OW2's Balance Philosophy has some major flaws

General Discussion
1

The Game Balance Philosophy used by the devs has been wrong at a fundamental level for the last 7 years.

Balance is about Choice Design. It’s related to the branch of mathematics called “Game Theory”.

https://quantimschmitz.com/2022/12/07/a-better-way-to-measure-game-balance-using-game-theory/?amp=1
https://archive.gamedev.net/archive/reference/articles/article1765.html

Basically, Balance is just a Numeric count of Worthwhile+UniqueNovelty choices for an individual player.

With “Worthwhile” representing how people typically attempt to minmax their Time, while playing Overwatch to get the maximum enjoyment out of it.

Which for Bronze, “Worthwhile” is largely “How fun is it to play a hero, in their visuals/audio/interactivity”.

And at Top500, “Worthwhile” is largely “How likely is picking this hero going to make me win?”

Which essentially boils down to “Worthwhile Hero choices, for a given Role, and a given ELO bracket”

To boil that down even further at high ELO, “Balance at high ELO is just a Numeric count of S,A,B tier tierlist choices, per Role”.

That has big implications such as “Role vs Role Balancing” mostly not existing.

It somewhat exists in a something like “Bastion and Ana limits viable Tank hero options”.

But Bastion and Ana are not on the Tank tierlist.

And the only other form of “Role vs Role Balancing” would be attempting to get the Role Queue population to roughly a 1-2-2 shape. And people typically pick their Role based on “Play for Fun”, rather than “Play to Win”.

The Role vs Role stuff, is largely just an “Aesthetic Preference”. Which needs to be weighed against all the other player preferences, for player retention. Which is important, but not Objective in a Numeric way.

Similarly the High ELO preference for more mechanically difficult heroes, is an Aesthetic Preference. Which is important, but not Objective in a Numeric way.

Additionally, Winrate math is broken for the devs, and has been broken this entire time for the last 7 years. Since it’s not a straight winrate for the player. It’s pickrate timeslices on a hero (I.e. Usage Rate), versus the total duration of the match, times winrate.

Which basically makes it so that “Quick on Attack” heroes have artificially lower winrate. (I.e. Widow, Sojourn, Kiriko)

And “Slow on Defence” heroes have artificially higher winrates. (I.e. Torb, Symm, Brig)

Or it’s a Push map, where Sigma is doing payload duty the entire duration of the match gives an artificially high winrate.

Which basically means winrate is mostly useless gibberish.

Because it only works in games that don’t have mid-game hero switching.

This is a gigantic oversight by the Devs, that has spoiled most of their balancing efforts. Because the math is wrong.

They’ve been following a “compass” that points in the wrong directions.

But also another thing to consider is that:

  • If they aren’t actively trying to sync “Play for Fun” to be the same as “Play to Win”.
  • Then the higher ELO you go, usually that means the less “Fun to Play” the game is overall. It’s a crapshoot on sheer luck if it’s not.

That’s a problem.

But that’s fine, if they stop using “Win/Loss Ratio Balancing”, and use “Popularity/Frustration Ratio Balancing”.

  • With “Popularity” representing how Fun a hero is to Play-As.
  • And “Frustration” representing how Frustrating a hero is to Play-Against.

If the devs minmax on that, then “Play for Fun” and “Play to Win” are as synced as possible.

That said, that means that heroes that …

    1. Aren’t that popular
    1. Are very frustrating

…“Should” be balanced for a low pickrate.

Where as heroes that…

    1. Are very popular
    1. Are not frustrating

…“Should” be balanced for a high pickrate.

That said, Balance doesn’t cover everything related to hero changes/adjustments. Despite players (and even the devs) often referring to basically every hero adjustment as “Balancing”.

Something that usually isn’t “Balancing”, is the Qualitative aspects of Counterplay Design. I.e. “How a player interaction feels”.

And as we’ve seen with a lot of heroes (I.e. Roadhog), they hard nerfed him, despite thinking he had a low winrate. Because his Counterplay design “feels” really frustrating Qualitatively, even if it’s supposedly “fair” from a purely Numeric standpoint.

This is the devs starting to pick up on the idea that Qualitative Counterplay Design is more important than Winrate.

Because the point is to make the game consistently Fun, and not too Frustrating, for the largest amount of players.

And if they aren’t doing that, they get less players, which translates to less dollars.

So the way Overwatch has been “Balanced” is wrong mathematically, qualitatively, and financially.

_

But they could easily fix that.

Since as one of the Co-Founders of the Civilization series put it.

“Given the opportunity, players will optimize the fun out of a game”; , “One of the responsibilities of designers is to protect the player from themselves."

So it’s a game developer’s responsibility to prevent situations where the way to win, isn’t fun.

Since a Game with perfectly optimized winrates, but it’s not fun, is a bad game.

And it’s a great way to get people to quit, if the better you get at the game, the less fun it becomes.

But like said, that’s a very fixable problem.

12 Likes
2

As much as I want to agree with everything you wrote, I can only agree with about 75% of it.

The last 25% has to do with previous mistakes, and the ability to resolve these mistakes.

I would start with GOATS. This is to highlight Blizzard developers repeated attempt to deny Occam’s Razor.

GOATS was created as a push back against high damage. This isn’t what it became, but the original intention was, high damage rules the landscape and we can’t compete, so what does our team do?

Well, that team picked 3 tanks and 3 supports to mitigate and sustain their way to victory. Later on, this was refined to include more damage aspects, and fine tuning for the largest stage of Overwatch league.

It had absolutly nothing to do with support numbers and AOE healing. This was not a consideration until the fine tuning of the composition. And instead of lowering all damage and healing, making DPS relevent in the game again in consideration to how much healing was possible. Blizzard decided to nerf Bridgette 15 times and buff damage, the original culprit, over and over. GOATS prevailed through everything. The reason it prevailed was because Blizzard didn’t understand the most basic function, the highest common denominator that created it in the first place. This type of absent minded knee jerk development creates more problems then it solves.

How did Blizzard finally get rid of GOATS?

The introduction of 2/2/2 role lock. Why?

Because they couldn’t figure out another way to balance the game, and went to the “break in case of lack of knowledge” development tool.

This is the same exact reason we got 5v5 in the first place.

Tanks were nerfed when role lock was implemented. It reduced the number of tank players in the game significantly. As always, it wasn’t Blizzard’s fault, people just need to adapt…right?

To reduce queue times, the only way Blizzard could think to do it was to remove 2 tanks from every lobby. The compensation was to give limited buffs to the tanks that exist. Like many games before this, even World of Warcraft, when you reduce the number required, you create a hierarchy of that role. One will be the utmost best, the majority will fall in middle and there will be a select few at the bottom. Look at that, its the Overwatch we play today.

Blizzard won’t fix this by moving back to 6v6. Blizzard doesn’t make mistakes, as we all know.

Blizzard has 2 ready options today;

  1. Homogenize all tanks. Make every last one of them roughly the same. It will totaly kill the role as it is, and no one will want to play tank anymore and Blizzard will have the authority to remove the role.

  2. Make every tank super powerful in their own right. Make it so each one is so strong that the very presence of the tank is what decides the game. This issue will please tank players, but nullifying the existance of the other 4 players on each team will more then likely have alternate reprecussions.

As I said, that is what can be done reasonably fast. There is another alternative that can be done in the long game.

  1. Populate the tank roster to the point were no one tank is countered by any other tank. This would require that Blizzard introduce an unknown amount of playable tank heroes over the course of an untold amount of time. I don’t actually think this to be a realistic measure.

You are right about the way they balance the game. They have the exact issues you claim them to have. When it comes to 5v5 and tank game play, I don’t think Blizzard has what it takes to dig themselves out of the hole they dug.

7 Likes
3

Well ironically they could have fixed most the problem with GOATs by just having a 2x limit on Support.
🐐 GOATs could stop new Content

But beyond that, just revert the knockback buffs, and nerf all the defenses behind barriers that can’t be blocked by Defense Matrix. Except for barrierless offtanks.
https://www.youtube.com/watch?v=CgeQidvigbo

I mean, the problems with Tanks basically boils down to

  1. AntiNade
  2. Snipers
2 Likes
4

This is all very well spelled out and organized. Well done.

It is also very complicated.

Allow me to ruin the sentiment by making a brash and less complicated statement of my own.

Dva is fun but can be more fun and needs more buffs to bring about that fun.

As I am a connoisseur of Dva, my authority on the matter is as follows:

Micro missiles have proximity tracking

Hp / armor split is 325/325

Defense matrix increased to 12 Meters

Armor is 40% reduction to beams from 30%

Booster damage from 15 to 20

Call mech now stuns on hit

I have outlined steps by which to make the game better. Better for me. More fun is the goal. More fun for me.

5

Where you are correct would be in my eyes something like;

In mid ranks you want to play Doomfist or Hammond, they are fun to play. But if you want to win and rank up, you need to play Orisa, or Zarya, who you may not enjoy playing.

There will always be outliers and people who can excell at playing Doomfist and Hammond at any rank. But for the base player in the game, these tanks aren’t an option because of balance decisions made by Blizzard.

And if you want to play a mid tier tank like Reinhardt, D.Va, Winston, or Ramattra, well, you will play them until the other tank swaps to counter you and then you will both end up on Orisa. Again, if you are looking to win.

Blizzard threw the tank players under the bus, for a standard 5v5 game, that will not be changing.

I feel this is an undeniable reality.

But the solution to these things, outside of nerfing those two things you mentioned, is still the same thing I posted about. You can homogenize the tanks to overcome Nade and Snipers. Or you can make them super powerful Souls/Borne bosses. Neither of those things will happen either. Blizzard is stuck when it comes to tanks, unless they are willing to change some fundamental portions of their game.

Then the enemy tank picks Zarya (I know I would if I want to win) and all your fun goes away.

6

Incidentally, the strongest variable for D.va balancing I’ve noticed, is the delay between DM usages.

7

Oh without doubt.

If I recall there was a point in time where the delay between dm usage was 0.5 seconds.

Reactionary usage of DM will allow for mind games as opposed to playing around the cooldown which is very strong.

Her worst state was when DM delay was 2 seconds after the goats nerfs. Absolutely destroyed her viability

Her current 1 second delay is the odd in between spot where you can time abilities such as grav from Zarya on reaction to have it not get eaten but isn’t so long that there is zero chance of mind games for the Dva player to potentially outplay.

0.75 second delay would be very interesting

2 Likes
8

Well, yeah.

But that would be really easy to do though for the Nade stuff, and it’s a nice catch-all for CC annoyance in general.

9

It changes the dynamic of how the game functions in a way that the average person cannot precieve.

If you give tanks the autonomy over their character faster then the other characters in regard to Nade or other CC. It will be the same argument on a different date;

“Tanks are the role that determine the outcome of the game because…”

They will not understand that for tanking to survive, these consessions need to be made. All they care about is their role, and their hero.

With the recent changes to support, which are laughable, Blizzard has no intention of such a drastic tank change. Also, Hog would still get decimated by Nade, even with the change you mention.

10

Well I think the recent Support change provides a good amount of rhetorical leverage that “number tweaking isn’t good enough by itself.”

Where as the Zen change, shows that they are starting to think about fixing how it “feels” to play against a hero.

Heck, the Hog changes, where they hard nerfed him, shows that they acknowledge that how it “feels” to play against a hero is important.

11

If I may add this, to some extent, is why we got double shield after GOATS.

Dmg was so high that dive comps got melted on arrival and playing guys like Winston was suicide. Rush comps did not even make it to the fight as they had to use all their barriers, abilities and defenses before the fight started. Poke comps died too as you could focus down the barrier/defense way too quickly.

This meant all three major comp styles had issues as none of them could survive the dmg storm, unless you took the two poke tanks with their barriers and put them on a team together, then you had some sustain so beat everyone else, not because you were good, but they all lacked something they needed. This could not (and wasn’t) fixed by buffing dmg as while yes that hurt the barrier bois it hurt everyone else too so really had no effect.

Why am I bringing this up, as this mindset still affects current balance issues. The two problems are promoting what you want to happen vs demoting what’s stopping it and simple and one-patch solutions.

Both GOATS and double barrier were issues and we all wanted to see the same outcome, dmg to define games more and feel more impactful. The problem was that buffing dmg would not bring that, nerfing the things holding it back would.

They also kept trying to fix the game with one move, be it nerfing Brig, endless different hitscan buffs, some of the reworked heroes like Torb, anything they did with Sombra. None of these worked as the problems were so ingrained and the balance so screwed you could not really fix it that easily, and every failed attempt left baggage in the game.

The answer would have been (and should be if we get such out of wack balance again) to hit the culprit hard and nothing else to blow up the issue. The result will be a messy and unbalanced game filled with unforeseen problems, but small ones that are easy to fix, rather than massive ingrained ones.

1 Like
12

As a precursor, I am not saying that damage was the only factor, or that what GOATS and to a smaller extent Double Barrier did not also have support and mitigation issues as they progressed. Only that high damage caused the thinking that became 3/0/3 GOATS.

As I recall, the Double Sniper and Grav Dragon META was formed around this time.

Yes. 100%

Though at the time I disagreed with most of the ideas coming from the Damage Player camp at the time. They were a retelling of, “too much healing, not enough pew pew” or “I don’t wanna do X,Y,Z and my role is all that matters.”

At the time, players were rejoicing because Bridgette was being nerfed so frequently. Problem is, GOATS remained. Logically Brigette was not the problem. And again, the damage player camp could ony say nerf her more, and Lucio, Moira, Zen too. Seems familier to what we have today.

The problem, as GF pointed out, Blizzard doesn’t balance the game with the right mind set. They have to change the mind set to fix the problem.

1 Like
13

I think there is however a way to make it a realistic measure: rebalance existing heroes (including non-tanks) so that no one non-tank hero totally shuts down a specific tank or largely shuts down >60% of tanks.

Tanks have one really big problem, only having one makes counter picking them really easy and low cost and the tanks will never be able to out-swap the other roles. As a result we need to build the tanks to be the most resilient role to counter picking, at the moment they are the weakest.

This means we cannot have things like Bastion over the last two seasons. He has been a decently viable option across the game vs every tank but Orisa and Dva. For tanks to be viable that cannot happen, Bastion can still be good, still be good vs some tanks but not like that, no role counters.

We also cannot have interactions like Ana vs Hog, which is basically an unwinnable matchup for Hog, or it was before the rework. Now I’d bet one Ana sure, but Hog has some wiggle room against her, this is a good step.

As for how we would end up countering opponent tanks, we target that tank specifically and we can’t do it with one easy swap. The 3 obvious ways; swap your tank, swap multiple non-tanks, swap your style.

I think having a tank hard counter a different tank is not the same issue as they can swap against each other one for one, which they cannot do vs other roles. If you gear the whole comp to beat their tank then fair enough if you beat it. Tanks currently have clear styles of play, play counter to those styles to beat them (e.g. play wide and distanced vs Rein or high ground and conservative vs Queen).

Also to reply XD

Agreed, these are complex issues and a lot was going on.

Yes, again we’re pulling from memory here but very early 2-2-2 seemed to be comps that won fights really quickly, like those, as the best way to stop yourself getting killed was kill em first, rather than endurance or mobility or anything else. Double barrier was sort of the only one that gave another option and won out.

And I have high(ish) hopes they will as…

  • OW2 (the real OW2) is dead and we now have a PvP focused game. I can understand going for quick fixes if you have your dev attention aimed at PvE stuff.
  • They have just gone after support pretty hard, after months of putting it off. I cannot think of the last time they went this hard, this wide on a problem, especially one with so many vocal defenders. It’s not quite what I think you and I are after but its a step in the right direction.
  • Similarly the hog rework is a good step
  • And we have a new step-dad called Microsoft, that will cause changes, likely not in this area but honestly god knows.

edit to avoid a double post

2 Likes
14

Remember early OW1 where DM was a cooldown on E? Those were weird times.

That being said, with the post at large I agree that like it or not, enjoyability is (or, well, should be) a large factor in the balance of competitive games. Sure, something could be competitively balanced, but if it’s not fun, why bother? This is why there’s so many arguments on the topic of hero swapping. Yes, you could argue that X strong hero is balanced because they get shut down by Y hero. But for people who don’t already play or don’t like playing as Y hero, it makes the game unfun, and if it’s not fun, why bother? I think Lincoln said that once. Or maybe I’m confused.
Regardless, the main takeaway is that a game being balanced doesn’t matter, if nobody plays that game.

1 Like
15

I also thought about this problem here
Does it even make sense to measure winrate for heroes? - General Discussion - Overwatch Forums (blizzard.com)

but for a different reason (distortion from matchmaking). I proposed a different measurement method and would like your opinion on that.

This is the reason I dislike battle royale because it’s part of the core idea imo. I don’t understand how it’s so popular

16

They’re actually not “balancing” to keep the game balanced, they’re “balancing” to force certain metas to keep the game “feeling fresh”.

17

Well kinda, but the other way around.

They are just moving stuff out of the meta that people complain about. But they aren’t fixing the root cause reasons why people complain about those heroes.

Or if they do it takes them a long time to get there.

18

This is unfortunately the only part I disagree with.

In a recent livestream with Cyx and the lead hero designer, the LHD said that he was a Roadhog main. This makes me think that the only reason this rework kept the “fun” of Hog was due to bias (as usual) and not an actual change in balance philosophy.

Not tryna be that Doomer guy, but this team has repeatedly stomped on my trust time and time again.

19

You know what you’re kinda right.

It’s less of them changing their philosophy, but more that community complaints are overriding that, objectively broken, philosophy.

Like that Zen change is good. But I could have sworn that was because Flats and others screamed their head off for months for something like that.

The community shouldn’t have to get really vocally upset about something for the devs to know there’s a problem, and fix it properly.

1 Like
20

The Zen discord change was good (to a degree, 7 seconds may be a bit excessive). Them giving him 25 more hp and the 5 second change to harmony just smells like trouble.

You’d think that they’d notice something is wrong with Ana being the most picked hero in every rank seeing as they “use statistics” to figure things out, but no…