Blizzard can very easily read stats themselves, saying that “x hero is fine, they have x% winrate/pickrate” isn’t bringing anything to the discussion, especially since their stats are actually more accurate.
The only reason we should be arguing about stats is discussing the variables that effect the stats. And how valid you would consider these stats to be. I.e what variables effect WR and PR, assuming those are the main stats they use at blizzard (considering what they have said in the past, those seem to be the main stats they use). For example, I remember a while ago a dev said “Genji and soldier are not as bad as everyone makes them about to be, they have x% PR and x% WR”.
Imo I think it’s actually better we just talk about the heroes themselves, and the validity of the stats, rather than stats themselves.
As I’ve said before. When it come to balancing and what is OP or underpowered, Its different for everyone.
But the issue is that most peoples balancing comes from “feelings” rather than logic.
“So and so FEELS un-fun to die to”
“So and so is Un-Fun”
“So and SO FEELS so oppressive”
“Im so TIERED of always seeing so and so in my games”
Im so fed up with this meta"
“Why cant these unpopular heroes be meta or popular”
“Its so boring playing against so and so heroes”
Ect, etc.
Basically it always comes down to feelings and each person wanting the OW balance to cater to their favorite heroes and the way they want OW to play like.
And whats funny about this is that Blizzard actually has accurate internal stats. Not the inaccurate mess that is Overbuff.
Which was perfectly shown in the last balance patch everyone was crying about.
They didn’t tweak those heroes by mistake. It was because of internal stats they have,
They know what heroes are actually underperforming and need help.
Just goes to show most people in these forums have no clue of what they are talking about or rely on faulty metrics to base their opinions on. Mostly feelings and popularity.
So as a big stat junkie I find these dev takes rather offensive/insulting since they never specify what numbers they’re exactly talking about and it’s fairly obvious that a lot of these takes are just some minor “PR stunts” where they don’t represent the accurate data (being highest tiered gameplay or other areas of gameplay where the realistic potential of heroes is game changing because of how important it is as a pre-requisite for victory) and are mostly catered towards an audience who can’t really tell apart one from the other. It’s like pointing out at Torb/Sym before they got their reworks and whatever and saying that “oh they were great heroes, some of the highest winrates in the history of this game” when almost all of those victories were racked from <2500SR games where their nearly automatic abilities had an insane amount of value by default in comparison to what they could perform on other heroes
The context is king and as we saw on the Pharah and Zarya topics when they had changes coming at them the validity of the data is definitely something you have to consider a lot when looking at things. Sometimes heroes might even have trendy winrates because if a meta is very strict for a long time (for example only 8 meta heroes or you get a defeat) and suddenly players need to relearn how to play other heroes again which might be counters to something that’s suddenly dominating
What I’d wish is that the devs quit acting like the playerbase doesn’t know what’s good in the game and what’s not and just hit us with roster wide patches. They said they wanted to avoid wide number crunches, but I think that that is exactly what OW needs and if they’ve done that since day 1 that we’d have a lot more lively game. After all massive imbalances between heroes and the devs reading the data backwards sideways upside down and making patches constantly more based on feelings for design than overall objective balance, is what caused most of the issues that ended up killing the game
Using Overbuff to argue is obviously wrong and redundant as you already said, but i do think feeling and real experience is quite important to voice.
Bliz had access to statistics and data, so they dont need us to quote Overbuff, what they need is each of our experience within our games, where they cant quantify it themselves.
A lot of these feeling are often blown out of propotion of course, but they can easily be dialed down and truth can be read between the lines.
i dont think its bad at all to discuss stats, i think its bad to use stats incorrectly and act like its fact. like the other day i was arguing about how nobody plays doomfist, someone said he has the highest winrate hes ever had in gm… hes picked by 1 percent in gm, the kids that have played doom since forever and never stopped, ofc hes gonna have a decent win rate there… these people dont understand that 1 percent is such a tiny number played by a few insane doom players, they cant fathom that that would lead to such a win rate… when on average doom is horrible to play in diamond thru masters-gm when ur playing into cree hog brig etc. every game and dont have the mental capacity to put all ur energy into grinding one of the hardest heroes in the game into your hard counters every single game because its overwatch in 2021, the same maps, same doom nerf for 2 years, same cree being buffed every few months. people do not know how to read statistics and they cite overbuff in every post about balance. its horrible misinformation that i think the devs actually listen to lol
I know, reading forum post will often boil your blood. But it s probably the same thing for the devs as well.
They probably also just laugh at bad takes and other perceived balance issues just as much as we do. They just read them to keep up with the complaining trend.
it’s much better to talking about stats alongside contextual knowledge of heroes within the game.
e.g. lets say you want to evaluate whether a niche hero is balanced or not esp when comparing to generalists or comparing to heroes of other niches. only talking about it in terms of contextual game knowledge gives you 0 baring as to the key question of “is this niche hero proportionally stronger in their niche for how narrow it is to justify the niche design?” esp when wanting to have every hero be approx even in terms of overall “power level”.
that’s where the stats come in to actually let you see “how much”.
Hero Prob of winning overall (agnostic of situation) = prob of niche occurring x prob of winning within niche + prob of out-of-niche occurring x prob of winning outside of niche.
ideally we want Hero Prob of winning overall to be the same for every hero and be at 50% (fair chance to win or lose).
using this alongside taking GM PR as an estimate of “prob of niche occurring” and the associated WR as an estimate of “prob of winning within niche”, one can estimate “prob of winning outside of niche” and compare that with contextual game knowledge for the hero to judge whether if such a niche is plausibly balanced or not.
for example:
^this month’s Gm stats used there and PR was normalised dps PR (i.e. of the dps heroes only, how likely was sym picked; calculated by hero PR/ (sum of all PR of heroes within same role))
I’m the opposite, i woukd rather talking to players who discuss through experience. We know where and how and why such personal feeling comes to pass.
People who can only speak through stats, on the other hands, is not to be trusted. They lack strength to play and form opinion on their own, instead they quote other’s achievements to make arguement. It s just so wrong. Players who dont go through the journey themselves have no right to speak of the destination, like i cant say Tracer is broken while myself still die to my own pulse bomb from time to time, that s just very wrong of me.
People who are factually wrong and try to imply that first-hand experience is a pre-requisite for criticism (which it’s not) are not the one’s to be trusted. You see this self-serving type of mentality is what’s the most harmful to the game. Every single decision they’ve made ever towards balance or any other aspect of the game for that matter based on similar scenarios: has been very awful to the game.
One of the more notorious examples is that 1,5 months during which they buffed Genji out the wazoo because they were in active inside discord discussions with Genji players like Shadder and Samito. So internally they decided to break the hero for a while because they knew they’d get a few sizeable content creators to whip out content again and to see what happens to the game. Well it turns out that: Everyone else hated it and we lost a massive chunk of playerbase from that time
You see it’s not that we should use objective statistics and data to dictate what happens to the balance of the game. It’s an outright pre-requisite that all the heroes are to be balanced around their realistic potential in the highest tiered gameplay for the health of the game. It’s the only way to have a large variety of heroes be actually playable and to not have the entire playerbase implode on hero elitism between those who want to actually perform well and those who want to soft-throw on their “niche” heroes, who are not even that when the balance is skewed
In short: There’s nothing to gain from subjective balancing and everyone just loses from it in the end. That’s why they’ve moved away from it veeeryyy slooowly over the years
I’ve said Goats , double shield, all that is boring and lifeless. And most that share that sentiment have already left OW completely.
But like I said people are complaining that the devs dont put out balance patches, when they just did, just that people in here complained about it.
So what that tells me is that not only are these people demanding frequent balance patches but balance patches that reflect their own bias view on how the game should be patched.
Only then would they accept it.
Im sure if the next balance patch included the following:
Buffs Sym to have auto lock again and her beam now has a 50 meter range. And also she has Shield Gen again.
They Make it so Bastion has a shield in turret form and also could headshot.
Mercy had 80 healing per second and 5 man rez is back.
Moiras damage is increased by 200%.
Brig is now back to her release form.
Hitscan heroes now all do 50% less damage and automatically start losing health if they are on the field for more than 3 minutes without dying.
All shields in the game have double the HP points now.
People in these forums would be like…
“BEST EXC CARD EVER! OMG YOU OUT DID YOURSELVES BLIZZARD!”
Experience is all well in good for context, but that doesnt cover bias. In my opinion Doomfist is incredibly irritating, and annoying to fight against. Im however not going to whine on the forums that he needs to be nerfed. However a lot of people here do because their experience and feelings are “this character sucks to fight nerf them.”
I think it’s fairly naive to say “every time Blizzard buffs something, it’s because their stats show that thing is struggling.”
Tbh, I call bull on that, especially with the Cassidy buffs in last patch. It’s doubtful that Cassidy was struggling anywhere near, say, Reaper or Doomfist.
It’s also doubtful that Tracer ISN’T overperforming in Blizzard’s stats, since she’s used literally everywhere, and has been for ages, at high ranks. But still no nerfs.
There’s a lot of reasons to think that Blizzard isn’t using a stat-based approach when dealing with balance.
Those same stats you are talking about? Lined up exactly with overbuff at the time. Another thing is blizzard isn’t transparent at all with anything. They hide stats cause that means they got to openly acknowledge how bad many heroes in this game are. And if the last year has shown anything it is how bad blizzard is at accepting responsibility and criticism from anyone, or actually listening to player feedback.
You really shouldn’t be advocating for anything in their favor until they are -fully transparent- and -fully responsive- to the community like some other game developers are.
Okay, let’s talk about what each stat really says.
Win Rate:
Most people are confused about what win rate actually tells you and what it doesn’t. Win rate is not an indicator of a hero’s power. Outside of the absolute top and absolute bottom the matchmaker will push win rate towards 50/50. You win more than 50% and you go up until you don’t. You win less than 50% you go down until you don’t.
What does it tell us? Deviations tell us what happens when people who do not play the hero regularly use it. This typically comes in two scenarios.
The first is a high play hero has a notably lower than 50% win rate. This makes sense as regular players get pushed towards 50% while irregular players are not going to be able to play up to their rank because of lack of practice. So, it tells us that despite being less practice a large portion of the player see an advantage in picking the hero. This would indicate the hero is an effective counter to some other hero(s) or strategy(s).
The second scenario tends to be really low played heroes with high win rates. This tells us that by and large the hero is not effective, but that in very specific situations they are really good. In other words, it is an indicator that the hero is really niche.
There is also a large problem that comes up. As a hero play gets close to the point where they are on both teams every game it starts to ruin the value of even looking at win rates. The hero on both teams means a win and loss to the hero, canceling each other out. The more games that fit that description the less games actually effect that stat. This means your effective sample size gets smaller as the play percent goes up. Smaller sample sizes are more prone to large swings and are less accurate, so less informative. So, the higher the play percent, the less information can be gleaned from the win/lose percent.
Pick Rate:
Simply put, Pick Rate is the best indicator of a hero’s power level. People tend to pick what works and gets them wins. For the most part the fun of winning overrides the play preferences people would otherwise have. We have seen the direct correlation time and time again. Every time a hero is buffed their play goes up. The bigger the buff, the more it goes up. Every time a hero is nerfed, the play goes down. The bigger the nerf, the bigger the drop. While there will always be some that buck the trend and play poorly performing heroes regardless of power, the percentage of players they represent is negligible compared to the percent that play based on power. So, while it is not a 100% perfect measure, it is a really good indicator of how strong a hero is.
Yes. But what is important to tease out is why those feelings are there. I get that some of the feedback, well a lot, will be heavily impacted by biases. But there is often problems that can be identified. For example:
This kind of feedback is often tied to a hero being played far more than others, leading to them in most games. And that is a balance issue. Sure, some of it is just noise from players who have an agenda. But often there is a real reason behind that feedback.
Or…
This can be a really good indicator of other problems. If, for example, most of the kills from X hero can’t really be stopped no matter what you do it can feel really bad. Maybe a hero is just giving to much value for the effort and risk put in. Or maybe it is a case where a certain hero is just way to good against the heroes it counters. Those kind of cases may be an indicator of problems that do need addressed. And they are things you cannot get from simple number crunches. The feedback can sometimes be a mess, but it is valuable.
There are some clear problems with this argument. First, using stats does not equate to lack of skill or understanding. Stats are objective, so people will often use them as a way of getting away from a purely emotional discussion. Someone who backs up what they are saying with evidence is far more trustworthy than someone that just tells you they know better. Second, stats are not really ‘other’s achievements.’ Stats are a reflection of what is happening outside a single person’s experience. To ignore them is very, very wrong.
If you have 100% of the data then win rate is the absolute most important stat and really the only stat worth analysing, obviously we don’t so it’s the most worthless stat unless there’s some extremely unusual outliers
But I think Mei is a hero that basically showcases how useless stats actually are, she has one of the lowest damage stats (or highest if we go by OB’s glitched stats), one of the lowest elimination rates, yet was one of the more prominent heroes in OWL this year because unsurprisingly, it’s hard to measure with units, the usefulness of putting a wall between you and your enemy
Also Overbuff has major flaws:
-Very little data on the Asian region which play so much differently to the rest
-GM stats get wrecked at the start of a new season when all GMs are reset back to 3900
-Private profiles skewer the data towards higher skilled players
-Still classifies Sym as a Support yet took the time of day to rename McCree to Cassidy (okay this is more a critique than a flaw but it annoys me okay)
I mean, yeah, but my point is that there’s not really any point in quoting overbuff, as the devs can already see overbuff + have a more accurate version of it.
And this part is the part I say we could argue about. Stats have infinite variables, no stat can be taken by itself. I would actually argue that stats like winrate are useless because of how they are calulcated which is why I disagree with the last experimental patch, clearly they used winrate only. Which was a really terrible idea.
Pickrate is also pretty terrible but at least gives people an idea of the meta.
This is actually really important. I know we all exaggerate a lot, but opinions are actually the MOST important metric to balance around.
I’m assuming blizzard takes exaggerated player feedback, then tones it down a LOT then puts suggested changes on an experimental.
Example:
Person 1 says Tracer is the most broken hero in the game rn, needs massive nerfs ASAP
100 people agree
Blizzard nerfs tracer, but not MASSIVELY. Just light nerfs.
Agreed completely
Agreed completely
Very well put!!
Completely true, a while ago I calculated winrate fluctuations.
(Data taken from september)
Bastion had 44.27% winrate fluctuations EVERDAY
Ana (most data/PR at the time) had a 5.36% winrate fluctuation EVERYDAY!
Thats crazy.
And we can prove this, last month zarya had a 61% winrate, now its down to 53%. That literally makes no sense since nothing has changed. Pharah was also at 61% at one point, now its down to 55%. Bastions was 62%, now down to 50%… etc… This is while nothing has changed
This is a massive hyperbole, of course that wouldn’t be the case.
