The Overwatch League's May Melee tournaments last week were the highlight of the league's 2020 season thus far. They brought drama, highlight plays and better viewership. Perhaps most importantly for the league, May Melee moved what had become a fairly large spotlight away from myriad problems that have plagued it all year, particularly a recent wave of defections of top pro players like Jay "Sinatraa" Won to other games or retirement.
These tournaments were exciting in a way that Overwatch hadn't been for the majority of 2020, at least since the first few homestands in the United States before the coronavirus outbreak put an end to in-person events. Part of this was due to format. Tournaments are by design more exciting than league play. This is a problem that also comes up in traditional sports in the United States, especially baseball, basketball and hockey, which all have lengthy seasons that can feel like a grind outside of a few key matchups.
Another part was due to the absence of hero pools.
What are hero pools and why are they bad?
Hero pools remove one tank, one support and two DPS heroes from the available pool of 32 heroes that can be played during a given week. Each week, these four heroes are pulled at random from the top few most-played heroes in the league.
Even typing out the above description requires massive caveats. The league has changed its mind multiple times over the course of the 2020 season. Hero pools were decided by different standards at different times this year as the league struggled to decide how to implement them effectively. At the beginning, hero pools were not united across the solo queue ladder and competitive play, making it more difficult for pros to practice with the hero-pool limitations.
Ideally, hero pools were supposed to ensure that no meta could become as stagnant as the 2019 GOATS metagame composed of triple-tank, triple-support hero compositions and named after the North American Contenders team that first began using this composition. The purpose of hero pools, as described by league vice president Jon Spector is as follows:
"You might have a situation where Week 1 [Zenyatta] is a little [overpowered] and played all the time, so in Week 2, Zen gets pulled," Spector told ESPN when hero pools were announced in January. "By Week 3, Zen is available again but there has been a [quick] balance change and Zen's discord orb is a little less effective than it was, and now maybe [he's] played half the time. I don't think any one of these things by themselves is necessarily a silver bullet, but when you start to pull all these factors together, I'm super confident we're going to see a lot of diverse team [compositions] and a lot more different strategies."
The immediate problem with hero pools is inherent in their description: they force teams to adjust week to week with little warning and practice time. While esports games have always had a push-pull between adapt-or-die and the perfection of a specific meta, hero pools took the former to an extreme that is mentally exhausting for players and particularly team coaching staffs. My first argument in this lengthy hero pool discussion is that dominant metagames aren't unilaterally bad. In fact, some of the best esports teams have showcased their unique skill in either perfecting a metagame or finding a solution to a meta after it has been perfected by other teams.
The GOATS meta
So much has happened since hero pools were implemented, it's worth returning to the question of why use hero pools at all? In order to understand why hero pools came to be in the first place, and why the league is so afraid of a dominant metagame, we need to go back to 2019 and the GOATS meta that dominated the Overwatch League for the majority of 2019.
This composition followed the release of support-tank Brigitte and was usually composed of three tanks (Reinhardt, D.Va and Zarya) and three supports (Brigitte, Lucio and Zenyatta). It required strong ultimate ability timing and coordination but didn't offer the same immediate viewer excitement of seeing a DPS hero like Widowmaker get three kills. Initial reactions from teams and viewers who watched GOATS in various Overwatch Contenders matches around the world was that it was comparatively boring to watch teams employing GOATS.
Prior to a single game of professional Overwatch in 2019, the complaints about the rising GOATS meta were already dominating discussion. As early as opening day, signs at the Blizzard Arena sported drawings of Twitch's ResidentSleeper emote alongside a goat.
For me personally, as a comparatively rudimentary Overwatch analyst to others in the scene, GOATS was actually a major building block toward understanding how Overwatch worked. It slowed down the game and coordinated ultimate and ability usage in a simple way that I could digest, and after a few repetitions, immediately break down.
This then led to certain teams, especially the ones I watched in Chinese and South Korean Contenders matches before the official Overwatch League season began, coming up with creative workarounds on certain maps, like StormQuake's quadruple DPS solution on Route 66. Chinese teams often seemed unconcerned that GOATS was the default way to play the game and sometimes ran whatever they wanted regardless. This was a harbinger of what would come from one specific Chinese team in the Overwatch League.
The Chengdu solution and 2-2-2 role lock
The main issue cited by the league (by implementing hero pools) with GOATS was that metas were too stagnant and teams were too scared to adapt. Hero pools would force more adaptation in 2020, solving the complaints of a boring meta.
Yet, the league already had arguably seen adaptation and a few solutions to GOATS courtesy of the Chengdu Hunters. While it's a bit specious to say that the Chengdu Hunters didn't play GOATS at all -- they did have to enter GOATS mirror matches against teams at certain points -- they focused primarily on playing their own style. This style was completely unlike GOATS in any way, and relied on tank player Ding "Ameng" Menghan's Wrecking Ball as a main tank over the ubiquitous Reinhardt.
It's too simplistic to say that teams should have just played their game to reckless abandon like Chengdu -- although from a viewer's standpoint, that's why I loved watching them. However, it shouldn't be much of a stretch to imagine a world where a team that was better than Chengdu in a GOATS mirror also tried out a variety of more DPS-heavy compositions.
Ultimately, the problem with GOATS was that the composition was so strong that it required mirror matchups because if one team went with GOATS, they would automatically win. Yet even this began to change with the addition of DPS hero Sombra to compositions and minor adjustments made by teams as the season wore on. Arguably, the solution or GOATS-breaking metagame was already being worked out by certain teams, albeit slowly. And this isn't even considering how the league's balance team could have addressed some of GOATS' dominance through fine-tuning certain heroes.
The winds of change were aided by the arrival of the league's implementation of a 2-2-2 role lock late in the season, which forced teams to pick only two tanks and two supports, thus breaking the GOATs composition. Prior to the announcement of 2-2-2 for Stage 4 of the 2019 season, teams like the Shanghai Dragons were already eschewing GOATs completely and running more DPS-heavy compositions with success.
ESPN Daily newsletter: Sign up now!
An overcorrection to a solvable problem
Before hero pools, there was the 2-2-2 role lock -- the Overwatch League's first top-down attempt at making teams play the way the league wanted them to, and a harbinger of what was to come with hero pools in 2020.
This change was to date the most widely accepted game-breaking balance change at a competitive level that I have ever seen in esports. There was naturally some pushback, but on the whole teams seemed ready to accept 2-2-2 and move away from GOATs permanently. Hero pools were met with a larger amount of criticism upon announcement, but the precedent for rewriting the competitive fabric of the league had already been set.
The first two weeks of competitive play featured a variety of compositions centered around the use of DPS hero Mei to control map points. She was used to set up a variety of hitscan options, two of which (Widowmaker, McCree) were banned after the first hero pool draw. Even within the guardrails of hero pools, one specific hero dominated conversation and drew audience ire. When Mei was finally banned (by desk analyst and host Soe Gschwind Penski's cat no less) people celebrated. Instead of Mei, teams began using Torbjörn to control points that weekend.
While it's true that hero pools have forced adaptation -- and in many cases, intelligent commendable adaptation -- it's come at the cost of coaches' and players' mental well-being as well as audience engagement (which was one of the reasons why hero pools were implemented in the first place). Coaches have frequently come out on social media, describing how difficult it is to prepare their teams when the meta changes so drastically from week to week. The addition of Overwatch's most recent hero, Echo, only complicated matters, as teams also had to learn how to play her among all of the other meta changes.
Players are exhausted, especially when there's little reward for them in perfecting anything when everything will change the coming week. And these complaints don't even begin to scratch the surface of problems inherent to hero pools that were also GOATS problems, like certain players being effectively benched by a given week's meta.
Although May Melee was also difficult to prepare for, the end result was a freer, vastly more entertaining product as teams were able to play however they wanted using whatever heroes they wanted. We saw a variety of compositions appear depending on team and map, alongside the natural narrative strength of a tournament format as opposed to a league format.
I think the league vastly overcorrected for 2019's GOATs meta, beginning with the 2-2-2 role lock but particularly with hero pools. Although May Melee is an admittedly small sample size, at this point I think the league should consider doing away with hero pools altogether. It's not like they haven't already made sweeping format and meta changes this year.