Sombra Leaves Damage, Roadhog Loses the One-Shot Combo: Overwatch 2 Season 5 and the Preparation Problem Nobody Has Solved

3:12 AM, and Nobody in the Chat Room Could Answer a Basic Question

My Discord channel still had 41 people online. An assistant coach in North America typed a single line into the general channel: "Sombra support just wiped our scrim block." Below it, a Korean data analyst replied: "How many minutes, how many maps?" Nobody answered. People just pasted a screenshot of the scoreboard.

I stared at that message for about two minutes. In thirteen years of covering this industry, I have learned to tell two kinds of noise apart. The first is community noise after any buff or nerf — loud, short, and self-extinguishing within forty-eight hours. The second is the noise of working professionals when they cannot find a common denominator to compare against. That night belonged to the second kind.

Sombra was once the name banned in nearly every match in the Damage role. Now she is a support. Roadhog was once a kill machine built on a single combo. Now he is a tank who holds the rhythm of a teamfight. Two heroes, two reworks, one patch.

Before you trust a number, ask where it was born. I tape that line to the top of every report I send to the analysis desk. In this case, the only number we actually have — "perma-ban" — has no clear measurement origin at all. It comes from the collective memory of the community, not from any statistics table I can open and verify.

That is why this article exists.

What This Patch Actually Says, and What It Actually Does Not

Overwatch 2 Season 5 arrived with two reworks at once. That is rare. Major updates usually split changes apart so the community has time to digest them one at a time. Not this time.

Rework one: Sombra moves from the Damage role to the Support role. She becomes the third character in the game's history to change roles. Her abilities operate similarly to her old version but carry new effects. The most notable point: EMP stays the same.

Rework two: Roadhog loses his one-shot kill combo. The Scrap Gun now fires in two bursts. He gains a new alt-fire called Trash Compactor, which absorbs enemy projectiles and fires them back as an explosive ball — a mechanic close to D.Va's Defense Matrix. The Scrap Hook Minor Perk is buffed, reloading four ammo instead of two.

Sitting on top of both reworks is the Perks system. For Sombra, players have options such as CTRL ALT ESC, Life Hack, Data Packet, and Cybersecurity. For Roadhog, there is Piggy Piggy, Toxic Exhaust, and Scrap Hook Minor.

That is everything the patch says.

What the patch does not say matters just as much. There is no win rate. No pick rate. No ban rate. No matchup data. No specific publication date for individual changes. And no information about which tournament, team, region, or competitive format this patch will directly affect.

At my firm we have a rule called the "silence threshold." If an analysis has more than three important missing metrics, we downgrade every conclusion to "hypothesis" and state that clearly at the top of the document. The Season 5 patch crosses the silence threshold about four times over.

That does not mean this patch is unimportant. It means anyone telling you they know exactly what this patch will do to the meta is selling you something that is not data.

Sombra: When a Hero Is Removed from the Field and Handed Back Through a Different Door

This is the structurally heaviest change in the patch.

Put simply: in a hero shooter, tuning numbers changes something inside the tactical space. Changing roles changes the tactical space itself. A Damage hero moved to Support drags consequences across three layers at once.

The first layer is composition. Teams that built strategy around Sombra Damage now have to throw away almost all of that understanding. Not adjust it — throw it away. The fight framework changes, positioning changes, engagement timing changes, and most importantly, damage output expectations change.

The second layer is the ban-pick phase. Sombra sat in the near-permanent ban group in the Damage role. When she moves to Support, ban pressure does not disappear — it relocates to a different category. Teams that used one ban on Sombra Damage now have to decide again where that ban goes, while another team may be preparing to deploy Sombra Support as a card nobody has prepared to answer.

The third layer is Perks. The fact that Sombra has her own Perk set in the new role means teams must rebuild her optimization table from scratch. CTRL ALT ESC, Life Hack, Data Packet, and Cybersecurity do not operate identically in every situation. Choosing which Perk for which matchup is a pre-match tactical decision, not an in-game reflex.

The detail I consider most important, and also the one most easily missed: EMP stays the same.

If you have followed Overwatch long enough, you know EMP is one of the heaviest ultimates in the game. Keeping it intact while changing roles means Sombra retains her strongest fight-swinging tool, but in a role where expectations for personal damage are far lower. This is the kind of change analysts call a "value shift": the same tool, placed in a context that makes it cheaper to own and more expensive to ignore.

I spent nearly two hours that week rewatching old VODs of matches where Sombra was banned. What I wanted to test was whether the ban pressure came from damage or from EMP. The answer I found was not enough to conclude anything, because the sample was too small and no statistics table confirmed it. But it was enough to make me doubt the popular assumption that Sombra was banned because she dealt too much damage.

Roadhog: Losing What Made Him Feared, Gaining What Makes Him Useful

Roadhog's rework has a far clearer logic than Sombra's, and is also easier to evaluate.

The one-shot kill combo is gone. That is a direct nerf to his ability to generate fast kills. For Roadhog players of the lone-hunter school, this is a major loss in feel — and in tactical value too, because turning a five-on-five fight into a four-on-five in an instant was what defined his role before.

In return, Roadhog gains three things.

First, the Scrap Gun fires in two bursts. This changes his damage rhythm: instead of a single explosion, the player gets a damage curve split across two beats. In extended fights, the double beat is usually more stable. In burst fights, the double beat is slower.

Second, Trash Compactor. The mechanic of absorbing enemy projectiles and firing them back as an explosive ball puts Roadhog in the same family as tanks with active defensive tools. Comparing it to D.Va's Defense Matrix is a grounded comparison, and it says something important: Trash Compactor is a defensive tool, not an offensive one. It does not help Roadhog kill people. It helps Roadhog protect teammates.

Third, Scrap Hook Minor is buffed to four ammo reloaded instead of two. This is a small numerical change with large rhythm consequences. Doubling the reload amount means Roadhog can sustain pressure for longer without leaving position to reload. In a game decided by who holds position longer, that is not small.

Taken together, Roadhog shifts from a lone kill-hunting tank to a fight-rhythm tank. This is a change in kind, not in degree.

And it creates a question nobody has data to answer: will Roadhog players of the old school actually convert to the new school? Mechanically, they lose nothing. Mentally, they lose almost everything.

Perks: A New Tactical Layer, and Also a New Preparation Layer

The Perks system appears in both reworks, but its impact reaches beyond these two heroes.

The principle is simple: when you add a customization layer to the hero selection process, you turn one pre-match decision into two. And when the number of decisions rises, the value of preparation rises exponentially.

For Sombra in the Support role, CTRL ALT ESC, Life Hack, Data Packet, and Cybersecurity open four different build directions for the same hero. For Roadhog, Piggy Piggy, Toxic Exhaust, and Scrap Hook Minor create three. Multiply two heroes together and you get a combination matrix a team must prepare for — not one composition, but compositions multiplied by Perk configurations.

This is the point I consider most important in the entire patch, and also the point most overlooked by mainstream coverage.

In a BO3 or BO5 format, the ability to adjust Perks between games is a genuine tactical weapon. A team that loses game one to a bad Perk configuration for the matchup can fix the error in game two without changing composition. Conversely, a team that wins game one with a specific Perk configuration must prepare for the opponent reading that configuration in game two.

This creates a new kind of advantage: analytical advantage. Teams with dedicated data analysis staff will exploit this layer faster. Teams with thin coaching staffs will have to cope with intuition, and intuition is a good tool for reacting but a poor one for preparing.

Based on my experience tracking matches and scrim blocks across many seasons, I have noticed a fairly stable pattern: every time a new preparation layer is added, the gap between the leading group and the mid-table group tends to widen for the first three to six weeks, then narrow again once the rest of the field copies the method. If the pattern holds, we will see the same with the Perks system.

But I must be clear: that is an observation from experience, not a conclusion from data. I have no statistics table to prove the pattern. And by the strict rule of my analysis desk, an observation from experience should only be used to raise a hypothesis, never to assert one.

The Fan Angle: Two Halves of the Same Reaction

That night I opened a short survey in my Discord channel. 214 people responded within six hours.

On Sombra, the reaction split almost exactly in half. The supportive group said they had never been able to play Sombra in the Damage role because she was always banned, and the move to Support finally unlocked a hero they loved but could not touch. The opposing group said they lost a Damage hero they had spent hundreds of hours mastering, and the role change made those hours worthless in skill terms.

Both groups are right. This is something I have to remind myself every time I read community feedback: two opposite feelings can coexist and both be reasonable, if they come from two different player groups with two different play histories.

On Roadhog, the reaction leaned clearly to one side. Most respondents were long-time Roadhog players and expressed disappointment. The main reason was not numbers, but feel. The one-shot combo is a ritual experience: you wait for the exact moment, you strike, and you are rewarded with a kill. Removing it does not just reduce the hero's power, it removes a ritual.

Without an audience, I hear the breathing of the match. And in a video game with an audience, that breathing usually comes from things not on the stat sheet. It comes from a player losing the thing that made them feel good.

I recorded both reactions in my report, in a separate section, under the heading "Qualitative data." That is what we call the things we cannot measure but cannot ignore.

Where the Story Gets Complicated: Correlation Is Not Causation

Here I have to say the thing many people in this industry will not want to hear.

A hypothesis is spreading very quickly through the analyst community: that reworking two heroes at once creates a "meta chaos window," and inside that window, weaker teams have a higher chance of producing upsets, because every team must adapt to two new variables simultaneously.

The hypothesis sounds very reasonable. It is also very easy to sell. And it has one problem: nobody has proven it.

To test this hypothesis you need at least three things. You need the upset rate before the patch as a baseline. You need the upset rate after the patch over a long enough period. And you need to rule out other variables — roster changes, schedule, format, competitive server version.

We have none of those three. No pre-patch upset rate. No post-patch upset rate. And no information on which tournament, format, or timing this patch will be applied to.

I understand why the hypothesis is attractive. It gives people a story to tell, and a story is always more comfortable than the silence of data. But if you are using this hypothesis to place a bet, you are betting on a story, not on a probability.

I am not stopping you from betting — I just want you to understand what you are betting on. And what you are betting on right now is an unverified hypothesis, with one very specific layer of risk: if the meta is not as chaotic as predicted, you will lose money because you believed a story with more appeal than evidential weight.

There is a second problem, and it is more serious.

The term "perma-ban" that the entire community is using to describe Sombra does not come from a dataset. It comes from many people remembering that they saw Sombra banned a great many times. Collective memory tends to retain emotionally intense events and forget ordinary ones — an effect widely documented in cognitive psychology research. That means "perma-ban" could be accurate, nearly accurate, or exaggerated. There is no way to distinguish without a ban-rate table.

And when you build an analysis on an unmeasurable premise, every conclusion downstream carries that error with it.

What I Do Not Want to Repeat

In 2026, after Korea beat Germany in Kazan, I wrote an analysis pointing out that the home team's expected goals were lower than the opponent's, and that the victory came from fifteen minutes of late pressing. The article was right on the data. It also got me called a traitor to a historic win, and I cried.

The night in Seoul 2026 taught me that the truth can be lonely, but never wrong. I still believe that. But I also learned something else: a truth presented without empathy will be rejected, no matter how correct it is.

Three years later, at Euro 2026, I wrote a piece comparing a star's pressing numbers with a midfielder from the champion team. The article got my company's page attacked, and I thought about deleting it. Instead, I held a public Q&A, opened all the raw data, and admitted the star was still the best player of the group stage.

The lesson I brought into this article is very concrete: when analyzing a hero the community loves, I must state their strengths before presenting numbers. For Sombra, the strength is EMP — an ultimate that retains its value after the rework. For Roadhog, the strength is the ability to hold fight rhythm, which the rework is actually trying to build rather than destroy.

And I must clearly note that data can change sooner than readers think. With a patch that has no numbers yet, this is many times more true.

So What Is Actually Worth Watching?

I have no conclusion about the strength of Sombra Support or the new Roadhog. Nobody does, at this point.

But I do have a list of things that will give us the answer, and their order of priority.

First is Sombra's pick rate in the Support role in official matches. This is the first metric that will tell us whether teams genuinely treat her as a viable option or just a scrim experiment.

Second is how teams defend against EMP in the new context. If EMP retains its weight and Sombra is now cheaper to own, we will see teams forced to allocate resources against an ultimate coming from a role they are not used to guarding against.

Third is the level of adoption of Trash Compactor in real fight situations. A defensive tool only has value if teams are willing to use it at the right moment, and that depends on whether players trust it.

Fourth is the frequency with which teams adjust Perk configurations between games in a BO3 or BO5 series. If that number is high, we are witnessing a new tactical layer actually working. If it is low, the Perks system will become decorative detail.

Fifth, and the one I care about most, is the upset rate in the early period after the patch. Not to confirm the meta chaos hypothesis, but to see whether that hypothesis survives its first contact with reality.

Data does not shout, it whispers — and I have learned to lean in and listen. It will take a few weeks before that whisper is clear enough to separate from the noise of the community. During that time, the most honest thing an analyst can do is say they do not know.

And the teams? They do not have that option. They have to decide right now, with less information than what we currently have on the table. That is the thing I have always respected about this profession, and the thing that keeps me awake until three in the morning reading messages in a Discord channel whose members will mostly never know my name.