidence but I suspect not.
No. You still seem to not know how statistics works. Cherry picking would imply that you only factor in people you want. There is no factor of who has the profile open and who has it closed. I have seen open profiles from Bronze to Top 500.
If you stand on the street and ask 1000 people what they rather eat for breakfast: cornflakes or toast, you also dont cherry pick. Its called a sample and the bigger the sample the more accurate the data. The data on its own does not completely change tho.
As for the apple pie example: you can guess how many apples were used for the pie based on one, two or more slices. The more slices the more accurate. But even with only one or two slices you will get pretty close.
By taking only into account ppl with open profiles only, you are already doing that.
You are relying on selected individuals who have chosen to show their stats and also have selected to engage with overbuff search. Overbuff can’t collect data of accounts that haven’t been searched.
That depends entirely on what you are trying to look at. The different data sets have value when looking at different information. There is even trending data as well as average damage, accuracy, etc. They all have value as long as you are using them for the right questions.
Context matters.
I take it you have never studied statistics?
It is called a representative sample. And it is how basically all statistics are done.
And the worst part about people claiming it is not accurate because it doesn’t have all the statistics is the vast majority of people making that argument are just trying to discount it in favor of their own ‘experience.’ Which is MANY orders of magnitude less data. And more than that, is almost always based on gut feelings of what they kind of remember from their games (not to mention ripe with biases). In short, they are discounting good, though not perfect, data in favor of effectively no data.
Well that answers my above question. No, you have no understanding of statistics.
Cherry picking data is specifically removing data you don’t like, regardless of sample size.
That is not cherry picking. Seriously, you need to take some time and learn even the basics of how statistics work and are collected.
Ok so how many active ow players are there? I mean the ones who actually play the game seriously? you can’t count me since I dont comp but seriously does Overbuff give this information?
Except people always use it in the wrong context. Such as saying Brig has the highest winrate across all tiers, and therefore does not need buffs.
If we go off the default information provided by Overbuff, you’re factoring in both quickplay and competitive play. You’re also factoring in the last 3 months. You’re also factoring in an average created from the lowest tier to the highest tier. This is where the “Brig has the highest winrate across all tiers” statement stems from.
Using this exact example, one could reasonably believe that Symmetra is fine with a winrate of 54.82%, yet we know that Symmetra has been underperforming in this season.
No, because the only ones who know the actual active player count is Blizzard. Overbuff is as close to accurate information as we are going to get? Is it perfect? No, but it is information we otherwise would not have.