27 February, 2019

17 Min Read

State of Balance: Season 8

State of Balance - Season 8

Hello Warriors! Welcome to this installment in our ongoing series of State of Balance, where the Fight Team analyzes the results from the just-completed season, and discusses where we are headed in the future.

Introduction

What is the "STATE OF BALANCE" about

This page shares some data about Win Rates and Pick Rates of our characters with you, the community. It’s also used to discuss our perspective on certain balance issues, and to describe a bit where we see it heading over the next seasons.

This installment focuses on the State of Balance for ‘Season 8: Marching Fire’, which ranged from October 16, 2018 to January 30, 2019.

You can read previous State of Balance posts here for ‘Seasons 7, 6, 5, and 4’.

What data is used to define "BALANCE"?

We use multiple sources to get information on the actual state of balance in the game. For example, we look at such sources as:

  • High Level and Casual Level Player Data (console, PC)
  • User Research studies (surveys, user testing)
  • Community feedback collected across multiple channels (Reddit, Twitch, Ubisoft forums, YouTube, Twitter, Facebook)
  • Top players feedback
  • Tournament Streams
  • Our Dev Team

How much should we rely on these sets of data?

The state of balance suggested by these sources may conflict with each other – this is normal. For example, in the collected player data we’ll show in this post for Duels, Warlord is positioned at the bottom – however, top player feedback puts Warlord at the topThis is because most players use only Locked combat in Duels, but the top players play Warlord very differently and include an Unlocked strategy focusing on ‘Crashing Charge’. When we consider the gathered data’s Win Rates in this post, we are most interested in using them like the proverbial ‘canary in the coalmine’ – to let us know when we may have a problem. It’s not critical to us that these stats show 50% across the board. Instead, we examine each one, then try to make the changes that make the most sense. -In Warlord’s case, in ‘Year 3: Season 1’ we buffed Warlord’s locked combat, and nerfed Crashing Charge. To best understand what actually happens in-game, all these sources have to be considered against each other on a case-by-case basis.

BALANCE CHANGES FOR 'SEASON 8'

Major changes to start 'Season 8'

We'll highlight what the major changes were that led into 'Season 8: Marching Fire'

Note: 'Season 8' was the final Season of Year 2

  • Wu Lin - we added 4 new heroes, which had a significant effect on the meta: Tiandi, Shaolin, Nuxia, and Jiang Jun.
  • Breach – we added a new game mode with new objectives requiring new strategies.
  • Perks – replaced Gear Stats, meant to improve the clarity of the system, remove abusive builds, and better compliment play-styles.
  • Advanced Tactics – meant to allow players to train to the full potential of their characters.
  • Nerfed Feats – particularly Traps and AOEs.

Major change during 'Season 8'

We’ll highlight one major change in mid-season as it nerfed the move that many felt defined tournament play in Dominion.

  • Nerfed Shaman’s ‘Predator’s Mercy’ – when the victim is hit by an external attacker, we reduced the amount of time the victim is vulnerable before recovering defense.

Data - 'Season 8'

Methodology

This data is gathered by tracking player matches from:

  • Games played during the entirety of ‘Season 8’ (October 16, 2018 to January 30, 2019)
  • The combined Console and PC player populations
  • Reputation 1+ players only
  • All players in a match must be in the same MMR bracket

We will show data based on these skill ratings:

  • Duel: Full Population/Platinum or Above Ranked Players
  • Dominion: Full Population/Top 4% (Based on Skill Rating)
  • Breach: Full Population/Top 4% (Based on Skill Rating)

We chose these skill thresholds (Platinum-or-Above for Duels, and Top 4% for 4v4 Modes) as they allow us a large enough population of upper-skilled players to show relevant data for all cells in the matrix. If we make the skill threshold too narrow, we can’t fill up each cell with hundreds or thousands of matches to be useful.

Full Population represents all players. We include this for the first time because it’s another data set we consult often, and so that all players can see themselves represented in one or more of these data sets.

Based on this data, we will show:

  • Win Rate
  • Pick Rate
  • Finally, we will also include the community-created "Tier List"

To be clear, the community-created Tier List isn’t managed or controlled in any way by the For Honor team – it is a subjective list created by top tournament players (currently it is owned by the tournament player ‘Setmyx’). We hope that comparing this against the gathered datawill foster beneficial discussion about balance between the Fight Team and the community.

Note 1: in this State of Balance, we use the version of the Tier List that was posted 2018-December-27, which was the final version of the Tier List that was posted in ‘Season 8’. All the Tier List images included are taken from the version on that date.

This should allow everyone to see themselves in one grouping or another: Full Population, Platinum-or-Above, Top 4%, or community tournament contenders.

Now that we all are on the same page about what data is about to be presented – let’s see the data!

DUEL DATA

 

S8DuelWLFull

S8DuelWLPlat

S8DuelSetmyx

Duel Win Rate

  • Jiang Jun is at the very top of the Full Population and Platinum-or-Above matrices. The supposition is that the match-up is tough to learn for many players.For example, many players find Sifu Stance makes Jiang Jun reasonably safe most of the time, and find 500ms lights that do 20 damage difficult to deal with. However, Jiang Jun is in the middle of the community Tier List, and not listed as tournament-winning viable. This represents the largest disconnect between the Full Population and the community Tier List. For the Fight Team, our answer is that we need to consider all these data sources, and weigh them depending on the situation. These disconnects illustrate why we consider many sources of data. We believe it’s important to consider tournament-level balance, but also the feel of fairness throughout the skill spectrum. Due to this analysis of multiple sources during the middle of ‘Season 8’, the Fight Team made certain changes to Jiang Jun (fixed animations on the heavy strikes, and nerfed ‘Dou Shi Choke’) to increase the feeling of fairness and readability for the Full Population without negatively impacting Jiang Jun’s position on the community Tier List.
  • Warlord is another major disconnect between the Full Population and the community Tier List. In Warlord’s case, we saw that the Full Population had little trouble beating him, while the community Tier List focused on the fact that Out Of Lock’s sprint unblockable was very strong. Here, the tournament-level players are maximizing an Out Of Lock playstyle, which appears Full Population doesn’t use at all. On the Fight Team, we don’t want the Duels to be based on Out Of Lock playstyle, so for the start of ‘Year 3: Season 1’ the Fight Team changed Warlord by buffing the Locked combat kit, and nerfing ‘Crashing Charge’.
  • On the Fight Team, we also recognize that part of the disconnect is the extreme efficiency of the playstyle of tournament-level players as opposed to Full Population playstyle. Warlord’s case shows that tournament-level players are incorporating Out Of Lock where it succeeds, and in Jiang Jun’s case (and many other cases), they incorporate Unlock Rolls to safely avoid mix-ups without any risk. As the Full Population doesn’t use Unlock Rolls whenever it’s safe to do so, the Full Population is taking risks in situations that the tournament-level players fully avoid for free. These additional dangerous situations that the Full Population attempts to defend against seems to be a big source of the difference between the perception of viability of certain heroes. In the future, we’d like to see that tools like Out Of Lock unblockables and escape tools such as roll nerfed. We believe this will make mix-ups more interesting for everybody. As an additional benefit, tournament-level and Full Population play would look a little more similar (in that they’d all be encountering similar situations, and it’d be their different responses that determined the difference in skill).
  • The top of the Platinum-or-Above matrix features the 4 Wu Lin, Conqueror, and Berserker. If you squint between this matrix (which is gathered data), and the Tier List (created by tournament-level community members), there are generally some similarities with most of these heroes.
  • The bottom-performing heroes for Full Population and Platinum-or-Above both feature the same set of Warden, Gladiator, Highlander, Nobushi, Warlord, and Peacekeeper. So there’s some consistency between those matrices – although it doesn’t match the Tier List at all.
  • Aramusha’s another case. Gathered data still has him performing reasonably highly, even with the re-balancing at the start of ‘Season 8’, but he’s at the bottom of the Tier List. For reference, in ‘Season 7’ Aramusha was 4th in Win Rate for the Platinum-or-Above players, while still bottom of the Tier List so the re-balancing hasn’t seemed to change his placement much for any population set.

S8DuelPickrateFULL

S8DuelPickTOP

Duel Pick Rate

  • Pick Rate is extremely similar between the Full Population and Platinum-or-Above players.
  • Generally, the working theory is that Pick Rate has more to do with the general fantasy of the character, rather than actual Win Rate or the Tier List. Players have always seemed to strongly identify with Warden, Orochi, and Kensei,and adopted Centurion readily.Now many seem to see the Shaolin as strongly representing their imagined monk-type warrior.
  • Perhaps the Pick Rate has more to do with playing a favourite character, rather than perceived power-level for many.

DOMINION DATA

S8HeroWinDominionFULL

S8WinDominionTop

S8SetmyxDominion 

Dominion Win Rate

  • All 3 sets now have at least 1 similarity: Jiang Jun’s at or near the top. Easily-accessible Unblockables and extraordinarily-strong feats (Soothing Mist) seem to be the general reason in group fights, although the 500ms lights seem to give the Full Population trouble (which would also help account for his power in Duels for Full Population and Platinum-or-Above).
  • Shugoki and Peacekeeper, for example, end up near the bottom of all 3 lists. Warlord’s pretty low on all as well. All 3 heroes got a buff in ‘Year 3: Season 1’. We’ll see how that goes for them.
  • Shaman is no longer S Tier for the community Tier List. It seems that our targeted nerf has been successful at nerfing her without deleting her.
  • Order is very interesting. If you look at Nobushi and Shaman, their rating increases noticeably as you go from Full Population > Top 4% > Tier List. It seems the higher-level you play, the more they bubble to the top.
  • However, we still see Aramusha providing a huge contrast between the gathered data and the community Tier List. For reference, in ‘Season 7’ Aramusha was 2nd in the Dominion Top 4% Win Rate, and still last in the community Tier List so it appears the re-balancing hasn’t changed the state much.

S8DominionPickFull

S8DominionPickTOP

Dominion Pick Rate

  • For the Full Population, the Pick Rate in Dominion seems to line up reasonably with the Duel Pick Rate.
  • For the Top 4% however, the Pick Rate starts to shuffle a bit more.

BREACH DATA

S8BreachWinFULL

S8BreachWinTOP

  • Generally, when compared to Dominion, the Breach Win Rate is very similar.

  • Nuxia conspicuously performs better in the Breach data than Dominion. It’s a little unexpected as Assassins seem to generally have a more difficult time due to survivability of reflex guard heroes against Pikemen, but perhaps the Caltrops makes up for that in her case.

  • Raider’s also interesting, because the theory on the Fight Team was that Raider would perform better in Breach data than Dominion because of ease of Pikemen-clearing. While we do see the heroes that are strong against Pikemen or Commander/Guardian performing well in Breach data, they seem to perform just as well in Dominion data (see Raider, Jiang Jun, Aramusha, Kensei). Perhaps the ability to play against high-damaging 500ms attacks is more important in the Full Population or Top 4% than efficiency against the objectives.

    • Note that we write “seem”, since Breach has only been around for one Season, and here we don’t have massive differences in data between it and Dominion - so hopefully these charts spark discussion.
  • Note: One omission we want to note here is that there is NOT a current community Tier List for Breach – so it’s not included in the ‘State of Balance’ for that reason.

S8HeroPickBreachFULL

S8HeroPickBreachTOP

Breach Pick Rate

  • Pick Rate generally seems to be similar between Dominion and Breach.

NEXT STEPS

'Year 3: Season 1'

We’ve made 2 major changes with the launch of ‘Year 3: Season 1’. We buffed Shugoki, Peacekeeper, and Warlord, and we introduced the Black Prior.

SHUGOKI, PEACEKEEPER, WARLORD

The balancing buffs to Shugoki, Peacekeeper and Warlord were meant to bring them out of the non-viable tiers. We expect a difference to be felt for all 3. We do recognize that the biggest problem facing Shugoki at this time is the ease that opponents can roll-away and avoid 100% of Shugoki’s mix-ups options. This was apparent to us before the buff as well. However, this is a problem that affects many mix-ups for many heroes, andthe fix must be a global one. We’d like to reduce the efficiency of roll-away, unlock back-walk, and backward strafe. These are actively being tested internally.

BLACK PRIOR

Black Prior has also sparked a number of discussions with most agreeing Black Prior is at the very least viable. The Bulwark Counter and Feats seem to give a really interesting dynamic to team fights, which is what we were hoping for. In terms of fight prowess, we have recently released Patch 2.05 that reduces the Out Of Stamina pressure of the Black Prior (removing the guaranteed Tenebris Rising after a Light Attack Opener). The other main point players note is that the shield attacks are quite safe in terms of recovery. The big question there is ‘should players always be awarded free damage for reacting?’ Not sure the answer should be ‘yes’ to that. If all attacks in the game are highly punishable, they’ll have to be very-low-risk and/or very-high-reward to be viable. There should be a spectrum to the ‘safe-ness’ of all attacks. In the Black Prior’s case, we’re keeping a very close eye on if we want to nerf recovery time, nerf damage on Light Finisher, or not make a change.

Mid-way patch of 'Year 3: Season 1'

At the mid-way patch of ‘Year 3: Season 1’, we want to change a couple of things:

SOOTHING MIST

Needs a nerf. Currently, it’s one of the strongest feats in the game, and it’s not a hard decision on when to cast it (since the cooldown is super-low for super-high benefit). This will nerf Jiang Jun in 4v4 modes, which is where we feel he needs it most. Full Population data, Top 4%, and Dominion Tier List all have this character at the top, and a Soothing Mist nerf will address this character’s power level without furthering the disconnect on the Duel data and Tier List.

OUT OF LOCK MELEE

We nerfed Warlord’s Out Of Lock melee attack because we don’t want Duels to be focused on Out Of Lock play-styles.Shinobi will see the same kind of change. We’ll compensate Shinobi for this nerf with the ability to Feint the Zone Attack.

REVENGE GAIN

We will remove the ability to generate Revenge meter during 1v1 fights. We feel that the current revenge gain in 1v1s prevents strong 1v1s characters from killing opponents before they can pop Revenge, and makes weird situations where you actually don’t want your friend to ‘help’ you (since they’ll just feed revenge to the opponent). The current state of Revenge gain also promotes ganking rather than helping against it (since efficiently ganking-to-kill an opponent before they can actually pop revenge is the strategy). The difficulty will be in finding a value that allows the Full Population to still get Revenge when they should, but not let tournament-level players stall out fights too long.We will continue to examine this once we make a change to Revenge Gain.

Closing note

We would love to hear more from you on the Full Population data, Platinum-or-Above data, Top 4% data, and the community Tier List. Please remember that we look at all this data, information, feedback, and a whole lot more when we make our decisions.

Keep sending us your feedback, and we look forward to sharing the next installment of the State of Balance at the beginning of ‘Year 3: Season 2’!

  • The Fight Team