Welcome to Weekly Recall! It’s our new blog series in which we’ll spotlight a topic close to our developers’ hearts, including details on some of the latest changes to Quick Play, Competitive, and Stadium. Expect data deep dives, design breakdowns, and surprises along the way.
Gavin’s helping to kick things off with a look into Hero Ban data. Who's getting the ban hammer dropped on them and what does that mean? Find out below!
- The Overwatch 2 Team
Hey everyone. As we’re halfway through Season 16, I want to dive into Hero Bans data from the last couple weeks provided by our Global Insights team and Server Engineers. Seeing how you all have approached Hero Bans so far across different platforms, regions, and rankings, has been incredibly fun. Are you ready to geek out over some numbers with me?
With the data only split by platform, the first graph is ‘PC’ and the second graph is ‘All Consoles.’
Sombra leads the PC pack with an impressive 85% ban rate, followed by Zarya at 59% and Doomfist at 43%.
On consoles, Sombra is banned even more, with a 93% rate. She’s followed by Zarya at 57% and Symmetra at 23%.
On PC, our newest hero is dashing through the wind to catch up with the pack, earning a 27% ban rate. Freja’s ban rate by day is climbing steadily!
She's remained at a pretty low ban rate on console, though.
The least banned Hero on PC is Baptiste at .16% and the least banned Hero on console is Hanzo at .24%.
There are some big differences based on Rank. Here's what the graph looks like for PC at Masters and above.
Here's what they look like for console at Masters+.
Sombra is banned more often by console players even at the highest ranks, but she isn't as highly regarded on PC. While overall, Zarya is the most banned Tank, when we look at Masters and above D.Va handily steals the most banned slot for Tanks for both console and PC. Ana is the most banned Support Hero at higher ranks on PC, while Mercy is the most banned here on console. Zenyatta also sees a lot more bans on PC than he does on console.
How do Hero Bans look in different regions? The following charts show the rates for Masters+ Role Queue, where the biggest differences appear and we have the largest population of high MMR players. Every region has the same most banned Hero in each role (Freja, D.Va, Ana), but there is some nuance. EU and US worry about Mercy and Wrecking Ball, but KR mostly ignores them. KR bans are much more polarized, where fewer Heroes get a majority of the bans. The US and EU ban Sojourn pretty often, while she’s ignored in KR.
Another important factor to consider with Hero Bans is that players don't just ban different Heroes at different ranks, they also ban different numbers of unique Heroes based on rank. There’s more alignment between players about which Heroes should be banned at higher ranks. At Grandmaster, the average team tries to ban four Heroes. At Bronze, the average team tries to ban six. The Damage role has the most disagreement upon Heroes to ban, followed by Tank and then Support. At Grandmaster, all three roles have very similar ban rates, which aligns with data we've discussed in the past that implied that all three roles have very similar impact at the highest ranks.
We know that one of the biggest questions you'll have is: "Now that we have all this data, what do we plan to do with it?" In talking to our Hero Designers, they explained that they've found the data we've collected so far very interesting, and we’ll be using it in the future to help inform balance changes. Heroes that are more niche or map specific may need to become more generalized if they're always banned where they perform well. Heroes that are very unpopular might see more adjustments regardless of performance. Ultimately though, there aren't any specific changes in the works based on this data yet, and Hero Ban data alone would never dictate how Heroes are balanced.
Speaking to the system itself, we're happy with how it works overall. We believe the system is meeting most of its goals, and we're excited to see how Hero Bans evolve over time, especially once we start releasing more data about our Hero win rates and pick rates in the future!
One way that we're considering changing the system is to add a feature to Preferred Hero that would allow it to act as a tiebreaker.
Around 42% of all games since we launched the feature have had 7 or 5 vote ties. A player's first Choice is worth 7 votes, and a player's second Choice is worth 5 votes. That means it's common for multiple players to be voting for different Heroes and for us to have to break ties randomly, which isn't awesome. As you can also infer from the ban variance data above, that's a lot more common at lower ranks where there's less alignment about who to vote on. So, the feature we're considering adding is making each Preferred Hero choice worth one negative vote.
One vote isn't much, since each player technically has 15 votes if you total all three choices. This will mean that if multiple Heroes are tied for a ban, the Hero that isn't preferred by any players will be banned. This negative vote would never count twice, such as if two Damage players preferred Widowmaker. These negative votes would also never apply to the other team's votes. Let us know what you think of this idea and keep sending us all your other feedback on Hero Bans!
Systems like Hero Bans (and our upcoming Map Voting feature, which launches in Season 17) are constantly improved by your feedback because we are making a great game—together. Also, don't forget to check out our next Drive starting on June the 19th. If you're still reading after surviving this tsunami of data, may your wins always be Uphill Battles, and your losses Consolations.