League Of Legends’ Greg ‘Ghostcrawler’ Street Talks Balance – Is It Even Possible?

League Of Legends boasts a huge roster of playable characters which bring the potential for vast numbers of potential compositions, ability interactions. It also has a player base quick to seek out what’s strong at any given time in the game’s ongoing development. So how do you approach that as a designer? Is the idea of perfect balance even possible or do you tweak characters so that batches rise and fall in relative popularity? Michael got in touch with Riot Games’ design director Greg ‘Ghostcrawler’ Street to get a clearer feel for the studio’s philosophy when it come to their champions:

One of the criticisms often levelled at League of Legends is that it’s not a balanced game, that the design team and the Champion creation team seem focused on creating a rotating meta instead of working towards an end-game of perfect equilibrium. This opinion largely stems from Riot’s new policy of patches that focus on a specific set of champions, adding new mechanics and giving them targeted buffs that can upset or heavily impact the meta of the game.

Prior to last season’s World Championship, a patch arrived called the ‘Juggernaut’ patch, where several tanky heroes who had fallen from favour or had never been competitively viable received a lot of attention. A number of fans felt these changes went too far, resulting in a World Championship tournament that bore little resemblance to the season which had preceded it. Riot have since introduced a Marksman update, changing the meta in favour of AD carry champions, again shifting the balance of power and resulting in a radically different way of playing the game. This summer will see the Mage update – are these seasonal changes what we can expect from League of Legends going forward?

We spoke to Greg Street, Riot Games’ design director, to try and understand the long-term goals for the balance and design teams and also to ask for a response to some of the accusations leveled at the game about its perceived lack of balance. Make no mistake, this summer’s Mage update is a big one for Riot and fans will use it as a marker to see if Riot can live up their lofty ideals. If not, they’ll almost certainly face further criticism.

Michael Johnson: In a previous AMA one of the ideas that came up was the notion that Riot is trying to enforce a ‘rotational meta,’ and it was something both Meaghan ‘TimeWizard’ Bowe and Brian ‘FeralPony’ Feeney rejected. I’d like to know firstly what was meant by a rotational meta from your perspective?

Greg Street: Philosophically, we try to make changes that improve the game. We always want to move forwards. At times, that does mean we have to take one step backwards for the proverbial two steps forward, but what we don’t want to be doing is just shifting the game from one state to another and back again. When Time Wizard said that it was someone’s “time to shine” or something similar [I believe it’s this section – Ed.], I think what she meant was that when we neglect a champion who is underpowered or overpowered for too long, we feel a lot of pressure to act.

What we try to avoid is proclaiming that it’s now the tanks’ turn to be overpowered and next it will be the assassins’ turn to be dominant, or whatever. While you can artificially force champion diversity that way, in the long term it comes at a pretty heavy cost. Players learn that the time they invest in experimenting and learning new champions or strategies isn’t worth it, because at the end of the day, Riot will step in and reshuffle the board.

Put another way, if playing League is like running an obstacle course, it’s okay to change or replace obstacles that are too easy or too hard to surmount. On the other hand, arbitrarily swapping the placement of obstacles in the intent of catching players off guard, or having someone unexpectedly shake the net they are trying to climb feels like a cheap way of making the obstacle course harder. Because those interferences I just described are so random as a player you don’t feel like you can get better at the obstacle course. The heart of League is how satisfying it feels to improve at the game, and artificially making things harder works against that mastery process.

I tell the designers to imagine, as a mental construct, that there is a perfect state of League of Legends out there. Realistically we’ll probably never reach it in our lifetimes, but we should strive to make changes that move us towards that goal, not just make sideways changes for the sake of change.

The follow up to that is that we’ve seen the Juggernaut champions kind of fade away at the top levels of play and replaced with the champions from the Marksmen update (as well as the counters to them i.e. Malphite [note, a more current example would be Nautilus]), is this what we can expect to be the new status quo in the game? Will the Mage update bring about another seismic shift in which champions are viable at the expense of the Marksmen, or is it difficult to judge the effect of these big changes?

GS: I think in many ways the Marksman update was more successful than the Juggernaut update, so that leaves us with more viable Marksman champs that didn’t exist before. I also think players get excited about seeing changes (especially when they’re changes that add depth) so it’s not surprising players would want to try out updated champions. We will also take blame for some of them just being overpowered as part of the update. So following that, I expect you’ll see a lot more mages being played if the mage update is successful, but I hope that doesn’t mean Marksmen no longer feel fun or viable.

Is the idea of ‘balance’ as this sort of thing where every champions is viable/unique still even possible with a game like LoL with such a huge champion roster? Or is it more about (as Time Wizard stated in the AMA) making sure that each champion has time to shine at one point or another”?

GS: Like I said above, I think it is possible in a theoretical sense to achieve perfect balance, but it is so very, very hard to do that we don’t set out with the expectation of ever actually achieving perfection. But keeping that framework in mind hopefully does guide the team to make *good* changes and not just changes. A good change adds depth and replayability and diversity, and is done with a purpose in mind.

We aren’t looking for a 50% winrate and play rate for every champion, because some champs may always be easier or harder, or more fun to play, or more niche experiences. That’s fine. But we don’t want players to feel like raw power is holding them back, either forcing them to pick a champion they don’t really want to play, or denying them a champion they really do enjoy.

Finally, for the long-term health of League of Legends, do you think it’s more important to change things up every now and then, with different champions being powerful and popular to keep the game varied and fresh for players/viewers, or to aim long-term for the idea of overall balance? Or is it a middle-ground between the two?

GS: Evolution is one of our design values. League is always going to grow and change, and players who are looking for a game state that calcifies aren’t going to find it with League. While we don’t want to make arbitrary changes, we do want to respond to problems in the game, so for example – we don’t view long patch notes as a scary thing to be avoided.

Now, we do describe League as a curated experience. It’s not some kind of wild animal that mutates on its own without direction: it’s something that Riot takes a hand in guiding along the way. When we say that we want to keep the game fresh, what we mean is that just when players start to get bored, maybe there is something new coming around the bend.

I’ll hit a couple of examples. We thought that top lane felt a little too isolated, and while we know that’s what some top laners enjoy about the experience, we want the actions in top lane to still have strategic relevance for the rest of the team. We added the Rift Herald to give a strategic objective (one that can change the fate of the entire team) to solve this problem. It’s tuning isn’t perfect yet, but we think the idea is solid.

Masteries was one of those systems that hadn’t aged well. Through millions of games, players had largely solved on their own which were the smart choices and which were the poor choices. Almost all players just looked up accepted builds and ignored the system after that. A system intended to provide some amount of player engagement was completely failing on that front. So we recently changed the content of specific masteries, as well as the structure of the trees themselves. That’s not change for the sake of change – it’s change to add depth where the depth had all but dried up. And what’s more, we know that this mastery overhaul isn’t going to provide enough depth to stand the test of time, so the changes we made at the end of 2015 were done as a short term fix knowing that we have an even more ambitious overhaul in mind that we’d love to deliver in 2016 or 2017.

The same arguments can be made for champions. If assassins are perceived as being too weak, unreliable, or poor strategic choices, then players are less scared of being caught out in the open alone. If specific champions are so weak or shallow that they are never seriously picked, then it kind of defeats the purpose of having 130+ champions, and we can’t count on new champions to offer a new set of challenges to master.

Using the obstacle course analogy again, there’s no point having 12 obstacles in the course if everyone knows of a shortcut after obstacle #2 that skips everything else. League players are smart, and there are a lot of them, and they play a lot of games, so it’s pretty easy for players to find short-circuits like this. As a player, it’s fun to find these short-circuits. But as they do, we try to respond as quickly as we can, without anyone feeling slapped on the wrist for trying to find new strategies to make themselves more effective.

Thank you for your time, Greg!

Sponsored links by Taboola

More from the web

From this site


  1. Faxanadu says:

    He looks so much worse without his cigar. Oh well. I suppose we should be glad there’s still butts in the game. Hope that doesn’t get Overwatched. What? I’m just being offended. I hear it’s all the rage these days.

  2. Chillicothe says:

    Im sure balance will just be fine once the heroes eventually all play the same 5 or 6 ways…

  3. pillot says:

    that ekko tho

  4. po says:

    As someone who was a regular poster on WoW’s forums, in many of the threads complaining about that game’s balance, during Ghostcrawler’s tenure at Blizzard, it’s strange to see him talking about game balance now that he’s working for a new developer.

    In WoW players were generally given the impression that the developers simply couldn’t be bothered with balance, with the term FOTM being thrown around liberally.

    However, since gaining a bit more of an understanding of how games are developed (as in, learning about things like release schedules, and Scrum/Agile development), I can now see that perhaps Blizzard’s failure to make more than token gestures at balancing the classes in their game, may actually have more to do with the corporate culture, and some developer’s mention of the company’s problem with ‘patch paralysis’.

    Blizzard likes to do big, irregular expansions, with very few patches inbetween. That makes it very hard to make balance changes, because they need to be done constantly, in small increments, so that the effects can be seen and further balances changes made.

    GW2 is a fine example of a game that was balancing in this manner, with regular 2 week patches, that contained mostly class balancing for PvP (2 weeks also being the usual length of a development sprint when using Scrum).

    Then we get down to the issues of trying to balance classes in a game that has both PvP and PvE elements, in the same open world (like in WoW). That gives the problem that you may achieve PvP balance, but then you might end up completely breaking PvE balance (and Blizzard’s bias among their studio’s developers is pretty much well known, even if they won’t admit it, thanks to that whole ‘Shut up PvP guy’ thing).

    Again, GW2 is an example of a game that doesn’t have a problem with this, as its PvE and PvP are split, so that class abilities can have different effects between the two, allowing them to be balanced separately.

    Personally, I think the way to go about class/hero balance is to start by creating a baseline to work from; selecting a hero that is very popular, fun to play, but not perceived as being overpowered.

    You aren’t going to be making changes to your baseline unless absolutely necessary, so you polish it to a much higher degree than the other classes, to make sure it stays popular due to its fun-factor.

    From there you can proceed with ‘tiericide’ (as CCP – EVE Online’s developers – put it), buffing and nerfing classes that are under or overpowered to bring them as close to your baseline as resources permit, the whole point being to avoid the occurrence of outliers, give players maximum choice, in essence making sure the content you’ve taken the time to produce isn’t being obsoleted.

    With the changes being an ongoing process, (largely due to the constant addition of new classes), the meta will still constantly be shifting, giving players a constant supply of novelty to keep them playing, but you’ll avoid that whole FOTM (or rather many months) problem that was so prevalent in WoW.

    So long as the extremes in imbalance are minimised, both in terms of scale and duration, then you won’t be driving players away because they feel forced to either pick that FOTM, and then be considered a less skilled player for taking the easy choice, or pick a weaker class that they may actually prefer the mechanics of, and have developed skill to play, but then be considered as not contributing to the team as well as they could be, had they chosen the FOTM.

    Lastly, I’m only speculating here, but I believe Ghostcrawler really does have a genuine passion for PvP, and that may well have been part of the reason he left a very prestigious position at Blizzard for at-the-time-newcomer Riot, because their studio provides a more supportive culture to make a balanced PvP game, while Blizzard were putting PvE first to PvP’s detriment.

    I may have had my issues with WoW’s PvP balance (moreso than many as I played several brackets), and said some cruel things, but while Greg was the one fielding the criticism, he obviously wasn’t in a position to solve everyone’s problems (and he also did some great work on both Cataclysm and MoP, outside of PvP, which in my book makes him one of the better lead devs WoW had. I’m pretty sure Vashj’ir, my favorite zone in the game, wouldn’t exist if Blizzard’s lead dev for Cataclysm didn’t have a PhD in Marine Science).

    Now, at Riot, he’s in that position, and his understanding of the underlying systems that affect balance is clear from what he’s said above.

    My only suggestion to him would be to take a leaf of out Linux developers’ book, and have both a stable and release branch for the game.

    Release is where all your balancing trials, experiments and new, (publicly) untested characters go, giving players access to everything new, but at the expense that not everything may be perfectly balanced. It’s called release because it gets released regularly, so you can get feedback to iterate on.

    Then you have the Stable branch, where you take a snapshot of the game when you have a pretty good equilibrium. It may be missing some heroes and newer features that are in release, but it’s dependable; the kind of thing you can use for competitions without there being too many issues. It doesn’t get updates other than to fix bugs.

    • Lukehei says:

      Riot have stable and release branches; there’s the PBE (public beta environment?) server where changes are tested before they go live. The new Taric rework is currently on that but not live, for instance. :)

      • po says:

        Except that the aims of live/test server development are significantly different from stable/release, in that stable doesn’t change for many months, where live will (also there’s a difference in which branch the majority of users will be using).

        In a game like LoL this means that teams have to deal with a constantly shifting meta, instead of being able to master the available heroes, and be able to rely on them being in the same state come a competition.

        With a new stable version only being released every 6 months, they would have a more stable environment in which to practice for competition.

        If you compare esports to conventional sports, they have far more frequent changes to the ‘rules’, even compared to something like Formula 1 racing, which sees changes every season.

        I believe competition would be possible at a higher level, if players were given a better environment in which to master the game, instead of constantly having to relearn it.

    • Frostfire20 says:

      I played WoW for about 3 hours one day and decided it was basically an upgraded version of Runescape before quitting forever. That was at least 4 years ago.

      I’ve been playing League since January of 2011, a full five years now. Although it’s been off and on the last two years.

      So I feel confident enough to completely blow your idea of game balance in League out of the water. No really, hear me out.

      Heroes in League are called Champions. And they are special precisely because they have no classes. League’s original designers were the very same people who originally built Dota. Remember that game? In that game, heroes didn’t just have classes, they had very specific build paths and only one single way to play each and every one of them.

      League isn’t like that. Out of the 120+ champions they have now, at least fifty rely entirely on auto-attacks. That number has been decreasing as they go back and “rework” certain ones (it’s my opinion certain older ones should be removed entirely, but tell that to the toxic playerbase.)

      Anyway, the core philosophy is that 90% of the champions don’t actually have a specific class. For example, one of my favorites is a monster called Cho’gath. He is listed as a mage, which is completely accurate. I like building mage items on him and playing him like a mage. However, if needed I can build TANK items on him and he works just as well. One of his abilities is a short-range silence, another is a long-range knock-up. He can perform very well as a cc-heavy tank. He can do even better by mixing the two, stacking damage and tankiness.

      These are simply playstyles. In-game, there are 3 main lanes right? I’ve played him in every lane except jungle. He excels in mid, supposedly is proficient in the jungle, and can do pretty well in top lane.

      It’s difficult to balance a game like League because changing one little thing makes waves. Nerf Cho’s ultimate (melee-range nuke spell, exceptionally powerful) and he’s forced to stack extra ap to make up the difference. More ability power = more damage for his other abilities.

      On the other hand, nerf his favorite item and you nerf 40+ other mage champions. Some of whom are weak enough as it is (Heimerdinger comes to mind, he used to be strong but the meta changed and now he’s useless) and others get stronger (certain fighters are really vulnerable to mages, nerf the mages and the fighters get even stronger. Played well they can carry their entire team to victory).

      My point is there’s a million things to factor in, and changing one iota of the game makes lots of other changes elsewhere. Never mind the fact that the learning curve skyrockets every time a new champion is released, and it rises even more if the champion’s abilities interact with other champions a certain way. Ie, Bard, Yasuo, Fiddle/Kennen/Nunu triple ult, etc.

      I can’t imagine how the pros must cope, having to relearn everything every couple of months. Lately Riot’s approach to solving problems seems to be sitting back/doing nothing, but I broke my habit a long time ago and rarely play anymore. Thanks a lot, toxic community.

      Tiers are something the pros get into, and in my experience, it’s not very relevant. Only the top tier matters anymore, and you don’t even have to know the tiers to guess which champions are in them. Even factoring that in, you can just play the champion a different way in a different lane with a different item build and get a totally different outcome. That’s why it’s attractive to people trying to learn the game. They can experiment with what works.

      And that’s why it’s so hard to balance. There’s so many things at work here, it’s impossible to every get it –just right–.

  5. Hmm-Hmm. says:

    Reading this type of articles I tend to wonder to what extent player perception plays a part in forming a game’s meta. A champion(type) may get a much-needed update and you might see people exclaiming they can’t now play another champion(type) because of that.

    That makes sense in some respects considering that people get used to the way it’s been. Changing the status quo can therefore have unintended effects.

    • Hmm-Hmm. says:

      I don’t play the game that way, personally. I like to have fun and stick with characters I like the playstyle and theme of. But then again, I’m usually not particularly competitive.

    • TheLetterM says:

      Perception’s definitely part of the issue. Lol has had several patches where a champion was stated to have something tweaked, but in reality nothing changed due to an oversight. Yet the playerbase was quick to shout “OP” or “so nerfed” regardless. Can’t remember specific examples at the moment.

  6. franciscasparks says:

    Fantastic work-from-home opportunity for anyone… Start working for three to eight hr a day and get from $4000-$7000 each month… Regular weekly payments…… ITS ORIGINAL SITE AVOID BY FUCK SPAMMER….

    ———->>>>>> http://www.worknow88.com