how does something like this still happen? just like when ludens came out and you could buy it in dominion, but no NLR

To answer specifically how this particular bug happened, we did catch it before release and fixed it, but a merge accidentally overwrote the fix. That kind of thing is not uncommon when you are trying to make changes last minute. We will fix it quickly.

Csgo uses something like dynamic Q in league, not sure if Dota is the same. Do you have any idea why people hate dynamic Q in league despite the fact that many games have the same system? Or is it just nostalgia of solo Q?

Ranked in League was largely solo for many years. I think calling it nostalgia undersells it a bit, because that implies folks are just pining for something they remember fondly. I believe that players who want ranked to remain solo prioritize the expression of individual skill above almost anything else. It's a very philosophical consideration, and I don't mean that in the sense that it's esoteric or academic, but that it really comes down to personal values. That is why, IMO, players who just want solo aren't swayed by the fact that they rarely encounter premades or that most players still queue solo. That's not the point. The point comes down to personally held opinions on what the point of ranked play should be.

I don't think that's a bizarre point of view at all, given that is exactly how ranked played until this season. If we had launched League with Dynamic Queue, I suspect you wouldn't see the same amount of controversy. That is probably the primary risk with making such a dramatic change to a live game.

While we do think Dynamic Queue is the future, I am very empathetic with players who don't like that future.

Kudos to the Playtest team that does the Q&A sessions. They show backbone. I like it when morons get called on their bullshit.

I'm a huge fan of that team. We had a playtest team for WoW, and they helped us out a lot, so I am glad to see the League one start to pay dividends.

Obviously you have to be careful not to talk down to players, even when they do lack information or just have a missed perception of things. So I would personally never use terminology like "morons getting called on their bullshit." But part of having very highly skilled players on the playtest team is that they can't speak from a position of authority and experience.

A lot of people say Riot is a bit lazy/slow and that games like Overwatch will keep you guys on your toes and make you work harder to compete. But I've noticed a significant increase in content quality lately. Do you personally feel you guys have slacked a little or that you've always tried hard?

I think it's awesome overall when great games come out and keep us on our toes. When good games come out and keep one-upping each other, it's players who ultimately benefit, which is a great outcome.

Blizzard has a really veteran team (I know that first hand!) and can pivot pretty quickly when they decide to change direction, which is something Riot can get better at for sure.

Granted I have only been here for about 3 years, but I do believe Riot always tries hard. The folks I work with do care very deeply about whether or not players are happy. When we are slow at times, and we are, I don't think it ever comes from a place of laziness.

So where does it come from? Here are a few things I see:

1) When you have a big hit, it's easy to become risk adverse. At some point the chance of losing players from any decision, big or little, can be almost paralyzing. To be clear: that's not a healthy attitude to have, but it is an understandable one. We have to keep reminding ourselves to be bold and take risks.

2) Riot is the opposite of hierarchical. My ability to get things changed in the game comes because I can articulate my point of view than because I issue orders. That is true of anyone who works here. Theoretically that should make us fast, because you don't have a lot of gatekeepers whose blessings you have to get to make a certain change. But the ambiguity can also hurt a team's ability to be decisive. Since there are no clear gatekeepers, sometimes folks will just ask a dozen different people, sort of trying to recreate gatekeepers organically. This is a problem we have acknowledged and are working to fix.

3) League (and Riot!) grew very quickly when the game exploded in popularity, which made it easy to prioritize fast solutions over sustainable ones. We have had to spend a lot of the past few years paying down this technical debt. The good news is we have made massive strides. Incidents of "spaghetti code" causing unrelated systems to break have dropped enormously. The new client we are working on will be a major step along this journey.

4) Philosophically, we can take a really long view, because we want the game and the company to last for decades. We also have a lot of young people for whom this may be their first real job, let alone their first job in the industry. To develop into a world-class team, we try to give those individuals a lot of room to make mistakes, because that's really the best way to get better at something. If I was an art lead, I could "save" a junior artist from making a mistake, and maybe that would result in getting her feature out sooner, but that comes at the expense of my bandwidth, which doesn't scale, and isn't helping to level her up as an artist. We're willing to pay some of that cost in the short term for long term benefits, so long as players don't suffer too much as part of that growing process.

Great question, and one dear to my heart.

Do you suppose any assassins will be reclassified as skirmishers or divers after their reworks? As an example, Diana doesn't fit the criteria of an assassin because she has no escapes.

That's possible for sure, but I think the champs we are looking at touching for the assassin update are all going to stay assassins.

I think you can make the case Diana is a Diver.

Hey Ghostcrawler, I used to love watching pros stream on Twitch and it encouraged me to want to be better at the game. Now with DQ, pros don't really care about ranked and only play seriously on the Tournament realm. Isn't this a good indicator that although DQ is fun for some, it isn't working?

Pros have always spent a lot of time on the tournament realm doing scrims. I'm also not sure the pros really took ranked that seriously. (I'm sure some did -- I'm generalizing.) They do take streaming seriously, and that seems like the problem to look at if any.

I know Oceania in particular has gotten special attention on this topic, but we've always had challenges with high-end Oceania queues because the population size is small. The quality of the matchmaker can only do so much if there just aren't people out there who want to play a game with me.

So, I think there are some problems here as you mention. I don't think the conclusion to draw is that DQ caused or even exacerbated these problems.

What is counterplay?

It has to do with agency. No matter what an enemy team or player does, you should have some ability to deal with it. A lack of counterplay could mean that if the other guys makes a great champion choice at champ select, that there is nothing you can do -- you essentially lost at champ select. It could also mean something like an assassin that gets the drop on you, stuns you, and then takes you to zero.

Counterplay doesn't mean that you should be able to shut down anything that comes at you. But it does mean that you should be able to act.

You should borrow a page from Cactopus. That dude is a God of PR.

Yeah, he's doing awesome.

I fight with my friends almost daily because they want us to play as 5 but are too lazy to join discord/ts/skype and i dont wanna play like that because the enemies always stay grouped and wreck us. A big part of the blame is yours.

Because we don't have a default voice option? Just trying to understand.

I decided to intentionally throw every single game if I have premades in my team. I don't mind the permaban as long as I can ruin premade players' game. Am I a bad person? :)

Well, you are going out of your way to ruin the game for 9 other players as a way of trying to punish us, I guess. I don't feel qualified to call you a bad person or not, but those other players are pretty blameless.

the problem is that ranked doesn't mean anything. it's not a gameplay problem. it's a motivation/interest/philosophical problem. there is a type of fun that focuses on competition and pvp. you removed the part of the game that caters to them. they don't have anywhere to play. so they leave.

I agree with what you are saying. There are players who feel like if premades can exist in any form, even if they are rare and even if they face matchmaking penalties, that it invalidates ranked completely as a measurement of individual skill.

I personally disagree with that invalidation, because individual skill still has an enormous and probably the largest influence on where you land in rank, but I understand that some players feel that way. So far we haven't seen too many of them quit, and I certainly hope they don't, and hope that we can find some systems that recognize their accomplishments and dedication as solo player while still allowing League to be a team-based game.

It can't be that rare for solo players to be matched up against large groups. Every time there's a 4 man premade in the queue a solo or a duo HAVE to be matched VS a 4 or 5 man premade. Robert Johnson

Correct, but statistically speaking, there just aren't that many players queuing as a group of 4 or 5. We saw 2s, 3s, 4s and 5s all increase with DQ (and 3s probably saw the largest percentage increase), but they were very rare to begin with, so even increasing them dramatically doesn't mean solo players will encounter them frequently.

Even when that does happen, the matchmaker can usually find another premade (or a group of smaller premades) to match against the premade. The situation of 5 solo players being matched against a premade of 5 is very unlikely.

Remember, the whole intent behind DQ was to make it more reasonable to form premades. They were very rare, even though players told us they wanted to use them and had more fun when they did. The intent behind DQ was never to try and convert all the solo players to group players. We just wanted to even the odds a little. We seemed to have done that, but only a little.

Where is Riot even getting information that a majority of players like DQ? Albeit I can not speak for other regions, but all I see on the boards are anti DQ posts.

Boards and Reddit aren't a great quantitative indicator of sentiment because it's generally the tendency for players to go to the boards when they are upset about something. You'll see some positive posts on almost any feature, but they tend to be in a minority for almost any feature, and players who are indifferent almost never go to forums to state that they are indifferent, which further tends to make almost any topic seem more polarizing than it might really be.

And that's fine, because that's not really what boards are good at. We can survey large populations of players in other ways. What the boards allow you to is to articulate an argument, taking almost as much time and room as you need, and then see if it resonates with other players.

I am simplifying pretty dramatically here, so please forgive me, but I generally see 3 categories of players who don't like DQ:

1) High elo players who have trouble finding matches in a reasonable time and any player who struggles with getting their primary position. (This tends not to apply to actual pros, who have access to scrims). We have made a bunch of changes to try to make this problem better, and it has gotten better, and we will continue to do more until it feels acceptable.

2) Players of any elo who "feel" like they are facing a lot of premades, or blame a bunch of their losses on speculation that the other team is a premade. Statistically speaking, the chance of this happening is pretty low. No doubt it does sometimes happen, and I can't blame you for feeling shitty when it does happen. We are adjusting the matchmaking modifier for premades to make sure their winrates aren't too inflated.

3) Players who feel like ranked as a true measure of skill is compromised by premades even existing. This begs the question that duos have existed for a long time, and presumably the players who aren't happy with DQ either weren't happy with the duo situation, or are just worried having 3s-5s will make it worse. Nothing short of reverting DQ is likely to truly address this concern, so there may be limits to what we can do here. We are working on better forms of solo recognition so that players do have a way to demonstrate that they primarily queue solo, if that's important to them.

Has anyone at Riot ever thought of this super simple solution to the problem with DQ I have? It goes like this, make a pure solo queue and then limit dynamic queue to groups of 2, 3, or 5 players. Its the best way in my opinion to satisfy as may players as possible. Losing 4-man groups is worth it OmniFlare

All of our estimations suggest that without solo players, there would be ridiculously long queues for the groups queuing as 2s and 3s. The likely outcome of that would be that the premades just decide to split up so that they could get in a game in a reasonable amount of time, further increasing queue times for those players remaining in DQ.

If you were born in Runeterra, where would you live and what would you be doing?

I would own a bait shop in Bilgewater.

Have you talked about making a more strategic champion? The sort that sends reinforcements to lanes, changes how the positions on the map functions, "global support"-abilitets, etc.

Yeah, stuff like that is intriguing and we'd love to do it sometime, but we would not want to celebrate something novel at the expense of the game overall. For example, being able to control a lane is a pretty core part of League mastery and we wouldn't want too many ways to opt out of it. Counterplay is really difficult when the guy casting the spells is far off of your screen. And so on.

When hiring a game designer at Riot, what are the core skills you are looking for? Also, how important is prior scripting knowledge?

Core skills are things like strong design analysis, good game taste, critical thinking, the ability to pitch multiple solutions and argue for which is the strongest, intuition about when to use data and when to use intuition, lots of games played, the ability to carry something through to completion, and stuff like that. There are not many specific skill requirements of the "must have working knowledge of X" variety.

Previous scripting experience is indeed helpful. Nearly all designers at Riot (and Blizzard for what it's worth) spend at least part of their time writing scripts to complete their projects. If you have any computer programming experience, scripting should be a breeze. If you have no experience, it's not that hard to pick up. You can get pretty far by copying a script that sort of does what you want and then modifying. Pro tip: just don't start with the Orianna script.

Do you have a response about all the talk about cloud drake that's been going on?

We aren't ready to give up on Cloud Drake. We'll likely buff what's there now before talking about alternative approaches.

i just wanted to say that perhaps the best way to explain what is wrong with dq is that your ranked ladder has lost perceived value. Challenger, masters, diamond and plat etc... when ppl tell you that they have said rank, you think, ehh probably boosted, not, wow, hes good.

The emblems should help with that. If you see a Diamond with the solo emblem, you know that player got there completely on their own.

I'm not naive enough to claim that boosting doesn't happen, but my suspicion is that most players queuing as premades just want to play with folks they know. It's kind of like "feeding." Not every time someone gives up First Blood are they intentionally feeding. :)

Emblems dont solve the fundamental problem of mixing solo players with grouped ones. I cant understand how you guys dont get something as simple as that. Already over half a season ruined by your stubborness, how much time are you guys going to waste resources on this? It isnt working

To be clear, you could always group with one other player. While you can group now with 3-5, it is rare for most players to be matched against a large premade.

I can understand how philosophically you might not be behind that direction. We know that there are definitely players who are not. But I have a harder time understanding how the season might be "ruined" or how it "isn't working." How do these problems manifest themselves in your games?

Any thought on Riot hosting a monthly/weekly discussion over various topics? Like state of marksmen, pick/ban phase issues in pro/regular play, or jungle pathing in different situations. Stuff that has depth and could create a discussion throughout the day with Riot/pros chiming in with extra data.

That's an interesting idea. Sort of focusing the conversation around one (hopefully broad) topic. We'd have to make sure players understood that players were going to have to be most of the conversation and not turn it into a Rioter Q&A.

Is it safe to say now, in hindsight, that dynamic Q was at least rushed? I remember a podcast with Lyte saying how he had the goal of revamping it all within 1 year and everyone was all like, "I don't think you'll be able to."

I would describe it this way: the development on the feature wasn't rushed, but at the very end we ran into some problems that took time to fix before we could go live. Our original timetable had DQ live for a long time before the ranked season started, but with the bugs, we burned through all of that time and were ultimately faced with the the hard choice of turning on DQ, keeping it off, or delaying the ranked season. We turned it on, but given the number of problems it still had, that was the wrong choice when viewed through 20/20 hindsight.

One of the great things about working on League is that there are very few hard deadlines. We patch frequently, so you don't run the risk of missing the box product release or even missing an expansion if your feature slips. Ranked season starting and esports events are the two big cases where you do have hard deadlines.

Solo players get screwed over the most with dynamic Q. An emblem doesn't fix the fact that competitive integrity is non-existent for them and many are struggling to climb elo (See Valrkin who has been a challenger for several seasons but fell as a solo player this season due to premades).

There are definitely some very high elo players who were hurt by the system. No question. We are fixing a lot of those problems, though we aren't done yet.

For the vast majority of players though, facing premades has very little to do with their ability to climb the ladder or not. They always had the chance of facing duos. Now they have a chance to face 3-5s, but it is a very, very small chance, and we do still apply a modifier to those solos to make up the potential winrate differential.

The idea behind emblems isn't to fix "can't climb" problems though. It was to recognize players who prefer to play solo in a way that other players could actually see.

If you could go to any country where would you go? What countries have you already been to?

Iran. I hope the political situation between that country and the US stabilizes at some point to where that is possible.

I've been to a lot of places in the Americas and Europe, fewer in Asia and none in Africa. At least half of that is work-related, for which I am very thankful.

But the emblem solution doesn't give solo players what they want, which is some sort of solo recognition that is only available to them. Not available in a ever so slightly tweaked form to everyone else. Exclusive. Just as ranked felt. Just like DynamicQ gets exclusive IP weekend and Hextech boosts

Are you saying you'd like the emblem if there were just a solo emblem but nothing to recognize a premade emblem? Like being exclusive is the one key to granting it value?

Or are you saying it requires a more tangible reward, like IP or boosts.

Remember, our original intent here was to provide more of an ability to queue with friends, to let premades make up some of the huge gap between even players who duo and players who solo (and I'm talking explicitly about players who want to do premades, not someone who feels like they have to). If we shift too many rewards to solo, then we just shift some of those players who now form premades back into solo again, and we potentially lose many of the advancements we've made.

The key is to recognize solo accomplishment, without encouraging players who are enjoying premades to go back to solo just to earn a reward.


Ask @RiotGhostcrawler:

About Greg Street:

Design Director, League of Legends @ Riot Games Before that: Lead Systems Designer, World of Warcraft.
Before that: Lead Designer, Age of Empires.
Before that I was an oceanographer.

Los Angeles, CA