We have a full list of questions answered by Blizzard from the latest developer Q/A.
Earlier today, Game Director Alan Dabiri, Lead Game Designer Travis McGeathy, Lead Content Designer Matt Cooper and Lead Systems Designer Joe Piepiora have been answering your questions on r/heroesofthestorm.
In order to have a better system reconnect and replay and spectate, does the game need a “new engine”, or is it otherwise impossible to make those systems significantly better? Ideally you’d want to be able to change the netcode to send some kind of “full state” so that the game doesn’t have to re-simulate every single command (which is what takes so damn long). Is this assessment even close to correct, and how hard is this to change in the current engine?
There is some validity to what you are saying, but I don’t think “impossible” is the right word.
Heroes of the Storm uses a peer-to-peer networking architecture. This architecture offers a lot of benefits, but also makes certain things more difficult. Unfortunately, reconnect is one of those things. Having said that, it is not impossible to make improvements to the system. We’ve actually improved the system on a couple occasions since the original launch of the game. But we’re absolutely with you guys that the system is far from perfect, and could use more love.
This again comes down to prioritization. The same people who would work on improving the reconnect system are the same engineers working on matchmaking and ranked improvements. After looking at the reconnect stats, we decided that matchmaking and ranked system improvements would benefit far more players right now.
The only simple improvement I could see with the current engine is to pause the game while a player is reconnecting, with a timer that shows where he’s at in the game and roughly how fast he’s catching up. A pause is annoying, but I feel it would be an overall plus over playing longer with a bot.
Regarding the pause idea: We do have the ability to pause the game which we use for esports scenarios. So far we’ve decided that this would be too disruptive to use in normal games if a single player had a bad connection/computer.
Will you change matchmaking rules for QM so that the games are more balanced and don’t have 4 assassins vs 3 specialists?
Yes, I actually discuss this in more detail here.
Are you considering allowing to downvote / remove maps, so I can only play the maps I like?
This is something that is brought up occasionally within the team. At the moment our focus is squarely on improving the matchmaking process, and while some sort of map preference system could be interesting it’s not on our radar in the near-term. We believe there are gameplay improvements that can be made for some of our existing battlegrounds in order to make them more enjoyable to play for the majority of our players, so our strategy is to prioritize making all the battlegrounds we have more fun to play rather than giving players more means to opt out of playing them. We have several battleground updates in the works already; Matt discusses them in more detail here.
Are we getting the ability to watch games of our friends live at some point?
The team is very interested in delivering this functionality to players, but the technology that drives our replay system and it’s peer to peer nature does not make this simple to deliver. Compared to other features we are discussing this is pretty low on the priority list.
Are you considering improving the tutorial system?
We’re considering several different methods of better educating our players. I touch on some of these in this comment.
Will you consider a surrender option, especially when team mates are AFK or inactive for a long time?
We don’t plan to include a surrender option to the match experience, and instead are focusing on improving our actioning for AFK or inactive players. We feel that the addition of the surrender option will ultimately yield a net negative social experience for players in the match. It serves as a point of contention during the match where the team cooperation can grind to a halt in the face of deciding whether or not to quit the game.
We do however have plans to include a Loss Forgiveness feature for players who are affected by leavers during a ranked match. I intend to write a blog post sometime in the next two weeks to describe this system in more detail.
Are there any plans to improve the bot AI? Currently the bots cannot help win a game if somebody disconnects.
We’ve actually made great strides with our AI tools over the past year. While an AI-controlled player will never be able to fully replicate a human player, we have improvements planned to help shore up some of our weaker AIs and ensure that they perform their roles more effectively in the match experience. This is an ongoing process that happens in parallel to other development.
Did you ever consider a full MMR reset with all the issues and effects coming with that?
We tend to consider everything when looking at solutions and, yes, we’ve considered what an MMR reset would mean. Man…it’s UGLY. The utopian view is that a reset would be a short period of utter chaos where everyone starts out equal and is essentially tossed into a giant thunderdome where the weak are slaughtered by the strong until everyone is sorted properly.
More realistically, it would be an extended period of utter chaos long after placements as those placement games would be almost completely arbitrary. With no starting MMR to use to match players up, it would be entirely luck-of-the-draw for team comps and where you end up after placements would come down to chance more than anything.
From there, the ranks would have to slowly sort themselves out as the GMs who ended up in silver/gold due to being matched repeatedly with teams full of bronze/silver players dominate those games where the bronze players who found themselves in platinum due to being in games filled with masters end up throwing most of their games as they slowly work their way back down the ranks. In the process, the GMs are inflating the win rate of the low rank players they’re playing with and the bronze players are tanking the win rate of the ones they’re playing with making it more difficult for everyone to end up at their deserved rank.
In short, it would be expected to be an awful experience for everyone.
An API for individual and global statistics has been another topic since forever…
Releasing a public API has been something the dev team has also wanted for a long time. We have an initial version of this API working internally that is used by our HGC and Heroes Game websites. The problem is that it isn’t complete, and it isn’t setup in a way to support the needs of a public facing API (i.e. reliability).
Unfortunately, the same people who would work on finishing this public API are the same people that are working on features like improving matchmaking and the ranked game mode. Because of that, we simply haven’t been able to justify prioritizing it above those features. As of right now, we don’t have engineers working on this, and until we get through a lot of the player facing improvements (matchmaking, ranked, etc), we wouldn’t prioritize this.
Regarding our philosophy of what we’re ok with having in an API, I would say that we’re ok with having almost everything you can see in your own in-game profile. On the match history side, we’d like to include basically every game stat that we track for a match. Everything you see on the scorescreen, and probably a bunch of stuff that we track but don’t show on the score screen. We’ve also ok with even adding “aggregate” stats across the whole player base like hero pick/win rates, talent pick/win rates, etc. Again, this is the philosophy of what we’d like to get in, now it’s just a matter of when we can dedicate resources to it.
Quick Match has been a hot topic in the community lately. What’s your take on it?
Quick Match is still our most popular game mode in Heroes of the Storm by far. We don’t think the right move is to shunt Quick Match aside, we would rather improve upon it. The problem that we see is that team compositions rarely feel balanced or competitive. In the blog post that we posted yesterday we discussed making a shift away focusing on getting players into matches quickly, and instead to focus more on getting players into better quality matches. That is absolutely applicable to Quickmatch and serves as the plan of action here.
The first improvement we’re going to make is to strictly enforce role compositions for the matchmaker. A team will not be constructed without a balanced team composition on its own. This means that assassins or specialists may potentially have longer queue times, while some tank-capable warriors and healers may have shorter ones.
This means: every quickmatch team will, at some point in the future, have at least 1 tank-warrior and at least 1 full healer, with 3 flex picks. Am I reading this right?
In our current thinking we believe that it’d require 1 tank-warrior, 1 healer and 1 assassin, so 2 flex picks. Pending testing and the queues not erupting in molten lava, yes. I think it likely that initially this will be the new standard for the vast majority of QM compositions.
The plan then is to follow up on this change by adding more incentives for queuing as roles that are currently underrepresented in the matchmaking queue which will then make this a hard requirement for every team.
“Teaching the community” is another big topic lately. Do you have thoughts towards that that you can share?
There are some big improvements we can make in the realm of educating our players and preparing them for competitive play. One thing we’re working on currently is a revised Hero Selection system that provides far more detail on things like their playstyle and main mechanics behind different heroes. This is something that is still a little ways out, but it will help players identify weaknesses in their draft composition and potential for synergies and counters.
As we bring the Player-based Matchmaking system back for ranked play we’re considering new ways to contextualize player performance based upon the stats we currently examine. It’s important that we celebrate areas where players have been successful, but even more important to identify opportunities for improvement.
Beyond these things we’re looking for more opportunities to provide actionable feedback to players through the end of match sequence. Right now we’re considering an “XP Missed” stat that tracks missed team opportunities for experience generation. This should help reinforce the importance of laning. There are places here for improvement, and if you have suggestions we’re happy to consider them.
We’ve also been creating additional educational content on places like the Heroes of the Storm esports website for some time now to help expose players to things like pro player insight. Every time a new hero is released now, for instance, we work with an HGC player to create a first impressions build guide that you can find on our HGC website. Here’s the latest one for Fenix, or example: https://esports.heroesofthestorm.com/en-us/news/21659396/fenix-tips-from-yoda. We’re also starting to do more things like Hero update spotlights to keep all our players better informed of incoming changes to heroes’ balance or playstyle.
As a followup to Team League, do you have plans about social features like a Clan System, a Party Finder or similar?
Lots of plans. As you’ll likely hear a lot today, it’s a matter of priorities and how we spend our resources. The plan is to start with an updated party finder which significantly improves that system. The major changes would be to decouple the party finder from chat channels, which would significantly expand the pool of available players, and allow players to look for others based on the game mode and role they want to play.
We feel that’s the important first step. From there, we can build upon that with clans, which would be a great addition to the game as well.
But, again, you can see the list of things we’re working on currently and we feel those items are higher priority than the social features right now. So, while we’d love to get to them, and they’re coming, they’re further out.
Hero Swaps and its alternatives
Unlike a lot of the things we’re discussing here today, hero swaps aren’t a clear win for the majority of the community so while it is something we’re investigating, it’s a lower priority item.
To be effective, swaps would need their own phase, extending the time it takes for drafts to complete. They also open the door to additional toxicity and, although the core of the feature is available in custom games, it’s reliant on the players on the teams trusting each other. For it to come to other draft modes, there’s a significant development effort involved to add a lot of validation between players. You wouldn’t want someone grabbing your first-pick treasure without your permission, for example, which you can do with the implementation in custom games.
On top of that, for swaps to be effective, they require significant upfront communication by the players. That’s certainly something we want to encourage, but that also means the feature wouldn’t be used by a lot of players beyond the top end of ranked play.
So, right now, we’d rather focus our development efforts on features that will more clearly be beneficial to all players.
For the alternatives, such as trying out first-come first-served (FCFS) drafting which provides a similar benefit, we can do that relatively quickly. There’s some skepticism about whether FCFS can work outside of a team environment, though. We had the same concerns when we allowed 2s and 3s in team league and were pleasantly surprised by the results. It’s an area where we want to gauge community interest and if the feeling is that players would prefer FCFS over nothing, we’d try it out in Unranked Draft first.
Could we not just do what League of Legends does and have a “request swap” feature, where both players need to agree to swap rather than giving people the power to steal people’s picks from them? I can see why the current swap feature works in custom games where players are assumed to know and trust each other, but clearly that is not going to work for Hero League. Instead it would make more sense to force people to communicate to optimize their drafts, this would further penalize people who AFK during draft because their team would be worse off compared to a team that communicated and organized themselves around their preferred roles.
I don’t see how people stealing their teammates picks is a concern with swapping, but there appears to be less concern around FCFS, which could just as easily be used to do the same thing. I can fully see people instalocking first picks without any communication.
It would be perfect if we could swap at any point during the entire duration of the draft so draft time wouldn’t even need to be extended since adding another phase to the draft was another concern.
Yes, that’s exactly the kind of validation that has to be put in place before swaps could go to the main ranked modes. It’s just a non-trivial amount of development time that we need to put towards higher priority issues right now, which is why we’re looking at alternatives, like trying out FCFS, that could be done quickly.
As of now, Team League seems to have a lot of issues to the point where barely anyone plays the mode. Do you have thoughts towards that?
This is a particularly difficult question to answer. Ultimately the population for this mode is lower than others due in part to it being difficult to organize a team to play regularly. Many of the ranked improvements we’re making will improve Team League without resolving this underlying concern. We believe that Team League needs some of the upcoming major social features, such as clans and group finder, to truly flourish. Since adding these social features would result in a significant update not only to Team League but to many other areas of the game as well, they’re still a ways out in terms of when we believe we can successfully deliver them. More on social features here.
We originally reduced division requirements for Team construction as a means of increasing Team League participation, which had a small effect on the population of the mode. I don’t believe reintroducing harsher requirements will alleviate the issue, but I’ll take it back to the team to discuss at more length. In the interim if there are other suggestions for Team League I’d like to hear them.
Hero Class Design
Tank (ETC, Muradin, Diablo)
Bruiser (Sonya, DVA,
Healer (Stukov, Malfurion, Uther)
Support (Tassadar, Tyrande, Abathur, Medivh)
Melee Assassin (Thrall, Malthael)
Range Assassin (Valla, LiMing, Zuljin)
Siege/spec (Azmodan, Sgt Hammer, Murky)
Would be nice to get a new role system.
maybe build in guide videos from known youtubers in the launcher like you did with HGC livestreams
an idea to improve quests like in Hearthstone where you have to do tasks to get the reward:
Gather X experience – 300g
Tank X damage – 300g
Heal X amount – 300g
Get X takedowns – 300g
replace “play 3 tank heroes” – “play 2 starcraft heroes”
This helps players to learn the game mechanics.
Information about values, timing and objective-behaviour
Can we get updated data on scaling bosses, camp timings and objective-behaviour like when does a boss use his root and when he casts the circle around him?
Thanks for the question Nyroku_Hots!
In terms of new roles, this is something we want to do. Our current design leaning is to keep it to your first 6 (Tank, Bruiser, Healer, Support, Melee & Ranged Assassins). We’ve always felt that the Specialist category is a little weird. In the best cases, a lot of characters like Nazeebo and Zagaras fit nicely into the Assassin categories. That does leave some oddities, such as Lost Vikings potentially fitting into the “Support” category (they do provide a lot of indirect benefit for your team). Curious what your thoughts are on this, and the rest of the communities thoughts, surrounding placing these odd heroes into more defined categories.
In terms of timeline for this: it is absolutely something we want to do. We think it’s correct for the game moving forward. Currently our priority is around improving the matchmaking and ranked experiences though, so this will be on hold for a little while longer.
For Quests: Many years back, we had a design very similar to this (most likely internal only?). We ultimately moved away from this as it encouraged players to farm out matches they had already won: “I need another 10k healing to finish my quest, don’t end the match!”. We also tried some quests such as “Land X Hooks” but that also encouraged weird player behavior. Ultimately, we stuck with a few simple quests of essentially “Play The Game” but with different requirements. We would love to update the quest system in the future and maybe we can explore some ideas here that could help teach players, as you suggest.
For updated data on bosses, scaling, etc. This is something the balance team has internally. I’ll make a note for us to send out the current numbers either here or somewhere else visible to the community. I will say things like Bosses and other mercenary camps rarely change unless we have a specific design or balance reason that we’re trying to address.
Any words on tags? If you have a minute, please read this. There is a reason many industries (porn, videos, stack overflow) use tags instead of categories. They just work. It is also the reason libs are better than frameworks. Right? It is my opinion that presenting tags would be extremely pedagogical, and, mostly, that matchmaking by tag union would greatly increase the quality of QM 🙂
Tracer: [Mobile] [Sustained-Damage] [Burst-Damage]
Falstad: [Global] [Sustained-Damage] [(extra: Displacement)]
Diablo: [Initiation] [Control] [Bulk]
Dehaka: [Global] [Clear] [Bulk] [(extra: Self-Sustain)]
Chromie: [Burst-Damage] [Poke] [(extra: Stasis, Control)]
Lucio: [Mobile] [Buff] [Heal] [(extra: Displacement)]
Nova: [Stealth] [Burst-Damage]
Genji: [Mobile] [Mobile] [Bullshit]
As part of the role update we will be introducing a tag system similar to what you’re suggesting. This definitely helps provide more context and information beyond what a higher-level role category can provide.
Do you plan to add any system of rewards for good performance? If someone played well, wasn’t toxic, wasn’t afk, there could be an opportunity to reward him somehow. In the current situation some players feel pressure because they can be reported for any mistake they may/or may not make. It could be a good motivation for players to play till the end, reduce the toxicity even when a match is lost.
We’ve definitely had conversations on this. While we’re not currently working on this right now, we’d love to incorporate a system like you mention into our existing set of “tools” to combat toxicity. While we think a commendation/reward system can be a huge help, there are still some players who aren’t motivated by these systems. So we think it is important to have a combination of systems to make sure we can still remove players who simply can’t play nicely in the game.
What do you think about the option to show the number of players in the queue with a chosen role (assassins, warriors, support, specialists)? There could be a bonus reward for those who choose the least popular role.
Yeah, so we’ve already got a bonus for playing “Auto-Select” when the matchmaker is getting into trouble (a new hero launches and everyone wants to play that hero). We’re looking into potentially expanding that to offer more rewards for under represented roles in QM (e.g. Support).
Showing the percentage of players who have selected particular roles is interesting. I’ll take that back to the team and see what they think. Thanks!
The current observer UI is about 3 years old now and has not received nearly any significant changes (2015, 2017). The official UI is used in all HGC broadcasts for more than a year now, and despite that it’s still missing very important features that have been introduced in the three years of its life, like for example showing quest progress, and a lot of smaller stuff that I won’t list because it’s already a long comment.
A year ago we got a great in-game feature with party frames. While I can’t find a source, I’m pretty sure I remember someone from Blizzard saying that the goal is to eventually include the new party frames into the spectator UI, but that’s unfortunately still not the case.
And the lack of updates aside, it’s not even maintained well, with Heroic cooldowns not fitting into their place and covering the talent icon since the start of HGC this year (2017, 2018) and simply not being shown for the last few weeks, which has still not been fixed and will hugely impact at least 4 weeks of HGC unless hotfixed.
So, I’d like to ask about what work are you doing regarding the observer UI and can we expect any updates soon?
Yeah, we agree that our current observer UI is long in the tooth. As you mentioned we’ve wanted to bring stuff like the current in-game party panel into the observer UI. As we’ve gone on, our observer UI has diverged from the normal UI a bit, and it’s part of the reason why you’ve seen some bugs in the observer UI lately. That’s no excuse and we need to fix it, but I just wanted to provide some context.
The current plan is to better unify the in-game UI and the observer UI. We’re going to be using the same party panel UI that appears at the top of an actual game. Once we have that unification, we can start to consider more functionality that would benefit esports viewers.
Has there been a conscious design decision to give recent heroes a lot of tools for their kits (waveclear, mobility, etc) to make them more generalist heroes vs. designing them to excel in very niche situations?
Hey JK_roll, thanks for the question.
The short answer is yes. For a more detailed explanation:
In the past the design team went through a phase where we wanted to make sure every hero had a really sharp role. Internally we discussed this a lot, but we wanted to make sure that there was a unique gameplay reason to pick every hero at some point. As an example, we wanted there to be a unique decision point of why to pick Tychus over Valla, or vice versa, without players simply choosing whichever character has a higher win rate. This is ultimately what led to Tychus’ updated design that he deals bonus % damage on his base kit. We designed his role to be strong against lots of high health targets. We also did a lot of this on our Healers: Malfurion wasn’t supposed to have burst healing as he was designed as a sustained healer, Uther was intended to be a weak sustained healer but strong with burst, and so on. For the healers, we looked at not just their healing mechanics but everything on their kit.
Ultimately we ended up backing off from this. We found that it made the game feel like there was a lot more hard counters and you could win or lose in draft. We still want to have good design and gameplay reasons for drafting each hero in a game, but we’re letting those be a lot softer.
What are your long term plans for support viability? Do you want each support to be viable as a solo support now that you have decided to push away from the dual support meta?
Related to that, there has been discussion about bringing back duo queue now that voice chat has been implemented. What are your thoughts?
You’ve said that Hanamura is coming back soon with big changes. Are there changes coming for maps like Haunted Mines (where the dominant strategy is to ignore the objective entirely) or Blackheart’s Bay (where Xul and Ragnaros double soaking has been incredibly dominant)?
I’ve played a significant amount of Ana since her release (about 150 games in Masters) and I tend to agree with the majority opinion that her rework really missed the mark. She was considered underpowered at nearly all levels of play and her rework gave her a very weak self-heal in exchange for a slew of nerfs. Can you speak to any opinions about the current state of Ana? Would you consider reverting the rework or giving her grenade increased self-heal instead of the Shrike healing?
On the opposite side of the spectrum, the Medivh rework seems to have made Medivh more popular and hated than ever, even seeing a huge popularity spike in competitive play where Medivh is known for being unliked. Do you think mistakes were made with Medivh? Will he see more changes?
You’ve been talking about reworking heroes with frustrating playstyles – what heroes are in queue for changes? Chromie? Diablo? Garrosh? Genji?
Hey LuckyLightning, thanks for the questions. I’ll try to tackle the ones related to gameplay:
Support Viability: I would say that we want every healer to be viable in at least some situations as a solo healer. Some healers will of course be better at preventing sustained damage, burst damage, or even keeping themselves safe. But overall, yes, ideally there isn’t a healer that you feel like you could never pick without another full-on healer.
Map Update: The map team is currently working on revising a number of our existing maps. The big focuses right now include large updates to Hanamura and Gardens of Terror. We also have some smaller updates planned for a few maps but these are more akin to balance changes. Braxis Holdout, as an example, we’re planning some small changes to address how snowbally that map can feel. We would love to make larger changes to Haunted Mines & Blackheart’s Bay, but those are both further off ☹
Ana: The rework was intending to help address the idea that she couldn’t solo heal because she had no easy way to heal herself back up. I know internally we played with a wide range of values on this, and even had a playtest or two where she could outduel everyone in the game because her self-healing was way too high. Ana’s overall win rate is a little on the low side so we have room to bring her up a little bit. I’m hoping we can accomplish what you’re looking for via tuning though, or potentially some slight mechanic tweaks. I will say we love Ana as a healer because of how much skill there goes into doing her healing well.
Medivh: Overall we consider the rework a success. It is interesting to see how much more popular he is now (which is a great thing!) but it does mean that players who feel he is frustrating to play against will see him more often now. With our rework, we did attempt to reduce some of the frustrating aspects of his kit: There is more downtime between using Portals & Force of Will, which ideally gives the enemy team more options for counterplay. This is an area we could easily explore and try to do more with.
Frustrating Heroes: Chromie and Genji are both being looked at internally. Genji has some small stuff that will be coming out really soon (similar to Tracer). Chromie is something we’re evaluating a lot more internally. We’ve done a number of changes such as reducing the range of her Q & W abilities so she would have to be within enemy vision range (you could see it coming and try to dodge), showing the splat for her W ability, and so on. Honestly, it feels a little bit better internally but we might need to do more to really succeed here. Chromie is further off as a result. I don’t think we view Diablo or Garrosh as that frustrating to play against. The recent change to Garrosh where his Q no longer pulls enemies in has helped a lot. In general, every character is going to have some level of bullshit they can bring, but it’s important that you feel like you have counterplay options and ways to outplay your opponent.
Will Gul’dan’s Rain of Destruction be changed or replaced with a different one that can be a considered an alternative to Horrify?
We’re talking about this internally. We do want to do something with this heroic to make it feel less RNG and more competitive with Horrify.
Will Zarya receive some changes?
She’s in an okay spot right now. The balance team might make some small tuning changes but no immediately plans.
Will Artanis receive any changes or buffs in the future?
Similar to Zarya, no current plans.
Are you planning to add M.U.L.E. to other heroes?
We’re leaning closer to removing this from the game as it feels really bad when the enemy team resets your progress. We think this mechanic works decent on Abathur since he really needs those safe forward positions, so decent chance that something like this remains on Abathur.
Hello and thank you to the HotS devs for being willing to have this AMA. I have so many questions I’d like to ask (especially in regards to mobility, as I’m doing a piece on that soon) but I’m going to avoid being selfish and stick to the big one, the one that I feel many of us are the most concerned about: The third ban.
I’ll start with the question before I go too far into my little diatribe here: What made you conclude that the third ban is best served being added to the middle of the draft and not the start? And given the very large number of pros who feel differently (i.e. that we should have two early bans and remove the 3/2 split from drafts in favor of letting both teams draft 3 before the mid-phase ban), are you open to the idea of changing this to better fit what the professional scene feels is more appropriate?
In case you are still on the fence about another early ban, I’d like to offer my thoughts on the subject. And, as I often do when I’m passionate about a topic, I have crafted a spreadsheet to help supplement this discussion with HGC data that I feel will help shed light on why the ban belongs in front. That is located here. I apologize as it is not the prettiest or most easily read thing in the world, but I had to whip it up over night to be ready for this AMA, so hopefully you’ll forgive my tidiness failure. 🙂
This sheet helps showcase how HGC picks and bans have been distributed thus far in NA, EU, and KR in the regular season (up to Week 8). I’ve divided the figures by patch and by map to account for various competitive differences that could lead to more heavy prioritization on “meta” picks. What I found is that the first ban typically only includes about 2-4 “must-bans” on any given patch or map (i.e. heroes that are banned in more than 25% of matches) out of a much larger range (usually 12-24) of total early phase bans.
This suggests, at least to me, that the first ban is already being used very strategically by teams in the HGC to target enemy team hero pools as they should be, and not just to ban out “meta” heroes, as many people are afraid that a second early ban might create. Pro teams are currently using the first ban to target their opponent’s strengths more often than they use it to target “meta” heroes, and “generalist” picks will still be viable as early draft selections. As such, adding a second ban to the start of the draft could provide greater diversity in the overall draft process by allowing teams to influence the draft more at the early stages. If I’m playing HHE and I’m afraid of Khroen’s Hanzo, but also don’t want to give up Dehaka to Mcintyre on a rotation-heavy map, being able to ban both of these is a much more valuable use of two bans than to just ban generic top-tier choices like Malf and Stukov (i.e. the common counterargument to two early bans, a “role choke”). I can force them out of comfort picks and make them adapt to what I want, giving me more strategic value for my bans, not less.
The only heroes that tend to buck this trend of being picked by map or team are heroes like Maiev, i.e. the ones that crop up as “OP” for a time and then fade away as they become less of a balance issue. So most typically, the only heroes which show up frequently in the first ban phase are those which exceeded balance boundaries. And while obviously the goal is to avoid having any of these, it’s always possible that strong heroes manage to slip through. A second early ban provides a “buffer” against this by allowing teams to ban out the “OPs” and still giving them a chance to target their bans based on their opponents or based on the map.
To summarize this before I run too long (though it may be too late for that), I believe that a third ban belongs in front, not at the mid-phase, and that both teams should get to pick 3 heroes before the mid-phase ban. My reasons are thus:
1) Greater diversity in draft by allowing teams to prioritize more potential strengths on a given map or vs a given team
2) Provides an additional “buffer” against balance issues that could spoil draft
3) Greater strategic decision making and influence on the draft
We’ve been discussing adding third bans since around the middle of last year. The idea to place the third ban in the mid-ban phase came from feedback from both the community and the pros. The initial requests for the third ban started coming up because there were now enough heroes in the pool that could fill similar roles that a counter-ban in the mid-ban phase was losing effectiveness. Adding an additional ban at that point would bring teeth back to mid-draft bans.
Taking a step back, the core idea is that the mid-ban is a strategic ban based on how the draft is unfolding, while the first ban is primarily a meta ban. At the highest levels of play, some strategy comes into play during the first ban phase, as HGC teams have done significant research into their opponents, but for most players, the first ban gets used to remove whatever hero tends to be on top of the meta at that moment. In higher ranks, it gets a bit more variety as it will sometimes be the hero that is on top of the meta for that battleground, but that’s not the typical situation.
That being said, sentiment shifts over time, and we’re open to revisiting this and adding the third ban as a first-ban instead. The feeling is that doing so would be mainly addressing a more short-term concern, the release of overpowered heroes, where an additional mid-ban is more interesting long-term. We’ve been watching the community response and are interested to see if folks still feel strongly about adding it as a first-ban after seeing the reasoning behind the mid-ban.
The lack of MMR and rank being directly related to each other can be very helpful to give players a sense of progress, but it can also be an issue, especially on the highest level of play. Being high on the Grandmaster leaderboard often requires playing the most games out of the good players instead of actually being the best (especially this season), and can lead to complaints about matchmaking because matches look imbalanced in regards to rank even if they are fine by MMR.
My question is, are you still considering at least adding visible MMR alongside the current rank system, and is it a bit more than “we’re discussing it internally” which we’ve been getting since the game was released?
Yep, its coming. Over the years, there’s been a lot of back and forth about whether showing MMR is more valuable than harmful. In the end, we’ve come to feel it’ll be more helpful.
To go a bit more in-depth, it’s good to understand the purpose of both MMR and rank. You touch on this, but I want to use this post to also make sure it’s clear to everyone reading this.
Matchmaking Rating (MMR) is purely a measurement. It’s an indication of your skill as a player which is then used by the matchmaker to put together games. The more accurate MMR, the better the matches that can be made. Gaining, or losing, skill is a slow process that happens over the course of many games, so MMR does not change quickly. It purposefully doesn’t react strongly to short streaks of wins or losses as they’re not necessarily an indication of a skill change so much as a streak of good, or bad, luck. This is where issues come up with showing MMR. Once it’s visible, it draws a lot of attention and gets treated like a reward system. Players tend to focus on figuring out how to make it go up by any means necessary, looking for ways to push it higher as quickly as possible. That’s not a healthy outlook and leads to frustration when it doesn’t happen.
On the other side of the coin, Rank is a reward system. It represents how well you’ve done over the course of a season. If you go on a win streak, your rank spikes accordingly.
The issue we run into is that Rank is currently the only visible indication of skill in the game, so it can’t diverge too much from MMR. When players are put together in a match with players of disparate rank, the assumption is that they are also disparate skill, and that feels bad.
This had led to the need for Rank and MMR to stay relatively close, which is why Personal Rank Adjustment exists. That blunts the ability for rank to be a true reward system. It also means that matchmaking must consider rank, in addition to MMR, when putting together teams. Visible MMR would allow us to get rid of both of those, which we feel is more beneficial than the downsides that are likely to come about with visible MMR.
We’re working out the details currently, but the plan right now is to add it as part of the updates we’re doing for Performance-based Matchmaking, where we’re already working on new ways to provide you with more—and more clear—types of information on how you performed in a game. Joe went into more detail on that here.
Would you be willing to put together a roadmap similar to what the Destiny team has put together to solve similar community related issues and concerns? https://www.bungie.net/en/News/Article/46735
This way you could effectively and consistently show timelines for the re-release of PBMMR, Rank/MMR decay, and some of other Ranked announcements from yesterday.
So we definitely want to communicate with you more about what we’re currently working on, and what our view of upcoming priorities are. The blog post and this AMA are some steps in that direction. We’re also discussing lots of ideas about how best to communicate with the community going forward.
That being said, here’s some info for you in terms of timing/order of stuff:
New toxicity report validation system: We should be turning this on within the next couple weeks. This will increase the amount of reports we action by a significant amount. We’re going to put out a dedicated post on this soon.
Matchmaking improvements: Since the majority of these are on the server side, we’re going to be rolling these out independently of our normal client patches. Some will come quicker than others. For example, we’re going to try and get the hardening of rules to favor match quality more over queue time in the near term.
Loss Forgiveness / Improved Leaver Punishments: This is deep into implementation and is currently lined up to come out in a patch in the next couple months.
Ranked improvements: Several of these improvements make the most sense to release with a new season roll. So, we’re trying to get some of these changes into the next season roll (~June). 3rd Ban is high on the list.
Performance Based Matchmaking: We’re wrapping up the improvements we discussed in the blog post, and are planning to re-introduce this system also with the next season roll. We may initially introduce this as a purely informational system, and then turn on the actual point adjustments after the community has had time to provide feedback. If we do that, it technically doesn’t need to align exactly with a season roll.
I’d like to take this opportunity to ask a few questions regarding your stance on Vision Control in the Nexus.
In terms of power level, how do you compare vision-enhancing talents and spells to other forms of utility, such as mobility or defensive cooldowns?
Should vision spells always be tied to other effects, such as how Hanzo’s Sonic Arrow can deal damage and be talented to stun, or how Chromie’s reveal talent is also tied to a quest?
What are your thoughts on vision spells that don’t alert players when they’ve been spotted, like a Wisp hidden inside a bush that players walk past without checking?
Thanks for the question Beg_For_Mercy.
Vision is an interesting part of the game that we don’t really augment too often. We do some bonuses to vision that you mention, and have some heroes (Dehaka) who can also limit enemy vision. Overall, I think it’s a super interesting mechanic that can add fun and strategic depth when designed correctly.
In terms of power level, we typically start internally as a ‘feel’ thing, and then eventually we can use player data to validate those decisions. Many of the talents players are choosing from can offer a lot of sheer power. It’s hard to make a vision talent compelling to choose over 25% more damage, without at the same time making that vision talent feel unfair or unfun to play against. This is typically why we will add secondary bonuses to these abilities or talents. Vision can also be considered situational, so by adding damage as an example, we’re guaranteeing some baseline power on the ability/talent.
In general, I think we want vision abilities to alert the enemy that they have been spotted in some way. As an example, you can visually see Zagara’s creep, so there is likely to be a Creep Tumor near by. For Lunara’s Wisp, this is an interesting one. I think we could alert you that a Wisp has spotted you, but it does lose some of its power as a result. We are working on a Lunara update that has some changes to her Wisp and related talents, so I’m looking forward to seeing how that plays out. We can revisit any of these rules though!
I’d love to see some thoughts on this. I feel like sometimes, the design guys give a character a kit they feel is lore-appropriate and unique, and then toss it to the balance guys for number tweaking without them having any input on the kit design. Vision and mobility are the two big standouts -> it breaks the game when some heroes have vision as a core part of their identity and then a new hero gets it as just some sort of flavorful thing they can do.
Just to touch on this quickly: We absolutely involve a lot of the design team as early into the process as we can. The live/balance designers are paired with a hero designer from the beginning. One thing the balance designers will try to surface early on in the hero design process is if they feel they have tuning knobs for each talent and ability. In the past we might design a talent that simply increases charges from 2->3 on an ability. If this is too strong, we don’t have a lot of tuning knobs without adding a kiss/curse mechanic or something else.
Some summoned units are quite frustrating at low MMRs, but not that useful at higher MMRs (Raynor’s Banshees, Kerrigan’s Ultralisk, Malfurion’s treants, …). Is it something the team is looking at?
We’re not currently looking at summons as a whole but will likely try and tackle this on a case by case basis. An ability that is really high impact (and frustrating) at lower levels of play but not useful for higher tiers of play is a bit of a red flag for us. That being said, this is kind of the nature of a summoned unit. In some ways, you can think of a summoned unit like an ability with a huge variance in its damage potential, and this largely depends on how the enemy team plays against it. Obviously higher level players can downplay the value of these summoned units much more easily.
The question: how about Raynor’s rework? (really, just putting his level 7s on the W and E, and making marksman baseline would already be very appreciated)
It’s not the next set of reworks that will be coming out, but we are actively working on it. Overall we wanted to keep the feel and playstyle of Raynor, keeping him simple to play, but we’ve made a number of changes to modernize his base abilities and introduce a lot of new talents. The team is super excited to share more with you guys on this soon (next few months).
Are you overall happy with Hammer’s rework, even though some claim it’s a bit bland and could offer less options?
I think we added some fun new stuff to Hammer’s kit, allowing her to access some of her key defining talents earlier on. We haven’t seen a ton of Hammer play, but she’s also a character that probably shouldn’t be core to the meta. We don’t currently have more plans for Hammer but we will likely revisit her again in the future.
Do you plan on changing ranked seasons, so they have the same durations/times as HGC seasons: Two six-month seasons per year? Do you plan on adding a team system, for named teams (in team leagues for example)?
Regarding season length: We’ve discussed season length quite a bit on the team. We’ve actually considered everything from once a month to one season for the whole year. We think the current system is a balance between those, but I would say that we’re still open to discussion on this topic.
Regarding named teams: We actually had this in Heroes originally! 🙂 Team league worked this way when it was initially launched. The issue was that problems we currently face with Team League (which I think one of the other guys answered a question on that already), are actually worse with the fixed named teams. The tl;dr is that it’s tough for players to align schedules with their teammates. As a result, we saw even less participation in that system than we see with the current system.
Is it possible we’ll get a second look at older cleave talents, or that modern/incoming cleaves will be toned down in power?
Great question hailcrest.
I agree that a lot of those recent examples have very strong AOE. In general, we try and look at the design of the kit and balance of the hero kit overall. Every hero will have some obvious strengths and weaknesses, but are hopefully balanced within the larger overall picture. For the Junkrat example specifically, I agree that his wave clear is insanely strong! Hopefully however his overall power is about in line with what makes sense. We of course have the ability to tune this down in a number of ways including the splash size, and the fallout of the damage (i.e. less damage further from the center).
I will say we will absolutely revisit old heroes and even these newer ones and make adjustments where needed. Nova has never been intended to have strong wave clear, so I’m not sure we will make changes there. It may be possible that we need to tone down some of the newer heroes wave clear, bring some of the older heroes up (Kael’thas in this example), or meet somewhere in the middle.
Can we expect a specialist balance pass similar to what you did with the warriors anytime soon?
The specialists as a whole have been struggling for a while in the meta and ended up being relegated mostly at low ranks, with only a few being competitively viable. If you exclude the pseudo supports the number of viable specialists is even fewer.
On that topic, can we also expect a role rework anytime soon?
For Specialists: Most of them actually tend to have higher win rates, so in terms of that, they’re not currently struggling. As part of the pass we’re working on to reduce frustrating moments with various heroes, there are a number of Specialists who are on that list. Usually its surrounding when they’re split pushing and it feels like the enemy team has to constantly respond to them.
For the role rework, this was covered in another part of the thread. TLDR is we want to do it, but its lower priority then some of the ranked and matchmaking stuff we’re currently working on. Definitely dig up that other reply though as there’s more detail!