Sombra Becomes Support, Roadhog Loses the One-Shot Combo: How the Season 5 Patch Reshapes the Overwatch 2 Meta

The last hack of that match is still the thing I remember most clearly. I rewatched the VOD 47 times — not because the play was beautiful, but because I could not understand how an action that dealt zero damage could decide an entire teamfight. Sombra appeared at the edge of the screen, locked out both enemy supports' abilities for a few seconds, and vanished. Not a single point of damage. Not a single elimination. The match was already decided.

At that point, in professional play, Sombra was banned in nearly every match. Banning her was not a counter-strategy — it was the precondition for the match to unfold according to its own design. I spent a long time wondering whether the community ever realized it was banning a character precisely because she made the game more honest.

Season 5 of Overwatch 2 wipes that state away entirely. Sombra no longer sits in the Damage role. She moves to Support.

This is the third time a character in Overwatch 2 has changed roles completely, but the previous two did not come with a dual restructuring. Alongside Sombra, Roadhog was also reworked on a comparable scale. Two characters, two opposing directions, one patch.

What makes this one different is not the two characters themselves. It is the Perks system — a new customization layer that lets players attach additional effects to individual abilities. Sombra receives CTRL ALT ESC, Life Hack, Data Packet and Cybersecurity. Roadhog gets Piggy Piggy, Toxic Exhaust and Scrap Hook Minor.

In other words, the patch opens an extra strategic axis rather than simply adjusting numbers. And when two characters are reworked simultaneously, the professional community must absorb two new variables within the same preparation window. The history of large-scale reworks always shows the same pattern: early-season periods carry a higher upset rate, because the advantage held by top teams is compressed when everyone has to relearn from scratch.

On Sombra's side, most of the old mechanics remain: Translocator, hack, and EMP. The effects were redesigned to fit the support role, but the ability structure itself is unchanged. The detail most coverage skips is that EMP remains fully intact.

That is the pivot point. EMP was once the highest-impact ultimate in the Damage category. When it exists inside a Support kit, Sombra becomes a priority target at the first pick slot — not because she deals damage, but because she can neutralize an entire enemy formation in a single beat.

A support hero carrying a team-wide ultimate is the kind of character that every ban system has to recalculate from zero. Sombra did not get weaker. She changed where she stands.

Roadhog moves in the exact opposite direction. The one-shot combo — the thing that defined him for multiple seasons — is gone. Scrap Gun now fires in two bursts instead of one. The new alt-fire, Trash Compactor, absorbs enemy projectiles and fires them back as an explosive ball, operating much like D.Va's Defense Matrix. Scrap Hook Minor gets a buff: it reloads four ammo instead of two.

This is a value swap, not a power reduction. Roadhog loses the ability to finish kills quickly but gains a tool for teamfight participation. His individual stats will drop. His contribution to the team will rise. For anyone reading a pure stat sheet, that is a visual illusion — and it is the most common distortion I encounter when reading reports on character reworks.

I cross-checked two sources — the official ability description sheet and community VOD records — and found no change to Scrap Gun's base damage. Roadhog did not get weaker. He simply stopped working alone.

There is an assumption spreading through the community: Sombra was "demoted" from Damage because she was too strong. That is reading correlation as causation.

Being permanently banned does not equate to being overpowered. It can equate to a character design that conflicts with the structure of the match itself. Hack disables abilities, turning every teamfight into a reflex test rather than a tactical test. Blizzard did not reduce Sombra's power. They moved her into a space where the ability to neutralize is treated as the core value, rather than a side effect of a damage kit.

Before trusting your eyes, check what your eyes already decided to believe.

The real risk of this patch is not Sombra or Roadhog. It is the Perks system. When every character has multiple customization options, the team with the deeper analytics department optimizes for specific matchups. In BO3 and BO5 formats, Perks can be adjusted between games. That is an entirely new preparation dimension, and teams without analytical staff lose ground before the match even begins.

At the same time, there is a personnel gap few are discussing. Players who specialized in Sombra as a damage dealer must relearn the character from scratch, because the new role changes positioning, teamfight timing and strategic value. Those who can leverage their existing mechanical familiarity — Translocator, hack rhythm, EMP timing — will hold a significant edge over players learning from zero. With Roadhog, the edge tilts toward veteran tanks with teamfight experience, players who have already piloted defensive characters with similar mechanics.

During a window of meta instability, the advantage of strong teams is compressed. Every team must learn two new characters at once, in two different roles. This is a window for early-season upsets.

Numbers never panic — people do, and people are the variable.

But it must be said plainly: this analysis lacks win-rate, pick-rate and ban-rate data. Without them, every conclusion here is a grounded inference, not a conclusion. I spend 30 percent of my working time cross-checking data, and that 30 percent tells me both what is true and what cannot yet be asserted.

The signal to track in the next cycle is not Sombra's win rate. It is the upset frequency in the early season. If that number rises and then falls quickly, the meta has stabilized. If it holds flat, the problem is not the patch.

Two things never lie: data and time.