Overwatch Tier list and Hero Meta Report #18: A Meta out of time

Opening Thoughts

What’s up guys and gals, CaptainPlanet here to present the Overwatch Hero Tier List and Meta Report #18: A Meta Out of Time. This week, the Report is a little bit late -- but it’s worth the wait. This past weekend I spent….let's call it an unhealthy amount of time overhauling how I collect the Report’s Data. While I still have many more features to add, I have finally developed a process for collecting and presenting a very important statistic: Time. Yes, Time...my old nemesis. Consider this your moment of hype, and hold on to it until later in the Report because as I write this, Blizzard is announcing their new Hero: Ana. Let's take a look at her Hero introduction video, and I'll have a quick live reaction to Overwatch's newest (oldest?) Agent. 

Not to get too hyped, but Ana is everything I that I have been advocating for in order to help create and maintain a healthier Meta. Overwatch presently has 21 Heroes, four of which are Supports, three of which can Heal, two of which are presently viable: Lucio and Mercy. Enter Ana, and the Balance Patch that she will release with. Along with buffs to Zenyatta’s health and Ultimate, Ana will become Overwatch’s fifth Support and fourth Healer: a Sniper with an extensive kit of CCs, buffs, and debuffs. Healers and Tanks have always been staples of the top of the Meta due both to their scarcity and their essential roles in a Lineup. With the release of Ana, the buffs to Zenyatta, and the retooling of Mercy, we can finally expect a bit of a shakeup at the top of the Meta and among the Support Role in particular. 

Back to business -- let’s jump to this Week’s Hero Tier List, because that is the highest hitting google search term of my content and I am not afraid to admit it. 

As always, Raw Data for this and subsequent charts can be found HERE The data is now presented as separate sheets for each Map, for better and more detailed tracking. REMINDER: This Data was collected from the Beyond the Summit Overwatch Cup, where 1 Hero Limit, 2CP Map Bans, and Stopwatch Format was used for Payload Maps

The Tiers

S Tier (>=100% Usage Rate*): No One! Sayonara, S-Tier!

A Tier (>80% Usage Rate): Mercy

B Tier (>40% Usage Rate): Lucio, Reinhardt, Pharah, Soldier 76, Reaper, Zarya

C Tier (>20% Usage Rate):  Roadhog, Winston

D Tier (>5% Usage Rate): Widowmaker, Tracer, Genji, Junkrat, McCree

F Tier (<5 % Usage Rate): Hanzo, Symmetra, Torbjorn, D.Va, Bastion, Mei, Zenyatta

*What is Usage Rate? It is NOT the percent chance of seeing a particular Hero in a Match, it’s ...actually now that my Data is Time based it's exactly that :^)

The Tier Ranges I’ve chosen** reflect different states of “Balance” in the Meta. This week, the Meta that is being analyzed is the Meta of 1 Hero Limit, 2CP Map Bans, and Stopwatch Scoring for successful attacks on Payload Maps.

The S Tier: It finally happened: Blizzard is removing Hero Stacking from Ranked Play. The Pros have been asking for 1 Hero Limit since early Closed Beta and while I was slow to jump on the bandwagon, I had no choice but to join them after last week’s Tracer and Lucio-filled 0HL experiment. It seems that Jeff Kaplan and the rest of the Dev Team have also joined us on the 1HL express and next Season’s Ranked Play will be devoid of Hero Stacking. With this in mind, I will be removing the S Tier until further notice.

The A Tier: The “Non-Negotiable” Core Heroes. This week only Mercy reached the A Tier, clocking in at a 97.15% Usage Rate. Reaching the A Tier, not to mention reaching nearly 100% Usage Rate is even more impressive than usual because this week’s Data is entirely Time-based. Mercy’s 97% Rate does not mean that she was used in 97% of all Sides -- but that she was used for 97% of all Time Elapsed in Matches, normalized by Match Length. Needless to say, it did not matter what game type was played, or if a Team was on Offense or Defense: if you did not have an Angel on your shoulder you were doing something wrong.

The B Tier: The “Balanced” Heroes. The Balanced Tier of Heroes encompassed an interesting spread in types of Usage this week. Reinhardt, for example, was used far more often on Payload Maps -- evenly on Offense and Defense -- than he was on King of the Hill. Heroes like Lucio and Zarya by contrast saw a huge spike in King of the Hill Usage, while Pharah, Soldier 76, and Reaper saw relatively balanced usage in all modes as the current Overwatch Meta's primary DPS Heroes.

The C Tier: The “Meta-Dependent” Heroes.The two Heroes who fall into the Meta-Dependent  C Tier are victims of the return of 1 Hero Limit. Both Winston and Roadhog actually exhibit differences in Usage Rate type comparable to their B Tier neighbors: Roadhog was used much more often on Payload Maps than King of the Hill, while the opposite was true for Winston. It seems like any shift towards either game type’s direction would be enough to boost one of the pair into the B Tier -- both fell just short of its 40% Usage Rate cutoff.

The D Tier: The “Specialist, or Weak” Heroes. Once again, the Specialist Tier seems to indeed be filled with Specialists. Widowmaker leads the group as Overwatch’s only Sniper (until Ana!) -- and has still not quite recovered from the Balance Changes she received several weeks ago. A well-played Widow can still be devastating, just like a well-played Flanker like Genji or Tracer. It was fun to watch Tracer’s usage spike during 0HL rules, and tank now that they have reverted. Balancing around 1 Hero Lineups will give Blizzard more flexibility to make appropriate tweaks if they think Tracer or any of the other Heroes who fall into the D Tier need a little love. Speaking of which, McCree finds himself once more in the D Tier, but things are looking up for him if reports from Pros playing the PTR are to be believed. McCree has received some reduction in his damage falloff at midrange, which has Pro Level McCree mains like IDDQD chomping at the bit in excitement. Like Widowmaker, a well-played McCree can be devastating and I expect his Usage to spike as soon as the Balance Changes on the PTR go live.

The F Tier: The “We Exist, Barely” Heroes.  One side effect of doing my Reports as Time-Based instead of Pick-Based is that the truly “Unused” Heroes take a beating in Usage Rate. Previously, if a player picked Hanzo as a joke pick at the beginning of the match, that pick would count as one whole pick. The same phenomenon would occur for a somewhat successful D.Va stall at the end of a long Payload Map. By swapping my Usage Rate calculation to be Time-based, these Heroes receive counts for the time they were actually played, and as such the F Tier is a bit more crowded than in previous Weeks. I would caution readers to keep an eye on the Heroes who are constant denizens of the F-Tier in the coming weeks -- if they’re truly only active for <5% of Match Times, then they are in definite need of some Balance Patch lovin’. Luckily for Zenyatta and D.Va, both are getting direct buffs in the upcoming Patch -- something I’m extremely excited about as a (not-so) closet D.Va main. 

** I do not chose the placement of Heroes in a Tier, only the Range which defines the Tier. By determining Usage Rate directly from Hero Time Played  in Tournament Matches, my data is Objectively determined, and not subjective at all. 

A Meta out of Time

This weekend, I lost my mind

For several months, I’ve been getting a lot of the same feedback -- questions like these:

• Captain, can you track when / where swaps occur? And who plays what? 
• Why can’t you tell me what Map X happened on? 
• Why do you only track Picks, not Time Played on Hero?
• I want to know what Heroes tend to win more than others, can’t you do that? 
• Captain, can you be our Team’s Analyst?*

Up until now, the answer to all of those questions has been No -- I don’t have the time and don’t have the resources to do so. However, something "clicked" this and I spent the next 36 out of 48ish hours of my own weekend making...a spreadsheet. I have re-built my Data-Collection process from scratch, and I’m proud to share with you the result: the "POW" (PlanetOverwatch) Sheet. (screenshots below)

*This was actually the genesis of the idea for the sheet -- I want to help all teams equally!

First off, you can find a V1.0 Link HERE

Secondly, if you would like to see an extensive demonstration of this sheet's capabilities, I recorded all of this week's data using it -- housed in this Google Docs Folder. 

I developed this tool for my own use, but also for all of the aspiring Overwatch Analysts out there. This spreadsheet is meant to be used by Overwatch Enthusiasts of all levels -- from the Amateur going over their Team’s scrim VODs all the way up to Envyus' Army of Team Analysts**. Let’s go through the POW Sheet’s current, and eventual features:

**May not actually exist

Current fields that can be filled in:

• Teams
• Players
• Team Region
• Maps
• Sub Maps (KotH)
• Heroes linked to Players
• 6 Sides worth of Data
• Up to 12 Swaps per Player/Lineup Slot per Match
• Time Swaps Occur
• Rough Location on Map Swaps Occur (Payload/2CP)
• Match Start, and Endtime
• Map Winner

Current Summary Data that Emerges from the above:

• Heroes played (Offense, Defense, King of the Hill)
• Hero Play Duration
• Match Duration
• Percentage of Match Spent on Each Hero

Future Features I will be adding:

• Swap Counts (Already have this, just not in a pretty summary form)
• Summary breakdown of Swaps based on Map position, Offense/Defense, and tied to Player
• Player-Hero interaction Tracking (swaps to and from Heroes, most played Heroes)
• Automatic Win Determination
• Point-Win Tracking -- leading to Checkpoint-based Lineup Win Share Tracking
• -INSERT YOUR SUGGESTIONS HERE- (keep an eye out for my Reddit threads, or hit me up on Twitter @CaptainPlanetOW)

This sheet is my baby, and I’m very excited to share it with you all. Please let me know what you think of it, both the good, but especially the bad! I’m looking for bug testers, format feedback, and everything in-between to create the best resources for the community that I can -- here’s your shot to help me out. 

What does this mean for the Data?

Previously, my Data Collection Process was governed by three rules:

• Count Heroes at the beginning of the Game
• Count all Hero switches that occur throughout the game
• Do not count Obvious, opening cheeses

This could (and did) create some strange occurrences with Trends that emerged from the Dataset. There would often be strange spikes in Tracer and D.Va Usage, simply because Teams would mount a somewhat successful stall Defense on a Payload that I would feel forced to count this as a “Switch”. This would count as a full Hero “Pick”, and boost these and other Heroes’ numbers accordingly. A similar phenomenon could occur with opening Lineups. By and large, I always counted the initial Lineup as full picks (barring extreme examples like the old 6 Hanzo King’s Row cheese) -- as this represented a Team’s initial strategy for taking the Point, no matter how obscure. If, for example, a Team wanted to try 5 Winstons and a Lucio on Hollywood -- by golly I was going to count all of their picks to give them credit for the idea. 

Now, my Data collection proceeds as follows:

• Note the Lineup
• Note Swaps that occur, when they occur, and where they occur. 
• ???? (Formulas do a bunch of adding and subtracting)
• Profit (Duration of Match spent on Hero gets spit out)

This means the Data much more accurately represents each Hero’s relative “Usage Rate” in a given match. If a Team only uses Zarya and Pharah for the First Point’s Attack and swaps to Reaper and Roadhog for the rest of the match, Zarya and Pharah will only get ~0.25 of a Pick, while Roadhog and Reaper get ~0.75. A Hero can only attain a “full” pick if they remain on a Team’s lineup for the entirety of a Match. 

“But wait, what if Player A starts on Mercy, then swaps to a different Hero at the same time Player B swaps to Mercy?”

The above example would still count as one "continuous" Mercy play. In fact, my POW Sheet can account for any number of stacked Heroes -- I built in this functionality before Blizzard’s announcement removing 0HL -- so I got that going for me if 0HL ever returns. If, for example, you had two Soldier 76’s on your lineup for an entire match, their “Time Played” stat would be double the length of the Match -- still allowing for a 200% “Usage Rate” by my old definition of the term.

Let's Talk Swaps

One of the advantages to how I can now collect my Data is that I can now track much more nuanced statistics -- like Swaps. I actually collected this week’s Data twice -- once with Icarus under my old method and a second time as a dry run to test my POW Sheet’s functionality for bugs. The differences the two  Datasets seem to roughly this chart I’ve created, which tracks the number of times Heroes were "Swapped to" for each Hero in the Beyond the Summit Cup:

The top several Heroes in this Infographic were actually the Heroes who lost the most Usage Rate % when converting from the old, Pick-Based System to a Time-Based System. This makes a lot of sense: these are the Heroes who would get a full “Pick” under a strictly “Pick-based” System, but lose part of their Usage Rate in a Time Based System since they were not part of the starting lineup. This is also balanced somewhat by Overall Usage Rate -- while Heroes like Zarya are very effective on singe points of Maps and thus likely to be swapped to, her Usage Rate was not high enough to boost her Total Swap Count. In the future, I will probably do some normalizing of this Data instead of a Total once I build more functionality into my POW Sheet. 

Looking at just Swaps, we can see that the Heroes most swapped to are Reaper, Tracer, Winston, and Pharah. For all but Tracer, most of these swaps were due to Map Design. Teams will often start out with a specific Lineup on Payload/Hybrid Maps in order to take the First Point, upon which they will immediately change Heroes once the Map’s area changes significantly. Hollywood in particular is an excellent example of this phenomenon. Regardless of what Lineup Teams use to take the first point, both Teams have a high likelihood of swapping to a Winston during its Street Phase to help control the high ground. Reaper gains a lot of swaps for the opposite reason -- in the alleys of King’s Row and in close-together indoor areas Reaper is the DPS King. However, most Maps only have 1-2 specific Reaper-favorable areas, leading to his high Swap Rate. Tracer once again  stands as an outlier for both last-point Defenses and King of the Hill Lineup shakeups. Even in 1HL Format, Tracer will often be the first and only Hero swapped to whenever a non-Tracer strategy fails to take a King of the HIll Capture Point -- perhaps because players are still quite used to using her from their two weeks of having to train for ESL’s 0HL ruling (now reverted).

Historical Tracking

Speaking of ESL -- they, like Blizzard have reverted to 1HL which is why this Week’s report focused on Beyond the Summit’s 1HL Overwatch Cup instead of the first week of ESL’s 0HL Online Qualifiers. It's actually really entertaining now to look back on the one “experimental” 0HL GosuGamer Weekly sandwiched between two 1HL tournaments in the Historical Tracking Chart. Watching as Lucio, Tracer, Reaper, and Winston tank while Mercy, Reinhardt, and Soldier 76 bounce right back demonstrates just how favorable the former Heroes are on King of the Hill Maps. It’s been a whacky couple of weeks in the Meta, and with this new Balance Patch and Ana on the way it’s about to get even weirder. 

Final Thoughts and Shoutouts

As always, none of this Data could exist without the help of the Competitive Overwatch Community -- so remember to check out my Competitive Overwatch Beta Twitch Directory! This Directory features all of the Twitch Streams and Social Media of as many of the people who helped make the Beta Competitive Scene great that I could find. Get out there and give these people a Follow -- without them there would be no Competitive Scene and no Overwatch Hero Meta Report. This week I'd also like to give an extra special shoutout to you: the Overwatch Community. If it wasn't for all of your bellyaching ER I MEAN POSITIVE FEEDBACK I would never have been motivated enough to create this new content. Also, another shoutout to the very nice, colorblind Redditor who let me know that my Infographics were difficult to read, I updated my color scheme just for you <3

Until next time,

CaptainPlane

Icarus' corner

This week, as always, I partnered with Icarus to compile the data necessary for both of our Meta Reports. If you'd prefer your Meta Reports in audio/video form, head over to his YouTube channel, or check back here around every Wednesday of the week! Icarus finished early this week, check him out below!

Like this content? Tweet  and Share it with your followers!