Mel loses her mirror: Riot deletes one frustration and leaves another untouched

In a match recording I have watched many times over, Jhin fires his fourth shot at Mel. The bullet turns around and buries itself in his chest. In another game, Seraphine casts her ultimate, Mel presses W, and the entire enemy team is charmed by Seraphine's own ultimate. Riot Games calls Mel's W "the most frustrating ability in League of Legends." After multiple rounds of adjustment, they decided to delete it from the game. What made me stop while reading the announcement was not the new mechanic. It was the piece of the old mechanic left behind. Mel's W still destroys every projectile that flies at her. The frustration of "I fired and nothing happened" remains fully intact. Riot only removed the louder version of it: "my spell killed my own teammate." Patch 26.22 does not fix a system flaw. It fixes a viral clip. Here is what changes, concretely. W no longer reflects projectiles; it destroys them and grants a shield plus movement speed to Mel and nearby allies. The passive loses early-game damage and duration in exchange for health and armor scaling. Q shifts damage away from the first hit and into subsequent hits. E reduces root duration at close range and increases it at maximum range. R gains cooldown and loses per-level scaling. Reading the whole package, I see a power curve bent in a new direction. Mel moves from a reactive punish mage in the mid lane to a late-game utility champion. This is a rework-level change to the slot, not a numbers nerf. And this is where I split from most commentary: the crowd says Mel was nerfed, I say Mel was made to change profession. The biggest loss is not in the stat line. It sits with the cohort of players who invested in her. Her competitive value rested on a reactive read skill: deciding whether to hold W for Seraphine's ultimate or Jhin's fourth shot, and for how long. Deleting reflect deletes that skill outright. This is the harshest class of balance change — not a numbers nerf, which players can adapt around by shifting their play pattern, but a skill-obsolescence event. Mel players do not lose a few percentage points of power. They lose the thing they spent hundreds of hours learning. Riot says it wanted to "limit situations that are too frustrating when playing against this champion." It also says it tried to preserve the mechanic across multiple rounds of adjustment before giving up. Put those two statements side by side and the real problem appears: a mechanic with binary power — reflect or do not reflect — cannot be balanced with numbers. There is no parameter that sits between "blocks the ultimate" and "does not block the ultimate." That is why it was removed. It is also why nobody should believe it will be stable after this patch. People call me a traitor, but I am only loyal to the numbers. And here the numbers do not exist. No win rate, no pick rate, no ban rate, no actual frequency of the reflect plays that swing games. Riot offers two examples — Seraphine and Jhin — and both are among the most dramatic plays a champion can produce. That is salience-based sampling, not frequency-based sampling. A balance decision driven by standout clips rather than frequency data always carries the risk of over-correction. Based on my experience tracking champion reworks across many seasons, the same pattern repeats: when the design team does not publish frequency, they are implicitly saying the problem is feel, not frequency. Feel is the right thing to fix. It is not the right thing to measure. There is a layer of information nobody states but that reads clearly from how Riot handled this. E reducing root duration at close range and increasing it at maximum range is a very clear signal: they pay for keeping distance and tax the ability to save yourself at point-blank range. Combined with W now granting a shield and movement speed to allies, the whole kit reads as a deliberate push toward the support role. A passive compensated with health and armor is also a defensive signal — surviving a dive, not winning a duel. At the commercial layer, a champion branded by its own publisher as "the most frustrating in the game" is a discounted asset. A depressed play rate drags down skin attach rate. Riot spending design budget on a rework-level change to exactly one ability slot is consistent with protecting a monetizable asset. I flag this as inference, not disclosed information. But here is where I could be wrong, and I will state the conditions. In professional play, losing solo-lane power while gaining positional flexibility can be net positive. A champion who can play both mid and support is an asset in pro drafts because it defers the role reveal until the opponent has committed. Flex picks are always worth more than a strong but positionally fixed champion. Then there is timing. Patch 26.22 lands in the late regular season, after the World Championship patch-lock window. Pro teams get an entire preseason to adapt. A rework-level change landing in that window causes far less shock than the same change mid-split. The patch ships globally at once, so no region gets earlier access — regional differences, if any, come only from adoption speed and stylistic habit. And I could be wrong here: if Mel's win rate does not recover across three patches, if she vanishes from both mid and support, then the shield and movement speed did not compensate for the damage taken away. The threshold I am watching is not solo-queue win rate, but professional draft presence and how long she stays on the map before being swapped out. The crowd shouts about a deleted ability. I am waiting to see Mel's first professional appearance after the patch. And I will bet she walks on from the support role, not mid lane. I do not write to be loved, I write to be right — later, when the dataset is long enough to judge both sides.