I still don’t get how they haven’t updated sum into damage role. It’s been like 2 year’s.
Forgive the non-relevant conversation, but is your tag (Baja) pronounced:
- or Ba-YA?
I’d challenge that maybe it’s a good thing because too many people let those numbers rule their fate.
It’s the placebo effect.
Maybe overwatch wouldn’t feel so narrow minded.
It’s pronounced Baja
You should change it to “Helpful.”
technically the first…but i prefer the 3rd
I mean you both kinda right here, overbuff doesn’t have a bias in the sense that it takes all of your data (though it’s in average per game not average per 10) but it also only takes data from public profiles and while that’s not really overbuff’s fault it still puts a fraction of bias into the mix because people can hide their data, but again, that’s the players choice.
As far as the stats go I don’t mind them, but one thing that we can’t really make tangible is game sense, if there was some way to quantify/make game sense tangible, I’d like it more, because yea stats show " I do this per game" but it doesn’t say what you are doing with comms, callouts and other stuff or making big plays that effect how things go in game, because, for example, you could have terrible healing stats as a healer (healing per game/ten) but if you are making good call outs or making big plays with say Ana’s bio made and hitting a lot of clutch sleeps that also tells a lot about a player but you can’t see that in stats…just my thoughts, there needs to be some way to quantify/tangify (yup, tangify) comms/big plays I think
Oh and the “endorsements” ain’t what I mean
First off, you need to have proof, objective proof, that certain personalities, playstyles, or just people will have a public profile over a private profile, to say that there’s a bias in the information collected. You can’t call Overbuff biased because of the claim that specific people will have a public profile, when you have no data backing up that claim.
There is no sampling bias without incentive. When it comes to having a public or a private profile, it doesn’t matter which you choose. You don’t get anything. We’re not paying people to have a public profile, there’s no advantage or disadvantage to having it, you either have it or you don’t.
Without incentive, the people that have public or private profiles is essentially random. You could attempt to make an argument that there aren’t enough public profiles for the stats to be accurate. You’d be wrong. Ignoring yourself, there are 11 people in a match. At least one of them is going to have a public profile. That’s 9% of the playerbase. Millions of accounts.
That’s more than enough to give a trend of what’s happening in the game, as we’ve seen for the last 4 years. Overbuff matches the meta perfectly. When Blizzard releases data (like Soldier and Genji’s pickrates a couple years ago), Overbuff confirms it.
I personally wouldn’t use it to compare personal stats to other players, partially because of private profiles hiding the majority of the playerbase to compare it to and partially because I don’t really trust Blizzard’s personal stat gathering to begin with, but when it comes to something as black and white as pickrate and winrate Overbuff is amazing. It doesn’t matter that it’s not every player, that’s not how statistics works.
And if you don’t want to use statistics, then don’t. Have fun discussing anything without having any data to back it up, because without stats and without Overbuff we’re just telling each other “nuh uh I’m right you’re wrong,” and that’s much worse than “Sombra must not be doing too hot on the ladder, she doesn’t get picked often and when she does she’s more likely to lose than she is to win.”
Ok admittedly there is a bit of an incentive but it’s not from blizz, it’s from dick group leaders who just kick you if your stuffs private and don’t give you a second to open it up
Is that incentive to open your profile though? Or leave it closed? Especially if you’re using LFG.
Oh know I’m just saying that most people are a closed profile and the like “f, that sh*t, he’s out” and give them the boot, at least in my experience
With regards to bias specifically…
Any time you ask for people to voluntarily submit data you are introducing bias…regardless of whether or not you were asking random people to begin with
You’re right…it’s very hard to concretely describe in this case…but that doesn’t mean it doesn’t exist…
Take The following for example (just for illustrations sake):
Because of the community’s general opinion of symmetra (partly due to thIngs like OB)…selecting her at times would be considered grounds for reporting…just based on that alone…the fact that you dared pick her…A simple inspection of a symmetra main might lead to toxicity directed at said player…so what does that type of thing do to the symmetra player in the future when given the option to hide said info? Would it be a stretch to say that a sym player, because of said treatment, might be more prone/likely to keeping their profile private?
Would it be wrong to say that that may then lead to underrepresented stats on OB for sym players in general?
Do you think smurf accounts are more likely to have profiles public or profile? what about alts? what about certain ranks (“i dont want people to know im bronze”)
It doesn’t have to be something grand…like I said in a way earlier reply…these things can be subtle…but they still exist…and when it comes to data interpretation …little things matter…And bias is horrible for study purposes…and a major concern when collecting information…
SO you have that issue (again REALLY hard to prove HOW it exists…but its existence is there nonethless)
There are other questions as well…like HOW much of the community has a private profile? i sure as hell dont know…do you? My (anecdotal) experience says that whenever i go to check my teams profiles maaaaaaaaybe 1 or 2 have a public profile (im usually lucky to get 2)…IF thats the case across the board thats a giant portion of the community youre not factoring in
and then theres the whole issue of how all this information is processed…and how its used…i think ive covered it enough in other replies to not go into super detail here but I will just say that i stand vehemently behind my assertion that OB is chronically misused and misinterpreted by this community…and THAT is quite BAD (and the chief reason for the post)…
that symmetra example mightve just been me trying to help visualize the whole bias thing…but that experience i was talking about from sym (and other) players is VERY REAL…and you can easily point a finger at the misuse that i keep referring to as a leading culprit
For clarifications sake…it IS in fact all we have…but it is poor information in its current form (it really doesnt say all that much even when used well)…and its not OB’s fault per say (they work with what they have as well)…the simple fact is that when it comes to whats on OBs…the only real answer we CAN give is “we dont really know whats on there at this point”…and that sucks with regards to a data source
BTW one would HOPE thats the case as most profiles were public at the time (id love to see it the posts myself…i too have looked for these things as i have said quite the opposite above…but i cant find any of these posts despite my efforts…theyre really old - possibly deleted if they were old forums - and frankly im starting to wonder if they were on the forums or on reddit now or maybe even an interview/video)…im all for private profiles (they exist for a reason) but for the purposes of OB we wouldve been better off if the change never happend
But that’s the thing, No one is asking people to voluntarily submit data. People are choosing to, of their own free will.
In any case, the bias that comes from asking people to submit data comes from whether or not the people think they have an advantage to do it or a disadvantage if they don’t. That isn’t the case with public or private profiles. People do or don’t because they want to, not because they’re asked, not because they’re forced, but just because.
Isn’t your argument saying that just because it might exist, we should throw out Overbuff all together? Even though you don’t have any proof that bias strong enough to alter data exists.
That’s the point. It doesn’t matter what I think. I’ve seen both. It matters what the data shows. And you have no data.
Bias of a handful of players isn’t going to do anything to the average, not when the average comes from literally millions of profiles. You need large, widespread, obvious to a fault examples of bias before you can throw out data.
My experience matches yours. 1-2 people (not including yourself) per team is 9-18% of the playerbase. If we seriously lowball the number of active and unique players to 10 million (20% of the 50 million accounts Blizzard claims) that’s still 900,000-1.8 million unique, public accounts. And while I have no data on this, I strongly doubt that the playerbase of Overwatch has fallen that low.
And when it comes down to it, 900,000 individuals is a statistician’s wet dream when it comes to a sample size. All advanced forms of statistics are about getting accurate data from a small sample, because from a large sample data is almost perfect. Objects you use every day to keep you healthy and safe weren’t tested on samples that large. Airbags for example had a sample size of 70 crashes on dummies and real people before OSHA decided that “hey, these are good enough for people to crash with.”
This was in January. Of this year.
We’ve had private profiles for at almost two years by then
When it comes down to it, you have no evidence that the data given by Overbuff is incorrect or biased in any tangible way. Just because it might be doesn’t mean it can be dismissed without any evidence.
In the mean time, the data given by Overbuff has been shown to match pickrates and winrates given by the devs, has matched the trend of metas, the rise and fall of characters, and the trend of their usage both before and after nerfs.
Then go play AI? Stats are what drive competitive gameplay. The urge to do better than someone else. While fps was not the first game to have stats that showed your performance it is one that is dependent on kills, healing, KDA, and assists.
If you don’t care about stats just play AI with friends since essentially you just want to play the game without caring about how well you do.
It’s a good source of stats and if someone doesn’t like it, it can just not look at it, I see no reason to get rid of it really.
Best would be for Blizzard to release an official site with similar stats that would just put overbuff out of business.
Another thing that has an effect tho is that because it only pulls from public profiles and while this has already been talked about, if they remove the private profile function from OW entirely thus allowing overbuff access to all data the data could become even more skewed then it already is because as other people have mentioned there are smurfs accounts, alt accounts, throw accounts, and stuff like that or some people who just don’t care and so that could give a false sense of “oh hey I’m doing good on x character or I’m terrible on y character” when there could be 50 people with say 5 alts each throwing or doing exceedingly well on a character making it look like more people play that hero and have either good or bad stats when in reality it is 1 or 2 people with alts/smurfs/throw accounts skewing the data for a character (I know 50 people is tiny comparitively speaking i just threw a number out there, obviously there’s way more)
I’ll just use Ana for example: If let’s say only 100 people play Ana but 5 of those people have 3 alts and they throw or just boost people with them it will skew the data for Ana because it’s say 15 accounts but really only 5 people throwing or boosting thus giving someone a false sense of “man, I suck at Ana” or “damn, I’m good at Ana” when they aren’t/are so OW needs to eliminate smurfs/alts as well honestly
its still not random…there are reasons why people go public/private or not and that very well may or may not have influence on whats collected…i cant quantify that for you because it would be hard as hell to actually do that…im willingly giving you and anyone else who has accused me of this that…its not like theres an application process asking everyone why they chose to pick one or the other…its just one of several “we dont knows” when it comes to that site and while not necessarily damning its still something to consider…
Just because it might be doesn’t mean it can be dismissed without any evidence.
just because it might not be doesnt mean we should ignore it either…
regardless its not the chief reason why this thread exists…i responded/elaborated because you as well as 2 or 3 others have openly questioned that biasis exist…
if you read further up (like way up there) my chief issue is how its used and how useful the information thats on there actually is…and i believe its actually done harm to the OW community over the years
Take your quote/link (which btw is not the ones i was referring to…i know for a fact that i was talking about much much older stuff)…while you point to it as an example of matching stats (and im sure it does though i havent gone back and compared), i point to it as an example of its chief problem: Jeff pointing out how people will take what they see on there and kinda do whatever they want with it as long as it fits a narrative…the site as is (again not their fault) is a poor representation of what happens in game and people are even worse at interpreting the little it does give us…
Im pretty sure i mentioned this above somewhere but IF OB actually gave us better data…which probably entails blizz having to be more forthcoming with its data - some actual transparency (ive cited the OWL stats page as a example of a “better” more useful page - seriously go look at it - its awesome)…this thread probably doesnt exist…in its current state though…ugh
They will NEVER do that, as it will show their bias even further
No, it’s not. The fact that we don’t know why someone chooses to make their profile public or private makes it random. We can’t attribute anything to causing someone to make their profile public or private. People may have their reasons, but those reasons aren’t going to be shared by everyone and that makes it random as a whole.
You can bully someone into making their profile private, and I’m not disputing that, but you can’t say that every person with a private profile has been bullied into doing so, causing a bias in the public profiles.
It does, because the chance of “might” is astronomically smaller than the chance of “might not.”
So why are we trying to get rid of the site when people are just bad at interpretting or even using data to support their arguments. Even if Blizzard released all data on every player, the same people would still be bad at interpreting it.
That article is the only time the devs commented specifically on Genji and Soldier’s pickrates and winrates in one post. There is nothing older that relates to those two heroes together.