Usage Rate New Template jpg.jpg

Opening Thoughts

What’s up guys and gals, CaptainPlanet here on Thanksgiving to present the Overwatch Hero Tier List and Meta Report: Season 2 Review. Some of you may already know, but Season 2 of Overwatch was when I began collecting much more detailed Overwatch tournament data, ranging from hero swap locations, to map completion rates, to team winrates. Now that Season 2 has ended, this data collection has borne fruit: I will be analyzing historical hero usage, recounting meta lineups of the past, and will open discussion on individual Season 2 “Hero MVPs”.

I will also present many charts for your own investigation, including King of the Hill first capture win conversion, what maps each team excels on, and many more. Finally, I will also perform a deep dive into the Stopwatch vs. Timebank debate to determine what impact – if any – Blizzard’s preferred scoring format had on tournament outcomes. First and perhaps most importantly: there was a new hero released this week! Did Sombra make a splash in the pro scene? Let’s find out:


tiers.jpeg

THE TIERS

S Tier (>=95% Usage Rate*): No one!

A Tier (>80% Usage Rate): Lucio (90%), Ana (85%)

B Tier (>50% Usage Rate): Reinhardt (69%), Zarya (64%)

C Tier (>20% Usage Rate): Soldier 76 (40%), Roadhog (39%), Genji (36%), Winston (35%), McCree (24%), Tracer (23%), D.Va (21%)

D Tier (>5% Usage Rate): Reaper (19%), Zenyatta (15%), Mei (14%), Pharah (10%), Mercy (9%), Sombra (5%)

F Tier (<5% Usage Rate): Hanzo (3%), Widowmaker (1%), Torbjorn (1%), Bastion (0.5%), Junkrat (0%), Symmetra (0%)

Caution: take the F Tier with a grain of salt … it only represents the usage from a single tournament: the Overwatch World Cup. It is not meant to tell you that your favorite Hero is garbage nor is it meant for you to use as ammo to flame people in ranked play. Let’s be nice to each other.

Additionally: I do not chose the placement of heroes in a tier, only the range which defines the tier. By determining usage directly from hero time played in tournament matches, my data is objectively determined, and not subjective. I call these ranges “tiers” for SEO reasons, not because I enjoy making tier lists… Google just really loves the word “tier” for some reason


Tiers Discussion

I’m going to try something a little different this week by only discussing noticeable changes that occurred from last week, rather than explaining reasons why Lucio and Ana are extremely good over and over again. This was the first week Sombra was available for play but her release was accompanied by an extensive hero balancing patch as well. Soldier 76 received a direct boost to his damage output, and notice how quickly he jumped from the F Tier of usage to the C Tier. This roughly 40% usage swing demonstrates how razor-thin of an edge Geoff Goodman and his team walk every time they implement new balance updates. Similar to Soldier 76, D.Va shot up in usage thanks to her movement speed buff, 100 extra health, and reduced Call Mech cost. Ana received a nerf to Nanoboost – removal of its movement speed buff – but all this has done has changed who she uses Nanoboost on.

In an interesting turn of events, the newly buffed Soldier 76 is now one of the premium targets for Nanoboost, comboing his Tactical Visor to create the dreaded “NanoVisor”. But as EnVyUs proved over and over again with their D.Va stand-in Mickie, the Nanovisor is much more easily countered than a boosted Reinhardt. Oversight Episode 2’s title says it best: just shove the D.Va in Tviq’s Soldier 76 face.

Jumping to Sombra - she saw noticeable usage in tournament play in her opening week, a first since Genji’s release. Her overall usage rate was 5%, however this was skewed by OGN Apex banning her play due to a bug. Removing those matches jumps her usage to 7%, and most of that play came on Hollywood’s first point and on Lijiang Tower:

sombreh.png

A popular strategy for Hollywood attackers was to run essentially the “Ultimate Dive Comp”: a lineup of Winston, Lucio, Ana, Sombra, Tracer, and Genji. Misfits demonstrates the execution of the “Dive” below:

(source: https://www.youtube.com/watch?v=U7ThwNwChJc&feature=youtu.be&t=1583)

ezgif.com-video-to-gif (4).gif

The Winston dives deep into the cafe of Hollywood and before Reinhardt even lands a swing on him Zebbosai’s Sombra has killed LDLC’s Mercy. All three flankers make quick work of LDLC’s non-tank heroes faster than LDLC can kill Skipjack, easily securing the point. As teams get more practice on Sombra, expect more dive lineups like this one to pop up – they are risky but when they work they really work. Sombra also makes for a good defensive pick on Hollywood’s first point because of the proximity of many health packs that can provide supplemental healing. Her usage on Lijiang Tower is similarly motivated: controlling its various health packs near the point combined with Sombra’s close-quarters damage output makes her a viable pick here.


A Brief History of Overwatch Metas to Date

historical.png

Direct Link to Chart

Gif demonstrating Chart Usage

Overwatch is blessed with one of the more active* balance teams in gaming – or at least one of the more transparent ones. Throughout closed beta and continuing even today with Symmetra’s rework, Overwatch has seen many game-changing updates that have significantly altered its competitive landscape. As patches come and go, hero usage has waxed and waned. In tournament play, some heroes were more heavily affected than others, and even more others like Reinhardt and Lucio were constants throughout the competitive scene’s journey. I created the chart above using data dating back the very first LANs after Overwatch’s release and come bearing tales of the metas of old that resulted in the hero usage trends you see here.

*Debate about this all you want


Pre-Release, the Wild West

Before Overwatch, there was “Closed Beta” Overwatch – the source of much frustration and angst among those who did not receive invites (yours truly included). The genesis of Overwatch’s competitive scene and many of the changes that Overwatch players take for granted in its released version occurred during this time period. For example, did you know that Zenyatta’s Orbs used to last indefinitely? This lead to a Meta lineup dubbed “Orb-ital Destruction”, where teams would run two Zenyattas (there also was no limit on heroes!), a Tracer, a Genji, a Lucio, and a Symmetra buffbot providing 50 HP shields. Teams would load up Harmony Orbs and Shields to create unkillable super-flankers. Closed Beta was a silly place – so here’s a list of crazy things that used to be true prior to Overwatch’s release:

  • There used to be no damage falloff for hitscan classes, so McCree could randomly two-tap you from across the map.

  • Speaking of McCree, Fan the Hammer dealt full damage with each bullet.

  • Widowmaker + Mercy boosting was a thing: when Mercy damage boost was still 50% Widowmaker could 1-shot any hero with 200 health with a body shot.

  • Mercy’s 50% boost made Pharah that much more deadly, especially since Pharah could headshot with her rockets!

  • Mercy could reach players who died on the final point from spawn with her Resurrect.

  • Heroes used to generate ultimate charge from taking damage, incentivizing awkward games of chicken where teams refused to deal damage to each other.

  • Hanzo Scatter Arrow used to be able to headshot.

  • Multiple Zenyattas could stack orbs on a single target.

  • Genji dash used to deal 50 instant damage, and 25 damage over time – and his Shruikens used to deal 36 damage instead of 28.

  • There was a short time where both Genji and Lucio only had 150 health.

The lack of single hero limit lead to many strange compositions, from the 2/2/2 Winston/Lucio/Reaper composition that lead to Winston’s shield nerf from 1000 to 600 health, to double-McCree comps that abused Deadeye to create space to push payloads while being able to insta-kill any tank with a Fan the Hammer combo. Many of these changes were fixed before release, but some snuck through despite the pro scene mailing several eggplants to Jeff Kaplan himself to protest the lack of hero limit (like I said, closed beta was a very silly time).


Overwatch’s Release (May 23 - June 14, 2016) 0HL and the McCree-Widowmaker Meta

m w.png

Throughout closed beta, the competitive Overwatch community clamored for the implementation of a 1-Hero Limit, but still competed under 0HL rules for the first month following Overwatch’s official release. Team compositions at this point revolved around McCree and Widowmaker on escort maps, and Winston and Tracer on king of the hill maps. Lucio and Mercy served as the only viable healers while Zenyatta languished with 150 health. McCree and Widowmaker were highly sought after due to McCree’s ability to blow up tanks with an un-nerfed Fan-the-Hammer and Widowmaker’s ability to nearly 1-shot any hero with a boosted body shot. Running 2x Winston, 2x Lucio, and 2x Tracer was the king of the hill lineup of choice, and many of these heroes had high usage until the Widowmaker scaling nerf, and McCree Fan the Hammer nerf on June 14th.


The Post-Widowmaker/McCree Era (June 14 - July 19, 2016) – The Pharah-Soldier 76 Meta

s p.png

Widowmaker felt the June 14th nerf more immediately than McCree, dropping out of usage almost instantly in the week following the balance patch. McCree took more time to fall out of relevance and as McCree gradually shrunk in usage Soldier 76 and Reaper stepped up to take his place as the premier hitscan DPS. Soldier 76 was a common counter to the growing amount of Pharahs in lineups during this time period, who thrived without McCrees, Widowmakers, or Zenyattas in the meta. However, a certain grandma would soon arrive to knock her own daughter out of the sky.

Note: Tracer’s usage exploded on July 3rd due to Gosugamers running a 0HL tournament for one last time to prove a point to Blizzard. This tournament also included Bo5 KotH maps for the first time, and 3-5 Tracer lineups were not uncommon


The Discord Orb Era (July 19 - September 1, 2016) – The Zenyatta-Genji Meta

z g.png

The July 19th patch introduced Ana, but its most impactful change was actually buffing Zenyatta’s health by 50, bringing him and his 50% damage boost Discord Orb back to viability. The Ana release patch had the most drastic impact on the competitive meta to date: Zenyatta replaced Mercy overnight, buffs to McCree brought him back from the dead, and these two heroes’ resurgence forced Soldier 76 and Pharah out of the meta entirely while Zenyatta elevated Genji to a Discord-Orb-hunting, support-killing monster.


The Ana Era (September 1 - November 15 (?), 2016) – The 3x3, Beyblade, “Pick your Ana Lineup” Meta

ana.png

On September 1st, Season 2 launched and with it came the long-awaited nerf to Discord Orb. In the following weeks, Zenyatta would slowly fall out of the popular usage while Ana’s usage steadily grew, owing much of her initial popularity to the tank-heavy, Nanoboost-abusing lineups pioneered by Ninjas in Pyjamas. Mei saw an increase in usage to a certain level as a counter to these tank-based lineups but also due to the buff to her Blizzard she received with the Season 2 patch. Despite a nerf to Ana’s ultimate charge in mid-october, all lineups still revolved more or less around using and abusing this single hero up until the release of Sombra on November 15th.


The Post(?)-Ana Era (November 15th-Present) – Meta Unknown

post.png

Will the Ana meta continue to dominate Overwatch? Symmetra’s rework was just announced and Sombra certainly has potential to crack the lineup on some maps. It is too early to divine their significance, but just like with Ana’s release many heroes received balance updates in the patch that released Overwatch’s 23rd hero. Soldier 76 has already spiked in usage after receiving a significant buff to his damage and D.Va also seems poised to make a comeback for example. If D.Va becomes a viable pick and Ana stays un-changed, the strength of tanks could be at an all-time high moving forward. However, who knows what the release of an entirely new hero and a mostly reworked hero will have on the professional scene. I’m excited to find out.


Season 2 Hero MVPs

mvp_example.png

Direct Link to Chart

Gif demonstrating Chart Usage

Using win tracking, time played, and all of the Season 2 Hero data that I have recorded, I created a list of “Hero MVPs” for tournaments that occurred during Season 2. This list comes with some pretty big caveats, however, and is only meant to serve as a starting point for discussion of best players on each hero. Here are the assumptions and limitations of this exercise:

  • Currently, I do not have enough data to make any statistically “significant” claims. FaZe has the highest match count in my dataset with 64, which is far too low.

  • My dataset is incomplete: it does not cover every tournament that occurred in season 2. I missed all of Carbon Masters, for example. It includes data from 470 total matches, but I’m just one guy.

  • A minimum of 15 matches was considered.

  • Players are ranked by Winrate, but matches and time played on hero is also included for reference.

  • Dataset is skewed towards NA and EU regions, as APAC region only just started having large tournaments.

  • Some heroes did not have 15 matches played, and were excluded.

  • Individual talent is not the only factor in a player’s winrate – their team’s success is what is measured here. Objective ranking based on team winrate will muddy situations where a world class player plays on a less successful team.

With this in mind, here are the players with the highest winrates on each hero with > 15 games played.


Zarya

zarya.png

1. Cocco – EnVyUs

Matches recorded: 15

Winrate: 73%

2. Winz – Rogue

Matches recorded: 54

Winrate: 72%

3. Kryw – Misfits (former)

Matches recorded: 28

Winrate: 71%

Winz and Kryw are known for their amazing Zarya play, yet the Cocco snuck into the top spot for Season 2 by winning his fifteenth Zarya match against none other than Winz in the Apex quarterfinals. Cocco is known for his Reinhardt play on EnVyUs, but all three of these tank players deserve all the recognition and praise that they receive when they bust out their Zaryas. Winz’s winrate was very impressive, having played far more Zarya than the rest of his peers. Players searching for a Zarya to look up to should would be wise to emulate Rogue’s resident Zarya.


Ana

anaa.png

1. Kyynel – NiP, Unkoe – Rogue

Matches recorded: 15, 44

73% Winrate

2. ZuppehW – NiP

Matches recorded: 18

72% Winrate

3. Chipshajen – EnVyUs

Matches recorded: 32

65% Winrate

NiP pioneered the present day Ana-powered, tank-heavy lineups, so its no surprise that their support duo of Kyynel and ZuppehW are prominent in the Ana MVP rankings. Unkoe and Chipshajen fill out the rest of the top three Anas for Season 2, known for their amazing all-around support play on two of Overwatch’s most celebrated teams. Chips showed off his Ana skills against Rogue in their Eichenwalde Apex quarterfinals match by nailing a clutch sleep on a charging Reinforce – preventing him from putting up his shield and opening up Rogue’s players to Mickie’s Self-Destruct. Hitting that sleep dart just right can change the course of an entire match and these four players will be some of the best playmakers to watch as we move into Season 3.


Reinhardt

re.png

1. Fragi – NiP

Matches recorded: 34

Winrate: 74%

2. Reinforce – Rogue

Matches recorded: 55

Winrate: 71%

3. Ryb – Cloud9 + Misfits

Matches recorded: 37

Winrate: 68%

Reinhardt is the first hero that comes to mind in Overwatch when someone says “we need a tank!” and these three tank players are worthy winners of Season 2’s Reinhardt MVP award. Fragi’s Reinhardt was a tank among a team of tank specialists, often the recipient of Nanoboosts and the protection from which Hymzi landed many of his hooks. Reinforce is perhaps the most well-known Reinhardt, Rogue’s main tank who has a reputation for his uber-aggressive play. Ryb’s outstanding performance as a Misfits’ Reinhardt stand-in in their historic Overwatch Open win paved his path to Cloud9.


Reaper

rea.png

1. Tviq – Rogue

Matches recorded: 17

Winrate: 76%

2. Akm – Rogue

Matches recorded: 36

Winrate: 75%

3. Shadowburn – FaZe

Matches recorded: 26

Winrate: 69%

After several second place MVP finishes, Rogue struck gold with the two best Reapers in my recorded matches by far. Tviq and Akm are one of if not the best dynamic DPS duos in Overwatch, so their appearance here was no accident. Akm usually covers hitscan DPS – notice his much higher matches played – but Tviq’s incredibly deep hero pool allows him to play Reaper in situations where Akm’s McCree is too valuable to ignore. Much like Tviq, Shadowburn has a deep hero pool as well and often picks up Reaper alongside Twoeasy’s Tracer or McCree in situations where his world-famous Genji is hard-countered.


Winston

w.png

1. Winz – Rogue

Matches recorded: 21

Winrate: 86%

2. Internethulk – EnVyUs

Matches recorded: 19

Winrate: 74%

3. Reinforce – Rogue

Matches recorded: 24

Winrate: 63%

Many of you might be wonderomg: where is Miro?? Unluckily for Miro, Lunatic Hai is not nearly on the level of the all-star Korea team that we all watched at the Overwatch World Cup – recently dropping out of Apex in an upset to Kongdoo Uncia. Reinforce and Winz share time on Winston for Rogue, a team that has a 4-3 match record against Miro’s own – so it makes sense to see these two in the top three for Winston MVP. EnVyUs’ Internethulk on the other hand is their only Winston player, and has been monkeying around at a high level since closed beta.


McCree

mc.png

1. Akm – Rogue

Matches played: 41

Winrate: 80%

2. Gods – NRG (former)

Matches played: 15

Winrate: 73%

3. Taimou – EnVyUs

Matches played: 22

Winrate: 64%

Akm ran away with the McCree MVP and it’s hard to argue with his numbers. Taimou is widely considered to be one of the best McCrees in Overwatch – and we see him here at third place – but Akm has “quietly” racked up stellar McCree performances on LAN over and over again. Taimou is the much more visible and vocal player, whose streaming and tournament success has garnered him international recognition. The end result of this discussion is … let’s just say Akm and Taimou are pretty darn good at McCree and everybody knows it. Gods’ appearance at #2 in the McCree MVP standings could potentially be due to his low match count but any follower of the Overwatch pro scene knows his McCree is very strong. Gods is currently teamless, but we hope to see him back in the scene Fanning the Hammer once more.


Tracer

tr.png

1. Talespin – EnVyUs (former)

Matches recorded: 21

Winrate: 67%

2. Akm – Rogue

Matches recorded: 15

Winrate: 60%

3. Tviq – Rogue

Matches recorded: 17

Winrate: 58%

Another top three Hero MVP with Rogue’s DPS players, but topped by Talespin – Taimou’s former DPS partner. Talespin often played Tracer to pair with Taimou’s deadly Reaper and Roadhog on payload and king of the hill, explaining his spot here. Talespin is known for his Pharah play, but no players recorded more than 15 Pharah games in my dataset. Fear not, Talespin still held the top mark at 71% winrate over 11 matches recorded.


Genji

g.png

1. Tviq – Rogue

Matches recorded: 37

Winrate: 74%

2. Arhan – Afreeca Freecs Blue

Matches recorded: 18

Winrate: 67%

3. Shadowburn – FaZe

Matches recorded: 72

Winrate: 60%

Genji is one of the more divisive heroes in the pro scene, because every pro Genji thinks they’re the best Genji in the world. Tviq thinks he’s better than Shadowburn, Arhan has been quoted saying he believes he’s better than Tviq, and the world thinks Shadowburn is the best. What do you think?


Zenyatta

zzz.png

1. Forsak3n – FaZe

Matches recorded: 42

Winrate: 57%

2. Morte – Reunited

Matches recorded: 29

Winrate: 52%

Only two players exceeded 15 matches played on Zenyatta out of the 470 in my dataset. If you are interested in learning more about Zenyatta play, I recommend searching for FaZe and Reunited VODs!


Roadhog

ro.png

1. Hymzi – NiP

Matches recorded: 34

Winrate: 74%

2. Taimou – EnVyUs

Matches recorded: 24

Winrate: 71%

3. Gods – NRG (former), Leetaejun – Lunatic Hai

Matches recorded: 19, 24

Winrate: 69%

There is no reality where Hymzi does not win MVP of Roadhog, the man is simply too good at hooking. Both Team Finland and NiP to some extent built their lineups to abuse Hymzi’s skill as a world-class hooker, and both had great success with it. Maybe there’s something in the Finland air, because his fellow countryman Taimou is an expert fisherman as well. Lately, Taimou’s Roadhog has been devastating in the Apex tourney, even hooking Akm’s Tracer twice in a row on Hanamura’s first point. Talk about demoralizing! Gods makes another appearance as our third ‘Hog, with the same winrate as Leetaejun – Lunatic Hai’s Roadhog specialist.


Mei

Mei.png

1. Zappis – NiP

Matches recorded: 17

Winrate: 82%

2. Tviq – Rogue, Esca – Lunatic Hai

Matches recorded: 35, 17

Winrate: 71%

3. Talespin – EnVyUs (former)

Matches recorded: 20

Winrate: 65%

Tviq hates playing Mei, which is a shame because he’s so good at it – tallying up more Mei matches than Zappis and Esca combined. Mei players tend to be players who excel at other, mechanically intensive heroes and this group of players is no exception. Talespin is an amazing Pharah and Tracer, Tviq plays almost every DPS hero, and Zappis has lately been seen playing extensive amounts of Hanzo, Ana, and other supports. Esca bucks the trend, splitting most of his time between Mei and McCree for Lunatic Hai.


The rest of the heroes you can explore at your leisure by increasing the “games played” cutoff at the bottom of the chart!


Stopwatch vs. Timebank

tt.png

Full Data

In Season 1 ranked play of Overwatch, payload matches that ended in full completions of the map were decided by a replay of the map with sides chosen by coinflip. Coinflip quickly became a meme borne out of hatred of coinflip outcomes – the community felt the offense was much more favored than defense despite data allegedly showing no such bias. The players demanded blood, so Jeff Kaplan whipped up a developer update video announcing the end of coinflips and the introduction of the Timebank system with Season 2.

Pros had been clamoring for Stopwatch ever since early closed beta, where escort maps are decided solely by whoever completed the map fastest. In fact, all competitive tournaments used the Stopwatch ruleset to determine winners – refusing to decide their match results on coinflips. Timebank differs from Stopwatch in a couple of different ways, the primary of which being that the team who completed the map slower still has a chance to win the map on the replay. When Timebank was released on live servers, tournament organizers began to use the ruleset instead of Stopwatch, deciding that it was “fair enough” for their purposes. The other day, Cloud9’s Kyky asked me how many matches that would have been won via Stopwatch had their outcomes reversed due to Timebank. Now that I have a full season’s worth of winrate data, I can answer that question. Out of 470 matches that I recorded in Season 2, 64 matches invoked Timebank, and 21 were won by the team with the slower first-pass time.

Did this decision by Blizzard have a noticeable impact on the outcomes of matches? At 4.4% of matches with reversed outcomes, there were certainly some that decided the outcome of a particular matchup. Russia won a match against France in the Overwatch World Cup that would have otherwise eliminated them, for example. But, Blizzard makes the game, so they make the rules. Which would you prefer, Timebank or Stopwatch? I created a strawpoll that you can cast your vote in Here


Perform your own analysis!

I’m quickly reaching TL;DR status here, so I will present the rest of my Season 2 review charts here as-is for your own exploration!


Hero Win/Loss

w_l.png

Direct Link to Chart

Many of my readers have asked whether X hero wins more often than Y hero, and I have plotted how often a hero pick lead to a win or loss here, as well as the overall winrate on each hero. You will notice that for almost all high-usage heroes, their win/loss approaches 50% as both winning and losing teams in the match tend to use them in their matches. Hanzo and Bastion are outliers, potentially due to their usage as surprise picks.


King of the Hill first-capture bias

koth.png

Direct Link to Chart

This chart plots the rate at which teams capture the point first in king of the hill maps against how often these games lead to wins of the map. The size of the datapoints is correlated to how many matches were played. It does not seem like there is any noticeable relationship between capturing the point first and winning the map on king of the hill maps, but it was an interesting exercise nonetheless.


Team Map performances in Season 2

tea.png

Direct Link to Chart

Gif demonstrating Chart Usage

The left chart plots teams and their winrates ranked by how many matches I recorded them playing in Season 2. If you click on each team’s bar, you then filter the chart to the right to display their per-map performance, colorized by winrate. Mouse-over each map’s box for exact information on winrate.


Swap Locations on non-KotH Maps in Season 2

swaps.png

Direct Link to Chart

Whenever a swap occurred in Season 2, I recorded what heroes were involved, as well as their location. This chart ranks map locations based on number of swaps, then clicking on the bar lists every swap that ever occurred there.


Swap Success Rates

swapsss.png

Direct Link to Chart

Gif demonstrating Chart Usage

This chart plots the same swap data, but adds matches played and winrate data as well. Scroll through the “Offense” and “Defense” charts, and the dots you see correlate their size to matches played and color to winrate of matches where swaps at this map location occurred. Click on the dots to filter the menu below for more information about which players were involved in each swap!


Final thoughts and shoutouts

What a long strange Overwatch season its been! I’d like to thank all of the tournament organizers that have been putting on amazing shows – from the Battle for the Atlantic to the Overwatch Open to all of the APAC region tournaments that have given us the opportunity to see how the East and West compare. I’d also like to shoutout to the Overwatch balance team: many of you may not always agree with their decisions but their regular balance updates has provided me and many others with plenty of content to discuss throughout the year. Here’s hoping Season 3 will be just as exciting and interesting as Seasons 1 and 2!

 

Until next time,

CaptainPlanet