Reign of Gaming

This may be a strange thing to hear me say, given my very positive opinion of Riot (so be sure to focus on the word 'seem') but Riot's live balance team often behaves in a way that makes it seem like they don't quite have coherence of design. On the one hand, they insist again and again that they balance for competitive play (which they mostly do). On the other hand, they have a habit of nerfing strategies that are anti-fun; these strategies tend to be much weaker once you know how to counter them, but they negatively impact most non-competitive games. Notice that I'm using the word 'competitive', and not really discussing Elo. Why is this? Because pub-stomp strategies often are strong even at SUPER high-Elo; it's not the strength of the players, but the coordination and trust that are important. When Eve was strong, her high win-rate went all the way to high Elo, and the same goes for Shaco. Even now, Shaco's winrate is (slightly) higher in Platinum than it is in Bronze. But even when Shaco was at his most deadly, he never saw tournament play. So obviously, to characterize Riot's balance priorities requires more than one phrase, because these are complicated issues. But because of the mixed nature of the player base, Riot also has conflicting interests. So let's say you're Riot, and something OP has just come up on your radar. What do you do?

Step 1: Identify a potential problem

How do you identify a problem? Well, it probably starts with public outcry. People say 'I can't figure out how to beat X' or 'Y is so strong; I feel like you either have to ban it or do it yourself!

Step 2: Data Analysis

The next step is to get good data. But what kinds of data? Well, first you could look at perceptions. How many people think 'ban it or join it'? Look at the play-rate. People tend to pick what they think is strong, so pick-rates can give you information about perception of power, as can ban-rates. And how well do these work? Well, there does appear to be some correlation between Pick% and Win%:

as well as Ban% and Win%:

So let's look at the top 15 banned champions; these are the champions people perceive as so strong they don't want to chance playing against them. Is this a valuable rubric to go by for balance?

Well, we see some champions with above average win-rates, and some with below average win-rates. On average, the win-rate of these champions is .50666; slightly above 50%, but not hugely.

But Gentleman Gustaf, this is for all Elos; wouldn't some of these champions be much scarier in the hands of a great player, or much scarier against bad players?

Turns out that the win-rates are about the same in Platinum as in Bronze. This only enforces my prior point: when it comes to countering a lot of strategies, it's about team organization, not individual skill. This is why Riot will occasionally nerf something that has never seen tournament play; no matter how good a player you are, if your team doesn't know how to beat this strategy and have practice against it, it doesn't matter if you do.

If you're curious, here's the data sorted by pick-rate (it turns out about the same):

But what especially amazed me is how many of the top 15 win% champions were not must-picks OR must-bans:

Singed, Rumble, Tryndamere (until recently), Janna, Fiddlesticks, Varus, and Twitch all see pretty low rates of play, and yet are in the top champions!

This leaves Riot in an awkward situation. Everybody hates playing against Blitzcrank, because one grab can change a game, and similarly, Shen, Malphite, and Kha'Zix have been permanent bans for a long time, and yet their win-rates are nothing to write home about. So exactly what data should they be looking at? Release Darius had a notoriously high pick and ban rate, and yet a 47% winrate. Thousands of community complaints later, he saw a nerf, although one could argue that this nerf was less about balance and more about making him less frustrating to play against. But as far as community response, Riot has to deal with two schools of thought:

This is OP and ruining games nerf it now!

An example of this? The recent reaction to Blade of the Ruined King; it was showing up everywhere, and mathematically, it seemed pretty strong. On the other hand, there's a strong argument that it was only strong because the current defensive itemization meta prioritizes health, something which has been changing over time (Warmog's being replaced with Sunfire Cape + Locket of the Iron Solari, for example, and Randuin's Omen becoming increasingly revealed as extremely strong). And when they did that, many people immediately said:

Let the meta evolve and people will learn to deal with it!

Of course, on the other hand, when they DID try to do this in Pre-Season 3 with Black Cleaver, the complaints were everywhere. Another example of this? Darius, I guess. Darius' pick and ban rates have consistently gone down as people learned to deal with him in team-fights and in lane.

So here's a standard progression of something OP:

Step 1: Holy shit this is OP!

Something is discovered that people think is OP, and (for the sake of argument) let's assume it is. I hope I've shown above how hard it can be to actually tell if something's OP or just hard to deal with and we haven't figured out how yet. Soon, you start seeing it in every game, either as a ban or a pick or the item that 6 people on your team have. You start feeling like you should either ban it or follow suit. A few innovators start trying to counter the strategy.

Step 2: Addressing the Problem

One of two things then happens (often both).

• Riot gives it a small nerf to see what happens.
• Innovators (typically high-Elo players) figure out how to counter this strategy.

Step 3: The Interim

People still haven't really learned how to deal with the strategy at this point. They've (maybe) heard that it's beatable, but they don't know the specifics, or they may have heard the wrong specifics. If it was nerfed, they might still be getting it because they don't know what item SHOULD replace it; they have too many options, and they'd rather stick with what they know is strong (but not the strongest) than build something that could be stronger but might be weaker or have stronger counters). As such, from a balance perspective, it may still look problematic.

This is like Warmog's being nerfed, and still being bought 8 times in every game, or Black Cleaver being nerfed and still seeing teams of 4 physical champions, or Jayce showing up in every game despite nerfs. This is the point Riot is constantly worried they're actually in. If they already nerfed the strategy, they're worried they're in Step 3. If they haven't yet, they're worried that the counters are out there and trickling down, and so they're in step 3. Hell, even when they first discover the OP strategy, they're probably worried that they've already reached Step 3. Why? Because sometimes it takes so long for OP strategies to trickle down into mainstream play that they've ALREADY been countered in high-Elo/coordinated play. When a strategy is developed in high-Diamond play, and doesn't see much play below Diamond (nobody's thought of it), but doesn't see pro level play (too easy to coordinate against), by the time it becomes a balance problem in other Elo brackets, it's probably already been countered in high-Diamond play.

A Conflux of Problems

So Riot ultimately has to deal with two problems that make balance very hard to deal with.

On the one hand, they have to deal with the current game experience, making sure it is still enjoyable and balanced FEELING. On the other hand, they don't want to rush to make changes, because nerfs elsewhere, meta-shifts, and people learning to counter something can all turn an OP strategy into one that is understood to have strengths and weaknesses.

And to make those decisions, they have to rely on data. Unfortunately, the bulk of their data doesn't come from high-level play, which can make it easy for them to miss an OP strategy falling by the wayside, as it could take months for it to trickle down to their data.

 For more of my work:

-- Find old posts @ the RoG forums and new posts every Wednesday (3 PM) and Sunday (9 AM).

-- Feel free to find me in the "A DIFFerent View" chatroom on the NA server.
-- Contact me at GentlemanGustaf@ReignofGaming.net
-- Follow me on Facebook and Youtube for updates on all of my new articles, videos, and streaming.

Gentlemen of the Day