Riot Is Deleting Mel's Reflect — This Is a Retention Patch, Not a Balance Patch

One second of one match. Seraphine's ultimate is flying at Mel, and the moment Mel's W goes down it turns around and charms the entire enemy team. Or Jhin fires his fourth shot with perfect timing, and it comes back and lands on Jhin's own head.

Riot has now written, essentially off those two scenes: such situations will no longer occur. That one sentence contains the real purpose of the whole patch.

I am telling you this is not a balance patch. It is a retention patch wearing balance clothing. And I will tell you how I could be wrong at the end — first I need to show why I think it.

Let me be clear about what is actually changing, because when five changes land together people usually only see one.

The W is the headline. Mel's reflect is being removed entirely. Instead, W will destroy incoming projectiles, deal a small amount of damage, and grant a shield and movement speed to Mel and nearby allies. The passive loses early-game damage and duration and gains HP and armor scaling. Q's projectile gets slower, and damage shifts from the first explosion toward subsequent hits. E's root duration drops at close range and rises at max range. R's cooldown goes up and its level scaling comes down.

Riot says these land in patch 26.22, roughly three weeks after the announcement. I have doubts about that patch number — it does not match the current season's numbering convention. It could be a transcription artifact, or a forward-dated 2026-season patch. I am treating it as given, but I am writing it in soft pencil until it is verified.

Core point one: Riot is not citing a win rate for this decision.

Their stated reason is that the mechanic is too frustrating to play against. That is an experiential argument, not a statistical one. There is not a single number in the announcement — no win rate, no pick rate, no ban rate. Based on my years of watching matches, this kind of reasoning usually gets applied to mechanics that are balanced on paper but feel unfair in practice. Which means the pre-patch win rate probably was not the trigger at all.

An old habit helps here. When I wrote about Paulinho's transfer in 2026, I learned that what a club says is not its feelings, it is its incentive design. I followed the money until it became a mirror. I am doing the same thing with this patch: reading what the announcement says against what it needs to be true.

Core point two: the W change is not a nerf, it is a role migration.

The default reading is that removing the reflect is removing power. But the new W grants a shield and movement speed to Mel and nearby allies. That is the loudest line in the entire package. It is a support-adjacent clause. Mel stops being purely a solo-lane burst carry and becomes a champion who can build something for her team.

If anyone reads this as a simple nerf, they are missing the most important part of the package.

Core point three: the ally shield is the invisible compensation that escapes scrutiny.

I have a long-standing bias — massive signing-on fees for free agents are more toxic than transfer fees, because they slide past the core of financial scrutiny. This package has the same pattern. Everyone will talk about what is being removed, the reflect. What is being added — the ally shield — is free power, with no disclosed numbers and no verification. If Mel underperforms on live servers, the first dial Riot turns will be that shield's value. That is the row in the spreadsheet nobody goes to look at.

Core point four: the most pro-relevant change in the package is E, not W.

Root duration drops at close range and rises at max range. This is the most elegant part of the package. It trades free safety for a reward for correct spacing. But it has a reverse side — dive champions who used to struggle against a reflect-holding Mel now find her much cheaper to dive. At close range, E no longer saves her.

I have watched matches since I was a kid, and one pattern keeps repeating: picks whose safety depends on the opponent making a mistake generate disproportionate value in coordinated play. Mel's reflect was exactly that kind of tool — the more predictable the opponent, the more it was worth. Reflecting a key ultimate in a teamfight does not just deal damage, it rewrites the story of the whole fight.

Core point five: removing that tool hits high MMR harder than low MMR.

In low elo, nobody lands max-range E roots perfectly, and nobody reflects key ultimates. High elo and coordinated play extracted disproportionate value from exactly those two patterns. So the net effect of this package is a nerf at the top of the skill-and-coordination curve and roughly neutral-to-positive at the bottom.

Core point six: there is real risk of a split between patch lock and live patch.

Riot's standard habit is to freeze the tournament server on a designated patch for the duration of a split or international event. If patch 26.22 lands near a lock date, pro teams may never get to play the new Mel until the following split — while solo queue and content creators are already fully adapted. That produces the familiar gap between public perception and professional reality.

Core point seven: five abilities changing at once has a cost nobody measures.

Mel players must relearn Q's projectile speed, E's root timing, and an entirely new decision tree for W — previously W meant holding the opponent's cast, now W means starting the fight on your own terms. Historically, champions receiving multiple simultaneous ability reworks show a two-to-four-week statistical trough. That is not in the article; it is my own observation.

Core point eight: Mel specialists lose, mid-lane generalists gain.

This is a relative change, not an absolute one. No individual player gets stronger or weaker — only one champion's market value moves. Anyone who kept Mel as a pocket pick to counter an opponent's key ultimate will probably drop her rather than relearn her, because the reason to pick her no longer exists.

There is a counter-signal, though. A champion that can flex across mid and support always has higher draft value, even at lower individual power. That is a basic rule of draft economy. If Riot is deliberately making support Mel possible, this rework is not weakening her, it is widening where she can be played.

Core point nine: no rule is being broken here, but the transparency gap is obvious.

I have an old complaint about referees — when decisions are not explained in the stadium, the fans become the only ignored party. In balance governance, Mel's players stand in exactly that spot. Riot writes the rules, earns the money, and is also the court of appeal. There is no independent body to take this decision to.

One thing does give me pause. Riot says it tried to preserve the reflect, then gave up. That is a legitimacy-building narrative — the decision is presented as the product of reluctance, not haste. It usually means at least one internal prototype was built and rejected. The published version is therefore not the first draft but the survivor of an internal argument.

Core point ten: financially this is a publisher decision, not a club decision.

No club owns a champion. The transmission from champion rework to club P&L is essentially zero. But the publisher's maths is different. Skin sales correlate with pick rate and with how fun a champion feels to play. Removing a signature mechanic can reduce player attachment to that champion — that is retention economics, not balance. And if Mel is tied to the Arcane IP, that raises the stakes for Riot specifically. That link is not in the article; it is my outside knowledge, so I am keeping it as an inference.

Now here is how I will be wrong.

First, I am assuming the reflect was worth more at pro level. It may have been purely a solo-queue drama mechanic that no coordinated team respected. In that case this whole piece is a post-mortem of a problem that did not exist.

Second, the ally shield may turn out strong enough that Mel is effectively buffed, just in a different role. Then my retention theory is wrong, and the truth is much more ordinary: Riot shipped a boring balance fix that I turned into a mystery.

Third, I may be pointing at the system too much and undervaluing player experience. For Mel mains this is not politics, it is losing a champion they loved. I went looking for a culprit and found a spreadsheet with feelings — but I regularly forget the person sitting behind the spreadsheet.

One prediction, and I am putting it in my ledger.

After patch 26.22 goes live, Mel's win rate will drop for two to four weeks, a hotfix will land on the ally shield value, and the champion will settle into a lower-variance, lower-emotion identity. In solo queue, support Mel's pick rate will rise before pro play follows. If I am wrong, that goes in the ledger too — because predictions nobody remembers are not predictions, they are just noise.