All Minor and Major Perks in Overwatch 2: The Second Power Layer Rewriting the Competitive Rulebook

Seoul, 2:47 a.m. In the practice room of a tier-two Korean team, the head coach pauses the VOD at frame 143. On screen, a Bastion has just healed itself in the middle of a team fight — an ability Blizzard deleted from the game the day Overwatch 2 launched. The 19-year-old player beside me turns and asks in broken English: "Is that legal?"

It is legal. And that single word is what made every hero power ranking I had built over five years obsolete overnight.

I stayed in that practice room until 4 a.m. Not to watch more VODs. I stayed to answer a simpler question: when a game adds a second layer of power inside the match itself, does the thing I measure through win rates, advanced metrics and market value still measure the right thing?

That was the first time in my analytical career that a new system forced me to rewrite the definition of "roster strength" rather than update numbers in a spreadsheet.

Where the Perks system sits in Overwatch's patch history

Blizzard had never done this before. For nearly a decade, the way they balanced Overwatch was to edit base kits directly: increase damage, reduce cooldowns, change mechanics, or delete an ability outright. Those changes shared three properties. They applied globally, identically to every player in every match. They were publicly documented in patch notes, known in advance. And most importantly for my work, they were relatively static across a tournament.

The Perks system breaks all three at once.

Perks are in-match unlockable upgrades. Each hero has two tiers: Minor Perks unlock at hero level 1, Major Perks unlock at hero level 2. Heroes level up twice per match, and each level offers the player a choice between different options. XP accumulates from three sources: playing the objective, kills and assists, and healing or saving allies.

This point matters more than anything else in the system description: the speed at which you unlock Perks depends on whether you are present in team fights. A player who flanks, farms safely and avoids contact — the player who used to be considered "playing correctly" in many older compositions — will level slower than teammates. And when your team sits at level 2 with Major Perks while you are still at level 1, that gap is no longer skill. It is structure.

Blizzard names this design philosophy "hero uniqueness reinforcement." It sounds generous. But read closely and something else emerges: Perks are a soft patch running inside every single match.

Architecture: two tiers, two level-ups, three XP sources

Separate the system into mechanical components before discussing consequences.

Tier one, Minor Perks. Unlocked at level 1. Lower impact, earlier arrival. In a match with average pacing, both teams can reach this tier at roughly the same time, so it creates limited divergence. The value of Minor Perks lies in shaping direction: the option a player picks at tier one hints at, though does not determine, what they aim for at tier two.

Tier two, Major Perks. Unlocked at level 2. Noticeably higher impact, arriving later. This is where matches are actually decided.

In the system description, two perks are named explicitly, and both belong to the most interesting category: Bastion with Self-Repair, and Orisa with Barrier. I call this group "resurrection perks" — they restore abilities that were removed or weakened during the transition to Overwatch 2.

Anyone who has followed Overwatch for years must pause here. Bastion's Self-Repair was removed because it made the hero nearly impossible to dislodge when positioned correctly. Orisa's Barrier was removed because it turned a dive-oriented tank into a wall-builder. Both were cut for clear balance reasons.

Now they return, but not permanently. They only appear after a player has accumulated enough XP to reach level 2. This is a design that hands back old power conditionally. Technically, it makes the game more stable early and more explosive late. Analytically, it creates a problem I have never encountered before: a hero's strength now depends on timing, not just on the kit.

In other words, a Bastion at minute two and a Bastion at minute twelve are two different heroes. Same name. Same model. Different power.

Hero switching and the game of "perk banking"

The hero-switching mechanic is the part that occupied my thinking longest.

When you switch heroes, your level resets to zero. But there is a compensating detail: the second time you unlock Perks on the same hero within a match, the process is faster. A small rule with large consequences.

In the open workshops I run for the analytical community, we call the potential tactic "perk banking." If a team forces the enemy to switch heroes to respond, it strips accumulated levels from the opponent and forces them to rebuild from scratch while your own team approaches the Major tier.

But the "faster on second attempt" rule opens an inverse path. A team can deliberately switch heroes early, accepting the loss of level 1, to buy a faster unlock later — turning hero switching from a defensive decision into an investment.

I have not seen a team do this in official matches. In scrims, the signal has appeared. If it becomes standard, we will witness a behaviour never seen in Overwatch history: switching heroes not because the composition needs it, but because the Perks timeline needs it.

That is a genuinely new strategic variable. It is also one that current valuation models cannot handle, because it does not exist in any historical dataset.

The XP sources quietly punishing solo play

The three XP sources are: playing the objective, kills or assists, and healing or saving allies. All three share one trait — they reward presence in collective fights.

No XP source rewards creating flank pressure, holding an isolated position to force a response, or stretching enemy lines. Those contributions retain tactical value, but the Perks system does not recognise them.

The consequence is concrete. A Tracer playing well, creating chaos, forcing two enemies to check her — but securing no kills — will level slower than a player standing on the point pressing a heal button. Over time, this may push teams to favour players whose contributions are easier to measure over players who generate indirect value.

I have seen this pattern before. In sports betting analysis, we ran into the same problem when basic metrics like kill counts were used to price players while unrecorded contributions were ignored entirely. The result was mispricing, and those who understood the mechanism could exploit the gap.

With Perks, the mispricing is not in the market. It is in the game itself. That is far more serious.

The level-2 breakpoint is the real win condition

If I had to pick the single most important moment in a Perks-enabled Overwatch 2 match, it would be the instant the first team reaches level 2.

Major Perks are not minor upgrades. They are role-warping changes. When one team has three or four Major Perks active and the other is still at the Minor tier, they are not merely stronger statistically. They have more options. They can open fights in ways the opponent cannot answer with the same toolkit.

Traditional analysis is used to the concept of a composition's "power window" — early, mid or late game. Perks add another dimension: the power window of each individual within that composition, and how synchronised those windows are.

A team can have a strong early composition but a misaligned Perks timeline. Another can be weaker on paper but more synchronised at the level-2 breakpoint. In both cases, a traditional power ranking says nothing useful.

This is why I believe the coming period will produce a new analytical artefact: a Perks pacing map by composition. Not who is stronger, but who crosses the threshold first.

No win rate, no pick rate: a blind market

The Perks system, at this moment, has no data. No win rate per perk. No pick rate. No ban rate. No metric that would allow anyone to rank which perks are strong and which are weak.

In my work, that means every conclusion about Perks right now sits at hypothesis level. Even the ones that sound entirely reasonable.

I want to state this clearly because it directly affects how people should read analyses of new systems, including this one. When there is no data, what we have is not analysis. It is structured inference. Structured inference beats random guessing, but it is not evidence.

Before trusting a number, ask where it was born. With Perks, that question currently has no answer, because no number exists yet to ask it of.

This also creates a practical problem for anyone tracking an Overwatch betting market. When Blizzard inserts a new power layer mid-match, models built on historical data lose part of their foundation. Bookmakers will adjust somehow, but their adjustment will rest on assumption, not evidence. And bookmaker assumptions can be exploited — in both directions.

I am not stopping you from betting — I only want you to understand what you are betting on. If you are pricing a Perks-enabled Overwatch 2 match with an old model, you are pricing a different game than the one being played on screen.

Patch lock at LAN: a familiar crack

This is the largest systemic risk I see, and it has nothing to do with the strength of any particular perk.

LAN tournaments typically lock the game version before an event begins, to guarantee fairness and give teams preparation time. This works well when patches only change numbers — every team plays the same locked build.

With a fast-moving system, matters get more complex. Blizzard has said new heroes will receive Perks, and updates arrive every other season. If a tournament locks an older build, teams may be playing a game that viewers at home no longer recognise. But if a tournament uses the latest build, teams with less preparation time suffer a structural disadvantage.

Both choices carry a cost. Organisers will have to choose.

In esports history, this genre of dispute has appeared many times. Every time a title adds a new system, there is a period where competitive rules lag game design by months. With Overwatch, the gap may be wider because Perks affect individual fights directly, not just the match in aggregate.

I will track this signal more closely than any other. The answer to "are Perks enabled in competitive play?" determines the entire value of any analysis of this system.

Rosters: flex players rise, one-trick specialists fall

This is one of the few conclusions I am willing to state with relatively high confidence, because it derives from mechanics rather than performance data.

The hero-switching mechanic, combined with faster second-time perk unlocks, clearly rewards players who can play multiple heroes within one role. A flexible player can switch in response to the enemy composition while retaining reasonable access to Major Perks. A specialist who plays one hero has no such option — if their hero is countered, they must choose between enduring the disadvantage or switching and losing all accumulated perk progress.

In football, the same logic drives teams to prioritise versatile players in flexible tactical systems. Adaptability has value of its own, separate from raw technical quality.

But there is a downside nobody seems to be discussing. If certain heroes' Major Perks are significantly stronger than others, teams may be pushed into picking heroes not for composition value but to deny an opponent access to a specific Major Perk. I call this "perk baiting" — picking to block, not to fight.

In the short term this may produce compositions that look strange to viewers. In the long term, it could turn composition analysis into a far more complex combinatorial problem than before.

Casters and the audience problem

Overwatch is already one of the hardest games to broadcast. The number of abilities to track in a 5v5 fight is enough to overwhelm a new viewer. Perks add another layer.

Now a caster must not only describe who is where and which cooldowns are available. They must track hero levels across ten players, forecast who is about to hit the Major tier, and explain why that moment matters.

This is an entirely new skill in esports commentary. In my experience, new skills take time to form. In that window, the viewing experience suffers.

I have watched this happen before, when advanced metrics began appearing on football broadcasts. Commentators initially ignored them. Then they recited them mechanically, without explanation. Only years later did a generation emerge that knew how to use metrics to tell the story of a match.

Overwatch has less time than that. And if viewers do not understand Perks, the system will be read as unnecessary complexity rather than tactical depth.

Without an audience, I hear the match breathing. But only when the audience understands what it is watching does that breathing have anyone to listen.

Balance risk: a dark zone with no map

Perks have not been through a full competitive season. Nobody knows which perks will be discovered as overpowered, in what way, or how quickly.

In Overwatch's balance history, the average time for a new system to stabilise is several months, with at least a few emergency adjustments. With Perks, I expect that period to run longer, because the number of variables is far greater. Every hero has two tiers, every tier has multiple options. The combination space may be large enough to make comprehensive testing impossible.

The middle scenario, which I consider most likely: a few perk combinations are found by the community to be abnormally strong, Blizzard disables or adjusts them mid-season, and we get a "patch within a patch." This destabilises teams that invested time optimising around those combinations.

The worst-case scenario: an exploit lets perks generate unintended power in an official match, forcing results to be reviewed. This has happened in esports history. Not often, but often enough to be non-excludable.

This is why I would advise teams to prepare in advance for the possibility that the system changes mid-season. Not because I believe the risk is high, but because the cost of preparation is far lower than the cost of reaction.

Correlation is not causation: a new system does not create depth by itself

A fast-spreading assumption holds that because Overwatch 2 added a new system, the game has become deeper. That Perks are a tactical advance.

That assumption is unproven. And there is a technical reason to doubt it.

A new system increases the number of variables. It does not automatically increase the number of meaningful decisions. These are different things, and the distinction sits at the centre of many esports analysis errors.

If Perks are simply stat upgrades — more damage, more armour, more speed — they do not create depth. They create complexity. Complexity raises cognitive load for players and viewers without raising decision quality. A player only needs to memorise the optimal pick for each situation.

Real depth appears when choices carry genuine trade-offs, when no option is correct in all situations, when players must read the match to choose. That is the criterion I will use to judge the Perks system in a few months.

There is an encouraging signal. Some perks restore old abilities, and that is not a stat bump — it changes a hero's role. Orisa shifts from a dive tank to a wall tank. Bastion shifts from a damage machine to a damage machine with self-sustain. These are structural changes, not arithmetic ones.

But there is also a worrying signal. If most perks are slight variations of the same upgrade type, the system becomes decoration with a high cognitive cost. Blizzard has a design history that leans toward high-impact, ultimate-adjacent abilities. If Major Perks follow that path, they will create real depth.

The Seoul night of 2026 taught me that truth can be lonely, but never wrong. And right now, the truth is that we do not yet know which direction this system will take.

Ghost power and the trap of memory

One psychological dimension of resurrection perks matters more than the technical one.

When players see Bastion's Self-Repair return, they do not see an ability. They see a memory. And that memory has its own power, independent of how strong the ability actually is in the current build.

In betting analysis, this is one of the most common sources of mispricing. Markets price based on memory of how strong something once was, not on how strong it is now. Those who understand this hold a systematic edge.

With Perks, I expect a period in which community perception of resurrection perks diverges from their actual performance. If I am right, that window holds the highest value for data-driven analysis — and is simultaneously the hardest period in which to persuade anyone, because you are arguing against collective memory.

The skill of an analyst in that situation is not being right. It is explaining why you are right in a way that does not insult anyone's memory.

The fan's view

The first feedback I received from Overwatch communities in Vietnam and Korea about Perks was split. One group was excited: finally something new to learn, explore and argue about. Another was tired: the game has changed again, everything must be relearned, the new meta must be chased right after adjusting to the old one.

Both reactions are reasonable. And I do not want this article to flatten them into a single attitude.

To the second group I would say: your fatigue is a data signal, not just an emotion. When a meaningful share of loyal players feels overloaded by change, that is a variable on player retention. And retention, over the long run, affects the quality of the professional ecosystem — the talent pipeline, the audience, and the commercial value of the league.

In other words, fan sentiment does not sit outside analysis. It is part of it.

Community sources

This article draws on Blizzard's published Perks system description, along with discussions in the analytical community I maintain on Discord, where analysts, high-ranked players and some coaching staff share scrim observations.

I want to be explicit about this for two reasons. First, scrim observations are early signals, unconfirmed by official data. Second, my projections about "perk banking" and about the value of flexible players derive from mechanics, not from observed matches.

If you have real data from Perks-enabled matches, I want to hear it. That kind of contribution makes analysis more credible — not louder claims about certainty.

What happens in the next round

Three signals I will track, in order of importance.

First, Blizzard's policy on whether Perks are enabled in ranked and professional play. The answer determines whether everything above has practical value or is merely a theoretical exercise.

Second, the emergence of community-driven perk tier lists. When pick rates and win rates per perk appear, we will finally have a basis to judge this system with evidence rather than inference.

Third, evidence that teams are deliberately switching heroes to manage perk progress. If that appears in official matches, we will know the system has genuinely changed how the game is played at the highest level.

Data does not shout, it whispers — and I have learned to lean in and listen. With Perks, that whisper is still too quiet to hear clearly. But it has begun.