Heroes of the Storm devs explain when and why they make balance changes

We like putting out smaller balance tweaks, and want to do it more often. However, there are many levels of balance changes, and each come with their own risks and rewards.

Some examples of these, and their associated risks and rewards are:

Small, numeric balance changes.

These are changes like those made to Maiev in our last balance update, where we lowered her Health and increased the cooldown on Vault of the Wardens.

These are usually used to influence pick and/or win rates. Simply put, in making these changes we want to move a hero or talent in a direction of being picked more or less often, so we buff or nerf them.

PROS

• These are simple changes. The risk of unintentionally changing other parts of the game is relatively low. • As you mentioned, we can push out many of these changes in a short period of time, which can make the meta feel fresh and new.
• There is little Development and QA validation. We can make these changes in a relatively short amount of time, and the risk of bugs is much lower than other changes. This is a bigger deal than I think most people give credit for. More complicated changes just take longer to do to make sure that they come out in a polished state, and even then, the more complicated the changes, the more inherent risk that things will be missed.

CONS

• They can’t fix design-related problems.

1. To give an example, for Kael’thas’s recent changes, I had a strong desire to help his Level 16 Talent tier, which has been dominated by Fury of the Sunwell for a long time. From a pure win-rate perspective, his other talents on the tier were close. I could buff them, but I was pretty confident that the numeric buffs would have to be so high to offset the “cool” factor of Fury of the Sunwell, that they would have to be absurd to get people to pick them. This is partly why we eventually settled on Ignite as a competing talent on the tier.

• Pick rate and Win rate are two very different things, and it’s hard to make small, numeric balance changes without directly changing the win-rate of a hero.

1. A common problem we run into is that community perception simply doesn’t match what’s actually winning in the game right now, and if we tune up heroes who are “not performing well,” they would start to have unreasonably high win rates pretty quickly.
2. As an example, over the last year Probius, The Lost Vikings, Rexxar, Sgt. Hammer, and Samuro have pretty much been dominating in regards to win rate at high rank Hero League. I think it’s pretty fair to say that these heroes haven’t been considered overpowered over the last year. If we wanted to give them the Thrall small number buff treatment, as in your example, then these heroes would be pushing 60% and beyond pretty quickly.

Reworks

These are the large, dramatic changes to heroes, where we change a large part of their talent tree, and sometimes their base abilities as well. Our upcoming changes to Diablo and Lunara are examples of this.

PROS

• We can fix complicated design-related problems.
• We can take lessons that we’ve learned from playing with/against these heroes over time and apply current design philosophies to them.
• We can remove or mitigate frustrating aspects of these heroes and give them more healthy gameplay.

CONS

• We can alienate existing lovers of the hero. If we change them too much, inevitably some people are going to be upset that they play differently.
• These are a lot of work, and take a lot of time to validate and for QA to test. We can’t just push them out quickly to solve immediate cries for change.
• There’s a higher chance that reworks will result in unintended consequences for the meta due to the hero changing drastically—or because the hero is not being picked anymore due to the changes pushing them out of the current meta during their release, which results in a negative reception.

Smaller Talent Updates

These are changes where we change, remove, or add a small number of new talents to a hero. We do these less often, but we want to do more of them in the future. A recent example of this are changes to Kael’thas, where we changed a few talents but didn’t fundamentally change how he plays. I’m a huge fan of these, as we can have the flexibility of fixing smaller design problems without having to wait to do a full rework of a hero (say 1-2 talent tiers on a hero that we really want to fix).

PROS

• We have more flexibility. We can fix design-related problems that have plagued a hero for a long time without having to do a full rework.
• We can make a hero feel fresh and new without the risk of drastically changing them. We get a lot of the benefits of reworks with less of the downsides.

CONS

• They’re in that sweet spot where we can push these changes out much quicker than a normal rework, but that also means there are design changes that have less time for QA to validate. As an example, I’ll totally take the blame for Kael’thas’s Level 20 Pyroblast talent coming online at Level 10 when he came out. I made a simple mistake when implementing the new talent that we didn’t have enough time to properly test.
• It’s easy to go down the rabbit hole and turn a small talent update into a full rework. I also worry a bit that if we go in and fix some talents but not others, that there could be a negative reception around “well why didn’t you fix everything?”

You saw a bit of our reaction with the last balance patch, as we touched ~18 Heroes (a couple of them bringing very significant talent changes). There are a multitude of reasons that the amount of changes fluctuates each patch. A couple of these are:

Stability

• With the exception of a couple outliers, the current state of the game is actually pretty evenly balanced. All but a handful of Heroes are sitting between 45%-55% win rate (70% of the roster is actually between 48%-52%). If we inject too many changes at once, we’ll have a hard time narrowing down the cause/effect of any particular change. This can become problematic as every change has a ripple effect and we don’t want to upset the equilibrium too much at any given time.

Perception vs Reality

• This one is what keeps us live/balance designers up at night. Let’s say that I come read reddit and see 10 threads on the front page about the frustrations surrounding Hanzo and how over-powered he is. I immediately run to our data dashboard and see that his win-rate is sitting at 46% (this is what it was before the last round of nerfs). What is the correct course of action – technically he needs buffs right? This is where balancing on win-rate alone becomes impossible. The frustration of playing against Hanzo has surpassed the reality that he is winning a lot less than he should. To ‘fix’ the issue we would need to make design changes alongside number tweaks, which take a lot more consideration and testing. In this example we opted to send out more nerfs in order to relieve frustrations and are internally looking at design changes to fix the core problem. Once we are happy with that, we can release it alongside some buffs in order to stabilize the Hero. In a perfect world we want each Hero in the Nexus to feel powerful and rewarding to play, but just as importantly to feel fair to play against.

Razor Edges

• We have a handful of Heroes that walk on a razor’s edge of balance. Each member of the live/balance team has been working on the game since before it was launched. We have seen a lot of situations where a very small number buff can turn a borderline Hero into a nightmare. As the perfect example, let’s look at the crowd favorite—Chen. There was a time where Chen was completely un-bear-able (tee-hee) to play against. In certain compositions, he still is! His design is very binary in that you can either deal with him and his trait, or you can’t. This is something we can’t really fix through tuning and need to go into much deeper design changes to fix. We have a handful of these types of Heroes and they are very much on our radar.

Here’s our take on it:

• We always want to push the envelope of what our new characters can do. We think we can do that and still make these characters feel balanced and fair.
• Some Heroes probably pushed this envelope too far. We recently made balance changes to Genji, Hanzo and Tracer to reduce their mobility and force them to be a little more exposed when they try to do their thing. We’re still tuning them, but overall we think they’re in a much healthier state now than immediately after their respective releases.
• You can look forward to more Chromie changes in the future. I can’t give a date, but we’ve started to experiment with various ideas, like pulling in her range further but reducing her cast time, and making her W visible to enemies. All of this is still early in development, but the idea is to make her feel more interactive with enemies, and not create such a binary experience for Chromie herself (if enemies reach her, she dies and can’t do damage, otherwise she’s untouched).
• We’ve discussed coming up with creative abilities that can combat mobility, but haven’t really found one we liked that met all of our criteria so far. It would have to punish mobile characters without also hurting non-mobile ones, have clear implications of what it would stop and wouldn’t stop, and make sense from a fantasy and clarity standpoint.

I’ve been tanking a lot of the design work, and here’s a sneak peek at what you can expect (keep in mind that this is still a work in progress and nothing is set in stone):

Roles

• We are currently looking at six different roles: Tanks, Bruisers, Melee Assassins, Ranged Assassins, Supports, and Healers. This gives us enough bandwidth to classify the vast majority of our Heroes. That said, in Heroes we have several characters who bring something unique to the genre and as a result they don’t easily fit within these conventions. Lost Vikings are a good example. We’ll need to make some concessions in those cases.

Tags

• I have personally been working on the tag system and although it is still very much a work in progress, we can share a few things. We want the system to be easy to understand, but also robust enough to give quality details about how each Hero plays. We currently have about 15 major tags (things like Wave Clear, Burst Damage, Crowd Control, etc.) that can be attached to a Hero. These should help players understand the focus of each Hero while also allowing them to find similar Heroes in their comfort zone or specific qualities they might need in a given draft/match.

We have a few plans regarding Tyrande and Tassadar:

• In the new role system that we’re aiming to release later this year, we have a distinction between “Healers” and “Supports.” The current thinking is that characters like Tyrande and Tassadar would be flagged as Support, and would be treated differently for QM matchmaking purposes then someone like Stukov or Malfurion.

• We’re okay with compositions that utilize a support Hero like Tassadar, Zarya, or Medivh in addition to a dedicated Healer. We understand that most (if not all) team comps will likely require a dedicated Healer regardless of whether or not you take a Support character like the three I just mentioned. We think that teams with some kind of Support-Healer combination could lead to some fun and interesting strategic and tactical gameplay situations, much like team comps where you take both a dedicated Tank and an off-Tank or Bruiser Hero as well. We very much prefer these kinds of team comps to double-healer comps, which we already nerfed last year because they were creating unpleasant gameplay experiences for many people.

• Tyrande’s current play style and talent builds can be really unique and compelling for certain players, but she doesn’t really have a role in the current meta. Internally, we are currently exploring some design changes to Tyrande that would make her viable as a solo healer. Overall the changes have been received very positively by our team, but there is a tradeoff: She is losing a lot of damage and her ability to chunk people from long distances with her owl build. This is something that we recognize can be fun for Tyrande players, so we’re carefully debating if this is the right direction for her. I’m curious to hear any thoughts you all have on this. Obviously, we could go the opposite route for her and remove/reduce her healing mechanic and point her more towards a ranged assassin.

• As I’m sure I will be asked in response to what I wrote above: Why make Tyrande a Healer, why not Assassin? We’re exploring the healing side of things as the rest of Tyrande’s kit can really encourage a Support-y playstyle—she can provide strong scouting for her team using Owls (using them to deal tons of damage works, but it’s a bit odd), her trait debuffs an enemy and helps the rest of your team to set up kills. We could make the trait more selfish, but we like its current iteration that sets up your team and essentially calls out a focus target. For her stun, were cautious to put a hard CC (stun) on the base kit of most ranged assassins. We do break this rule from time to time so it’s kind of soft for either way, but we feel more comfortable putting strong CC on Supports and Healers. This allows them to set up your team do meaningful things outside of purely healing.

• Tassadar is someone we would like to look at more in the future but he’s not someone we’re currently working on. In the past, he could be viewed as a really frustrating character to play against and he was often paired with characters like Genji or Illidan who could essentially “hyper-carry” with his shields. We think the role of ‘enabler’ for Tassadar is pretty unique and fun, especially with someone like a Valla who is a little easier to deal with. Not sure what changes we will make to Tassadar in the future.