Overbuff Overwatch

Overbuff Overwatch
Something went wrong, but don’t fret — let’s give it another shot.

World of Warcraft Arena World Championship
World of Warcraft Arena World Championship
Unfortunately, your browser is too old to work on this site . Please upgrade your browser to view rich content, log in and reply.
Pretty please don’t turn this thread into an argument about current stats
First things first, Overbuff is a site which walks over peoples public profiles, and gathers data from them, which it aggregates, and displays.
When forumers talk about overbuff, they USUALLY mean the heroes screen on it.
Useful points about the data it gets.
At the top of the dashboard, you can select over how long you want to aggregate the data. Which platform it is for, if it is QP or comp, and a role filter (which you should ignore as it very out of date)
Under that, there is which rank you are filtering for. All of them, or some particular rank.
Under it, is the tabs - which people tend to not notice. But there is 3 more tabs than overview - 2 of which is useful. (Medals isn’t)
Pick rate - It is what it says on the box, how much a hero is being picked.
Win rate - How likely a hero will win in a game, subject to a large number of caveats.
The last point is a VERY interesting one, since, it means that their win rate is heavy biased towards “how easy is a hero to get value from for people who play more than one hero ” - they are represented in the win rate WAY WAY more than the one tricks. The more someone one tricks, the less they are represented in the win rate.
An Example of biases in the data on this screen
This gives things like Elims per game, etc. objective kills, whatever.
But like, the tab… just isn’t hugely useful. The Eliminations Tab what you are looking for.
Ok, first things first, what is an Elim? If you do damage to a target, and they die within a few seconds after that, you are credited with an Elim.
E:D - Mostly a measurement of spam - an abjectly AWFUL measure of power.
Like, don’t use this in any kind of argument, it is just a bad measurement of… well everything.
Elims / G similarly useless by it’s self - it measures spam more than anything. But, between this and E:D you can measure deaths per game, which is interesting.
Solo Kills - More useful than it first looks. Multiply it by Elims/G to get solo kills by game.
Final Blows - This is the king of stats on this page. It measures how deadly a hero is.
Ok, that should be a pretty good primer on overbuff.
It is a pretty good data source, has more than enough sample size (don’t make me have the sample size talk to you all again).
Just remember it’s limitations, and everyone should be all good.
Actually I don’t think this is true. If I recall correctly, Hog made it to 18% pick rate back in mid 2020
They should sticky this, maybe people would then start looking up stats before crying “This hero is OP nerf it now!” threads
Nah who am I kidding it wouldn’t change anything if they did
Actually I don’t think this is true. If I recall correctly, Hog made it to 18% pick rate back in mid 2020
It is 100/6 Like, you can’t get over this, unless you get more than one hog on a team.
I wonder if Echo could push it up though, I don’t know how her ult is allocated towards pick rate.
Actually I don’t think this is true. If I recall correctly, Hog made it to 18% pick rate back in mid 2020
That was caused by people swapping who played hog within the same game I believe
So say tank 1 played hog first half, then dva and ball for the other half
Then tank 2 played rein and zarya first half and only hog for the next half
Then Overbuff counted those two players as having played Roadhog within the same game
Regardless, anything above 13% is ridiculously high, I’d argue a healthy meta sees no character over 10%
Then Overbuff counted those two players as having played Roadhog within the same game
It is scaled by time played in the game. If you are hog for 1/2 a game, then you have hog gets only 1/2 a game allocated towards pick rate.
It isn’t touched the hero, it is weighed by how much they are played. Blizzards stats DO get messed up sometimes though (See Mei’s damage/g) so, it wouldn’t surprise me if they messed something up.
I wish everyone would read this and understand.
It is scaled by time played in the game. If you are hog for 1/2 a game, then you have hog gets only 1/2 a game allocated towards pick rate.
I thought Overbuff only used the “games played” stat for determining pickrate, in which case it’s awarded to the hero you spent the most time on in the match, which is why you’ll have different win percentages if you compare Overbuff to Overwatch’s built in stat screen, but I haven’t really looked into it so you’re probably right
Well, a lot of it is from talking to other people. It isn’t all my own work as far as working out different effects.
Like, I hadn’t put together all of the crazyness which was win rates by myself, there is a lot which influences it, and some I didn’t get into (because, hey, small effects, and complex to explain).
Mostly, I wanted people to understand that, there is context, and the stats do get pushed around by a bunch of effects. It doesn’t make them useless, far from it, but it does mean you how to keep them in mind.
I thought Overbuff only used the “games played” stat for determining pickrate, in which case it’s awarded to the hero you spent the most time on in the match, which is why you’ll have different win percentages if you compare Overbuff to Overwatch’s built in stat screen, but I haven’t really looked into it so you’re probably right
Overbuff is… well, a bit opaque in how it generates it’s stats. I’m working off when I ran my own scraper, and what processing I had to do to make it look like overbuffs data, but, again, we were sampling different populations, so… it is a bit of guesswork on my part.
For my sins, I am trying to have a writeup on all of the stats in the upcoming OWL viewer which will make the processing on how they are generated painfully clear so people will be replicate the work.
You can still see a bit of data from the Super Hog days. Here’s a pic of him reaching 20% in GM.
Super Hog’s also interesting for the line that regularly gets put about: “A character that’s picked every game will have a 50% win rate”. You’ll notice that Hog has times where he’s >15% and also >50% win rate.
Huh, that is just weird. I may have to change my post
Super Hog’s also interesting for the line that regularly gets put about: “A character that’s picked every game will have a 50% win rate”. You’ll notice that Hog has times where he’s >15% and also >50% win rate.
Yeah he is being picked in GM, there but, like, the GM’s are playing a lot of NOT GMs so, you would expect him to have average GM win rate, NOT 50% win rate. (they are different! :))
But a 20% pick rate looks like broken Blizzard stats.
It is a pretty good data source, has more than enough sample size (don’t make me have the sample size talk to you all again).
Amen! Outside of glitches and bugs, Overbuff is an excellent data source, and every single time we’ve gotten concrete statistics from Blizzard they have nearly perfectly tracked what Overbuff said.
has more than enough sample size (don’t make me have the sample size talk to you all again).
I think too many people forget the difference between sample statistics and population statistics.
It is basically impossible to properly gather the entire set of stats from a population. Blizzard probably can with the player data, but outside of them releasing that data we never will. So, we rely on these sample stats.
Overbuff is the sample data of people with public profiles. Is it indicative of 100% of the playerbase? Hell no, a lot of people have private profiles and many of them are one tricks who skew play rates of some heroes, even if only slightly.
Its hard to know the best sample size for OW since the population size is not public data to my knowledge. But Overbuff almost certainly gets data from a minimum of 6 digit figures (100,000 is by far the lowest estimate I could make, this is super conservative btw). That is easily a big enough sample size to make broad claims about patterns in the community even without the full data set.
I think more often than not, people pull out the ‘Overbuff isnt the whole population!’ as a last resort because they dont wanna admit their argument is wrong (we’re all guilty of abusing stats, dont feel bad). Im not saying this is the only reason people make this argument, but it seems to be the most common reason. For example, Ana used to have spent a huge amount of time as the most picked hero period (not just support), but when people brought this up to suggest Ana nerfs it was shot down because OB wasnt reliable enough.
TLDR: People who say Overuff isnt reliable because not everyone has a public profile dont understand that all statistics use only samples of the whole pop
It can’t see private accounts, but, there is no reason to believe that private accounts are biased in a particular direction.
Players who set their profile to public tend to be proud of their stats or have stats that are viewed as being acceptable by the community or they would face prejudice if they had it on public
You say that, but there is a massive amount of people with public accounts at all ranks.
It VERY much will slant it towards the higher ranks, but, since the stats are broken up by rank anyway, and the samples sizes are big enough… I don’t think it is a thing we have to worry about.
It WILL mean lower ranks may have less people public, but, that is a sample size thing rather than a stats thing.
now… SMURFS may not run with public profiles, and they may not be counted as much, but that is a good thing for balance discussions.
Although OB’s raw data is useful I find the more applied information on Overpicker more useful. I just put in the rank, map, our team comp and (ideally, if I know it) the enemy team comp and it tells me who I should play.
Although in practice the answer is always Rein unless the other tank is playing Rein already, then I should play Zarya. (Also there’s the 1% chance the enemy team comp is one that recommends I pick a different MT).
It’s sad that you post this in a “how to use overbuff” thread since you can’t know if these are true.
These two bullet points are not facts necessarily, and as such you don’t know if your sample is truly random so I take all Overbuff stats with a huge grain of salt.
You just say “there is no reason to believe private accounts are biased in a particular direction”
Yes, because resampling with too smaller sample size would give the data a lot of jitter, which it doesn’t have.
“there is no reason to believe private accounts are biased in a particular direction”
There isn’t, if you think there IS reason to believe that private accounts are biased in a certain direction, now is the time to explain why, and in which direction.
Because currently, there is no reason to think that the unsampled data is materally different from the sampled data.
You think a post explaining the biases of each stat in great detail is made by a person who doesn’t understand data science?
LITERALLY it is a post about biases in stats.
Powered by Discourse , best viewed with JavaScript enabled

No tanks, its fine. Don't protect me.
Subreddit for all things Overwatch™, the team-based shooter from Blizzard Entertainment.
Let’s appreciate how this beta Overwatch characters only...
Even after 4 years, still feels so good to get this message
Illustrations Of Some Of My Favourite Heroes
Rule #1 when attacking on Rialto: watch if the payload is...
So..... I made the fox girl "leak"/ fanart :D
Latest tweet from official PlayOverwatch Twitter is just...
their Junkrat Set me up perfectly! I would hate to be him!!
Everything confirmed for Overwatch 2
the strangest endorsement level i have ever seen
Anyone have any guess what this could be? I have none
Does this mean the all stars vault too?
This Mei/Rein bowling combo was nasty AF
Why is the mercy skin still available? It should have...
What Meka squad member should be / has the most potential...
Reddit Inc © 2022. All rights reserved
So I recently started playing overwatch again so I don't really care too much about my actual stats but I don't understand how they work, with the quick rank being a percentage is it the lower the better or the higher the better?
Also how accurate is overbuff can I trust it to keep up to date with how I am doing or is it better just to use the in-game stats and ignore my rank compared to others?
Our ranking system is talked about more on our About page. We don't share specifics, but we do share the four factors that go into our ranks. It's a combination of SR, your skill vs others on Overbuff on that hero, win percentage, and games played.
We use a percentile system to display your position in the rankings until the top 1000 or 2. You can always mouseover that number to see your exact rank.
The notion that data skews high is made up by Reddit and not true. Players from all skill levels check out their stats and if there is much of a skew, it's marginal. We don't have every player, but we have a large sample.
As for how accurate, we refresh the data provided by Blizzard whenever the profile is visited, and for active profiles, about once daily. All data is provided by Blizzard.
Question, on percentiles do you separate xbox, PS4, and PC from one another or is everyone clumped together?
It's been a while since I used Overbuff, but if I remember correctly, they use percentiles for everything - and a higher percentile is better. In terms of quickrank, you'll work your way up towards 100 until you get an actual numerical rank.
One caveat is that anyone with a private profile is no longer included in the Overbuff rankings, so the stats are now heavily skewed towards higher-level players and/or those who take Competitive mode super-seriously, and not representative of the player base as a whole.
I stopped using stat sites / apps ling before the private profile thing came into effect, because it's very easy to obsess over certain things ("My stats are amazing, why aren't I ranking up?") when in fact the stats by themselves tell you nothing about how a player's decision-making process and judgement, which is what really matters.
I see, I was wondering why I was only in the 4th percentile for roadhog when most of my hours are sunk into him and my friends who do play seriously say I'm a valuable asset when playing even though I'm only in gold, but then again they say my value is in my hook accuracy and being able to shut down enemy supports and tanks easily making most team fights a 6V5 😂

Not connected.
Connecting in {{retryTimeLeft }}s...
Try now

Developers

Getting started
Documentation
App creation process
Apply for funding

Gamers
Appstore
About
Careers
Blog
Privacy

Support

Help center
Get help on Twitter
Developer's Q&A

Browse by game

League of Legends

Hearthstone

Fortnite

CS:GO

PUBG

DOTA2

View all...

All apps

Guides & Trainers

Enemy stats

Fun stuff

Capture & Stream

Communication

Music & Video

Social

Utilities

Beta apps

All Mods Related

Mods Managers

Server Hosting

All skins

Featured

By the community

Creators

Build an App
Publish a mod
Why Overwolf
App documentation
Mod documentation
Creator services
Apply for funding

Gamers

Discover Mods
Discover Apps
Gameplay First
Download Overwolf
Suggest a feature
Overwolf Alphas

Company

About us
Our story
Our culture
Careers
Blog
Brand Guidelines

Work With Us

Partners
Advertisers
Influencers
CurseForge Core

Support & Privacy

Help Center
Supported Games
Terms of service
Privacy Policy
Licenses

Overlay settings
Manage subscriptions

Remove dock shortcut
Add dock shortcut

Remove desktop shortcut
Add desktop shortcut

Uninstall

No reviews yet... Secure your spot in heaven, be the first to rate this app! (you must first install it)

Rate this app | To rate this app please log in

No reviews yet... Secure your spot in heaven, be the first to rate this app! (you must first install it)
From now on, here's how it's going to roll:

As gamers ourselves—we want the best for you.

Your review is about to be deleted. Are you sure?

Naked Boy Ass
Junior Nudists Nude
Naked Porn Girls Video

Overbuff Overwatch

Report Page