About recent balance update

General Discussion
1

Following the article we see that a lot of surprising facts were revealed, such as:

Since shifting to a 5v5 format, we see many players focusing more on the tank role given the heightened importance of a single tank in role queue modes. One of the most talked about heroes on the roster right now is D.Va. With buffs after the last Beta to her Defense Matrix and Micro Missiles, she can put up a fight longer than she could before the launch of Overwatch 2. Despite many players saying she’s one of the best tanks in the game, Reinhardt, Sigma, and Zarya actually currently lead up the competition with an average 53% unmirrored win rate. Zarya’s new ability to choose where to place each of her two Particle Barriers often allows her to build up to high energy quickly, which is something we’re currently keeping a close eye on for potential changes next season.

and:

The damage role has seen many hero combinations across all ranks, but we’ve noticed some popular picks among our players. Genji and Sojourn started as the most popular heroes when Overwatch 2 launched, but they’ve since leveled out to the rest of the damage roster. Genji started off with a dominating win-rate, and although it’s lowered slightly since launch (currently at 52%), we want to make sure he’s not dominating the playing field. We’ll watch Genji carefully throughout the season to ensure he doesn’t slice up the competition too much and may make adjustments in Season Two. One way we are considering doing this is through an adjustment to the damage role passive, which has particularly benefited heroes like Genji.

TL;DR of which could be written as "Zarya & D.va are fine, almost all tanks are doing well + Sojourn & Genji are mostly fine"

(I admit - that's hard to accept but let's be pragmatic here, numbers don't lie) + Sombra wasn't even mentioned

However I’d like to know:

  • Is this across all tiers of play? (in which case the value of such data is dubious)
  • Is this on a tier they balance around (like they stated: ~GM and above)?
  • Do they also look at pick-rate numbers? 80% of win-rate with 0.001% of pick-rate is still “garbage-tier”

Because without knowing that data (at least) it’s not sufficient to conclude the state of balance, even worse - it can lead to some misconceptions. Blizzard themselves were prone to misrepresenting the data in the past - hence 3 reworks of Symmetra for instance.

Thoughts?

4 Likes
2

I haven’t stepped into Comp myself but going from talk on the forums there’s been a massive rank crush, with everyone squeezed into a small space. Wouldn’t that screw up the stats even if they went by high ranks only, since there’s even fewer people populating top ranks than usual?

3

I find the given data pretty meaningless.

And while on one hand I don’t think the game is busted and needs patching ASAP, I also don’t think waiting until S2 for it is good enough.

A new patch in… Lets say 15 days, would be good to shake things up.

I also think it’s weird how they mention WR… I presume those are averages? Meaning, if a hero was dreafully problematic and 2 data points were respectively 100% WR and 0% WR, the resulting average would be 50% WR, i.e., the average gives the illusion of a well rounded hero.

To be specific, if Sym has a particularly high WR in Lijang with basically zero pick rate in every other map, then that’s a problematic hero, it doesn’t matter what their WR averages to, since players can’t make the hero work outside of a very specific situation.

Similarly, Rein could have good WR with Lucio and few maps and bad WR in other situations, while Zarya and Dva have smoothed out WR across the board simply for being better at most things and more reliable.

3 Likes
4

This is a fair point. The only hope is that they refer to MMR and not much SR when talking balance. MMR was not “reset” so I still see some validity of data even if they indeed only took a look at the “upper echelon”.

The dilemma they have (also openly stated) is when they need some sort of representative set of data, i.e. at least for long enough. When they told they plan to patch the game very ~6 weeks balance-wise in OW2 I understood it will be much more feedback-driven than data-driven because the sufficient data set won’t simply be available in such a short time-frame. I’m not saying that one way is better than another - we all saw to where their “data-driven” approach led in OW1 (albeit to be fair: having the data and having the expertise / ability / talent to use that data are not the same thing)

2 Likes
5

It’s this for sure. Which is laughable. There’s no other reasonable way to explain them saying that Torb and Sym are the strongest dps.

They need to start fully releasing the data to justify their decisions.

3 Likes
6

I kinda like it but hate point where zen can’t fully use kitsune rush coze some DPS starts whining on forum that its “to op”.
Now zen have only mobility speed but fire rate is still same.

7

That’s the thing, 6 weeks is a lot of time already. I would personally hope for a major patch every season, but with very minor patches at least every month, maybe even every 2 weeks.

There was also a shake up in lead hero designer leadership and if the team needs a season to pick up the pace, that’s okay too. I can be patient - the well eventually dries, as it did in OW1, but we cool for now.

I would also like to point out being “disconnected”, “out of touch” and “dismissive of player’s issues” has always been my personal feedback to T4’s balancing and the latest post doesn’t strike me as any better. Data is good but you can’t talk legs into a fish, meaning if the WR shows X isn’t a problem but players aren’t enjoying the game, then players aren’t enjoying the game, the data becomes moot. Fortunately that’s not the case, game is good, issues are minor so far - however I don’t have trust in T4 to keep it that way, they destroyed one game already.

There’s also OWL - arguably the most limited meta we have where problems and exacerbated and highlighted… Anyone can see what’s going on in there.

8

That’s not the case. The patch note was a tad misleading. Zen was just bug fixed to have the proper benefits applied (he was getting stronger effect previously).

1 Like
9

Well it looked that it wasn’t in OW1 because people also were asking “how much is needed” and they were hard-locked onto “several months”.

The biggest issue is one they pointed out themselves - what’s balanced isn’t always fun and what’s fun isn’t always balanced. It’s odd to see them forgetting their own piece of advise.

10

I do like the idea of an additional minor patch or two mid-season. Every 2 weeks sounds extreme to me, but that is just because I don’t want the game to become league of legends where every 2 weeks everything you learned becomes worthless, even if Overwatch balance patches aren’t anywhere near that hardcore in changing characters.

The data they showed is meaningless. But they are not looking at just that tiny amount of data, literally in the same blog they mention that they saw players talk about Dva as one of the best tanks in the game, but for most of the playerbase that statement isn’t backed up by facts. They are instead locking at Zarya and Doomfist, likely both at opposite ends of the WR spectrum.

Sym has historically always had one of the highest winrates, that is not a secret. People that play her either play her well, in the right situations or know when to swap off her. If a team plays against a Phara+Mercy and swap to Dva+Ashe+Widow they might still get rolled if those 3 players are not used to playing those characters on a regular basis like a Sym 1-trick would be used to playing Sym.

Blizzard has the stats and player feedback, they are looking at them both and see which is the case. During one of her reworks I remember them mentioning that they know people were swapping off Sym after 1st point in 2CP was lost, because her old ult TP was practically useless on 2nd point.

Some of the beta blog post show just how in-depth they are going, and all of them still don’t show the full picture and amount of effort that goes into this decision making.

We should remember that this is “just” a blog post. They could release all the stats per hero, unmirrored, with and without mid-game swaps and on all ranks. But people throw games when there is a Mercy on their team and a random streamer says “Mercy bad”. Imagine how bad it really would be if Blizzard officially tells its playerbase “you are 2% more likely to lose with a Mercy”, or the other end “you are 2% more likely to win” and then that player gets flamed for playing the “easy character”. There is no good ending.

If all characters have a WR between 45-55%, I think the devs are doing a good job at this thing. That means if I play the worst character in the game and play 100 games, I only lose 10 more games than somebody playing the best character, that might have less impact than my personal performance and could probably still climb rank consistently (as 1-tricks have proven in the past).

Doesn’t mean we shouldn’t still be critical, and the questions are fair, they could definitely clarify that a bit more. But they are clearly doing more than that single blog post implies.

2 Likes
11

I would double this. I think patching any more frequently than 30 days is too often personally. Basically doing it every 30 days would give you a start of season patch and a mid-season patch. And that would be just about perfect.

I think that most tanks are basically OK though. Maybe Zarya could use a slight nerf. JK and Doom need a buff, but we have to be careful with JK because it could be easy to overbuff her. Supports also seem fine to me personally.

I think DPS might be the most out of balance, but even then I don’t think anyone is so OP that they have to get an instant nerf or anything. Sombra and Sojourn are clearly the best though. Genji isn’t as OP as people think

1 Like
12

I love the people that argue …“Blizzard knows better they have the Data that we dont”

They’ve had the data since day 1 of Overwatch. How has that worked out. Their data told them the game needed Brig to stop dives, that we needed more shields and stuns. Now their data tells them we need hit scans dont need recoil, Tanks should decide who wins.

13

I personally don’t think D.Va is a problem by any means but according to several available opinions of the high GM/top-500 she is S-tier (i.e. almost a “must pick”). You can find several examples, quick look leads to this list. While the numbers are important - I believe just averaging out all tiers of play does not do balance enough justice.

I feel like 45% vs 55% is too big of a difference. I think occasional 45% or 55% are okay but the spread should be tighter on the balance side - because again, we are talking about statistically significant (check wiki) set of data and 45% or 55% on that reveals a strong bias

For context: if one hero has 45% win-rate and another one has 55% then on average you’re giving up 2 games out of 20 if you play the undeperforming one. Those are big differences.

I am not questioning their data, but rather their methods of dealing with said data.

14

I don’t think D.Va is a problem. She has historically, anytime she is even close to good, had all the high level players flock to her. People in the top 500 and OWL can use her in ways that even people at high masters cannot.

She isn’t a problem because for the vast majority of players from (let’s say bronze all the way to diamond, possibly even masters) are going to get more value out of other tanks.

I think the problem with this community is they tend to focus way too much on GM pick rates and ignore almost all other statistics when they have statistics available. It’s lead to some very poor community balance suggestions. I’ll also note that GM opinions didn’t often even match up with GM statistics. The top picked hero didn’t always have a great win rate necessarily. Group think is a weird thing.

1 Like
15

Which is my point number 1:

and

Which is my point number 3:

I advocate for a more careful treatment of data and not reading too much into the one they presented so far because without context it can be extremely misleading.

1 Like
16

Not sure I get what you meant here.

You won’t have to relearn anything, literally absolutely nothing, after a minor patch. Minor patch should be just very fine tuning.

Then why should anyone care about it? I still have no context for it and showing cherry picked data only strikes me as the same old problematic attitude that ravaged one game already.

30 sounds good too.

We’re getting 60 though. Guess I can hope it’s just because it’s S1 of a new game and there’s a new lead hero designer. Pace might pick up as the game develops further. There’s nothing wrong with nothing right now…

I’m really, really, just very doubtful of the devs capabilities after what happened in OW1. They say they’ll do better and they showed some good work in the betas, but I’m not about to give them a vote of confidence again, they don’t deserve it, and when they say “they’re watching closely” I just get flashback from Moth Meta and GOATS where devs did in 12 months what Respawn does in 4 weeks in Apex.

1 Like
17

I agree with you. I just don’t actually think the community should have the data. We are just not good enough at interpreting it and it would add more fuel to the fire in balance discussions instead of calming it.

I personally love data. And I would like to have it. But I can already tell that the community would use it to beat team 4 over the head to reinforce their already preconceived notions of balance instead of actually changing their minds based on what they see in the data.

2 Likes
18

The community was asking them why they take so long to do balance patches and they replied they need time to collect the data. And then when people asked how long they need for a balance data collection cycle, the response was in the ballpark of months (which also correlates with their rate of significant balance patches in OW1)

Same here. Yet again, I think the data they collect could be perfectly fine while what they do with the data might be way off what’s actually needed. Data Science is a real thing (expertise track, job, career - you name it) and data analysis is crucial for things like that. I feel it’s sorely missing when anyone with a glimpse of education on the subject can ask questions like those I mentioned in he main post.

19

The way they were talking before launch, I was expecting the following balance update schedule:

Season start: Big update to shake up prior season meta.

Week 3: small changes to adjust new hero for comp play. Skip in no new character.

Week 5: Mid season patch of mostly small number adjustments to tweak current meta.

Instead we’re getting a season start patch, and that’s it until next season. Unless OWL develops a boring meta. Then changes will fly fast and furious. Sigh.

20

You can always misuse a tool, that doesn’t mean one shouldn’t have it.

Devs are welcome to share data and then let CMs handle forum stupidity,it’s basically what already happens. I don’t think anyone is gonna vandalise pick rates in sole dev’s house door. Fuel this fire, nothing wrong with it.

Heck, for years pros complained about the lack of communication channel.

You mean “significant” as in “I was hit by a car and almost died, that’s a significant event”?

Because OW1 was nothing but stupid and slow patching. In one breath I dreaded every patch while also wishing I had a one every month rather than every 6.

1 Like