I mean, how does the “work cycle” behind a balance patch work?
Blizzard takes their perspective of the current game and decides what to do with it via stats, a little input from Reddit & these forums? (doubtful) and maybe some input from the “pros”
Throws it on the PTR to get an opinion and ultimately releases to live without much changing even if theirs crazy uproar against it. Cause the PTR doesn’t give reliable statistics and apparently that’s there way of efficiently balancing the game.
Aka why this game has so many issues.
Edit: Oh, yeah they apparently do some internal testing as well. 
4 Likes
There’s a lot we know and a lot we don’t. For example, we know:
- The pros have multiple ways to provide feedback. Either they’re talked to directly by devs, or they talk through discord channels.
- The devs are always planning far ahead. They are always working on multiple heroes (last figure I heard was 6, with many “Ready to go”) so they’re considering any future heroes. At 3 heroes a year, that’s at least 2+ years in forward thinking if the pace doesn’t change.
- They are taking into account any ruleset changes that might be coming. For example, they had to balance goats around the potential for 2-2-2 for almost a year.
- We know they do internal testing, and that at least one of the developers are in grandmaster, some in diamond, presumably some lower.
- We can reasonably assume that the PTR is mostly for bugfixing and not really for balance testing.
What isn’t known is exactly how much iteration is done or how they weight different opinions (pros vs dev feelings, forums, streamers, etc). It’s also hard to say how they choose who gets buffs or nerfs sometimes, at times it just feels random.
It involves a Ougi board and a roulette wheel.
5 Likes
This was more of what I wanted to know. How do they acquire data, and what is their source, and how is it legit?
and a bowl of California’s finest medical grade.
1 Like
Snoop Dogg Approved™
3 Likes
As i said in another post, PTR works like a Pre-Show of everything that’s about to go down. It doesn’t really matter that much if the community provides constructive criticism/feedback.
I used to think that too. Like firing up an engine for the first time after a rebuild, to make sure they didn’t miss any functionally important parts
1 Like
They have a lot of really powerful monitoring and stat gathering tools at their disposal. Way more than what overbuff has. If it’s any comparison, Overbuff operates purely on what it can scrape from your career profile on their website. You can kind of get a glimpse at it when watching the OWL, there are tons of different weird stats they have. MonteCristo even sometimes talks about how you have to have special access to their databases and all of the stats they can provide, it isn’t something public. I can only assume that the developers have even more to work with.
I’d assume many of them are handled with internally developed programs. I say that because in a video advertising Kibana (I think it was kibana?) the Blizzard team showed quite a bit of their internal dev-ops stuff and there was a great deal of in-house tools being created. That, and pretty much everything else seems to be specifically made for the game. Map editor, scripting tools, game engine, etc.
I know they also sometimes send out feedback emails but honestly I’d be surprised if they cared much about those. I’ve never felt companies paid too much mind to those. In chronodota’s “is overwatch dying?” video he speaks about how even the pro feedback channels in the OWL discord feels like it’s ignored most of the time, so people aren’t even bothering as much.
any tools are as good as their user aren’t they?
Being the ones behind the game, they must have all the data.
But what do you watch for? Playtime? Pickrates? Win Rates?
Or is it even more specific stuff like “Teamfight winrate probability when X ability is activated at Y time?”
How do you set the criteria for a teamfight being a win or lose?
What is the time window for the ability to be activated?
I mean that’s really the question every dev has to ask. For example, soldier. His pickrates are godawful, but because they’re so low, his winrates fluctuate between 40% and upwards of 60%. Then you’ve got heroes like moira with huge pickrates and fairly tame winrates. I would be surprised if they only balanced around stats because they’re so easy to misinterperet.
It’s kind of arbitrary but the OWL does this a lot. They’ll say things like “___ wins ___ percent of fights when ____ happens first” or something along those lines. It’s usually just to point out weird things with teams playstyles though, like how vancouver would win almost every fight as long as Bumper died first lol. I don’t remember how they determine what’s a teamfight based on the stats for the OWL. Like there will always be weird edgecases that don’t make sense but it seemed decent enough.
I dont think its ever been revealed if the OW devs do this too though. From what I understand, the people who work in the stat department for the OWL are completely seperate from the OW team. I dunno if they watch for things like “Players are 80% more likely to win if ____ does ___ first” that would be interesting to hear from them on. Maybe they have a machine learning tool to look at trends? “a disproportionate amount of teamfights are decided by who casts ____” that would be kinda cool.
That would shift the study from emotions to numbers. And until its done, its effectiveness is questionable.
Also, the criteria is a lot of grey areas
That’s what every tool does though, they’re purpose made to boil everything down to easily digestible trends. The important part is responsibly balancing artistic vision vs what’s fun vs what’s reality.
I think Blizz get a lot of negative feedback because people think they either care too much about their vision or too little about what’s fun for more than just a small percentage of players.
If anything I give them negative feedback becuase it always feels like the current patch is being balanced for a patch that we won’t get for months. 
Could be foresight.
But why would a future proof patch work in the present where the future patch doesn’t even come online.
Also, feedback can be questionable.
For Meta’s we could have them look for different data. Like TTK, Health pool and DPS effect on it. Point capture time of GOATS vs GOATS and then others.
Its like CSI: OWL but with 8Head computers and miles of excel sheets
Ya that’s true. I wonder how much they have to consider their slow patch cycle too. They get a lot of hate for having metas that last for months because they patch the game so slowly, but do they embrace it? Pretty much since the game’s launch we’ve seen metas that last for at least an entire competitive season, if not more. I’m sure it forces them to try and theorycraft about what might happen a lot more… this might just be hindsight but it seems like they miss a lot of obvious stuff lol and it makes me think they do embrace their slow patch cycle rather than fear it.
But sometimes to me it feels like they are waiting, seeing if any new trends emerge themselves without the devs having to step in, watching for patterns. And when the graph goes flat, they put a new balance patch in motion
I wonder what their waiting period is. I remember at the pro level at least, for a DPS comp to be viable enough to win the grand prize, it took 3 full stages of the OWL along with however many months teams were playing moira goats in contendors. Yet that was called a success. I dunno if I’d call that a success just because it happened eventually…
It would be nice if they would at least drop some more hints. Like for example, halt hook was an extremely solid meta for quite a while. Even if players didn’t organically figure out a “counter” playstyle (using dva or mei to deny the combo) they probably would have still chosen not to nerf hog or orisa because they knew Sigma was coming, who would likely force a new meta. It just bothers me because it leaves people guessing if something’s actually going to change or not.
Blizzard has what they call, the triangle, its composed of 3 points
The community
The developers
The stats
If 2 or more of these agree on something that something will be changed