Riot is quietly rewriting how League handles trolls — and this time the bans actually bite

League of Legends is about to feel very different for anyone who treats ranked like their personal therapy session – and a lot safer for everyone else. Riot has essentially decided that running it down and hard-griefing are no longer just “bad behavior”; they’re being treated like cheating, with automated detection, harsher bans, and systems that can straight-up end ruined games.

This isn’t another “we care about toxicity” blog. Riot is reporting a tenfold increase in daily bans for inting and griefing since patch 25.9, a claimed 99.95% accuracy rate on the new detection system, and more punishment tools queued up, including stronger chat sanctions in patch 26.5 and a vote-to-instantly-end griefed games on the horizon.

Key takeaways

• Riot’s new automated griefing/int detection is already issuing around 10x more bans while tying into anti-smurf and Vanguard/TrueSkill 2 systems.
• Ranked players get more protection: LP refunds in some griefed games, an upcoming “end the game now” vote, and quarantined queues for repeat offenders.
• High-elo players will be judged more harshly, and specific troll patterns (like spamming Anivia walls or Trundle pillars) are directly targeted.
• Riot is betting hard on automation to clean up League’s culture – and the line between “creative play” and “griefing” is where this could blow up.

Riot is finally treating griefing like cheating

For years, League’s worst games fell into the same black hole: someone runs it down, or spends 20 minutes sabotaging fights, and the only response is /mute all and a vague “we’ll review your report.” Riot’s latest behavior update quietly flips that script.

Since patch 25.9, Riot has rolled out a much more aggressive automated detector for “antijeu” – intentional feeding, griefing, and outright sabotage. Internal live tests show daily bans jumping by around 10x as the system picks up patterns like:

• Repeated intentional deaths with no real damage dealt or objectives gained.
• Ability spam that clearly harms teammates: Anivia walls trapping allies, Trundle pillars to body-block escapes, etc.
• Obvious “run-it-down” routes and AFK-in-disguise behavior where a player mechanically moves but refuses to participate.

Riot claims a 99.95% accuracy rate on this system. Even if you take that at face value, at League’s scale that still means thousands of decisions every day where the system has to be right. The point, though, is not “perfect justice” — it’s volume. Riot is finally using the same industrial-scale tooling it uses for anti-cheat and smurf detection on the people who just want to burn your LP.

This isn’t happening in a vacuum either. The griefing detector is wired into a wider integrity stack: Vanguard on PC, experimental TrueSkill 2 matchmaking tests, and new smurf controls like account linking and LP refunds. Riot’s basically admitting what players have known for a decade: it doesn’t matter how good your MMR system is if half your losses come from someone else’s meltdown.

More bans, but also more protection for everyone else

The interesting part is how Riot is structuring the fallout. This isn’t just “ban more people and hope queue times survive.” The new tools change how ranked risk works for the average player.

First, LP handling. In games where Riot’s systems detect a clear griefing or inting outlier, LP can be refunded or partially protected for the innocent side. That means fewer nights where one troll nukes a promo series and Riot’s answer is a copy-pasted support ticket.

Second, the “ruined game ejector seat.” A future patch in Season 2 will add a vote that lets your team instantly end a match when Riot’s detection flags extreme disruptive behavior. According to early details, if the vote passes:

• The griefed team keeps their LP and hidden MMR.
• The enemy team still gets full LP as if they won normally.
• The offender — and any premade partners — eat the LP loss and face a separate punishment on top.

Crucially, this isn’t just a fancy surrender. The system has to detect something clearly game-breaking before the vote even appears. That’s Riot trying to avoid weaponized dodging — premades shouldn’t be able to spam this button because their draft went sideways.

Third, Riot is done letting repeat ranked griefers hide in Normals when queue bans hit. Future changes route serial offenders into a separate ranked queue rather than just shoving them into other modes. That’s the closest League has come to a “prisoner island” for toxic players: you still get to play ranked, but mostly with people the system thinks are like you.

And because Riot finally accepts that context matters, high-elo players will be held to a stricter standard than Bronze or Silver. The logic: if you’re Challenger, you know the difference between contesting a wave and sprinting it down. The margin of “benefit of the doubt” shrinks the higher you climb.

Automation has teeth — and edge cases

Any time a massive live-service studio says “don’t worry, the algorithm is 99.95% accurate,” the next question should be: accurate compared to what? Riot says it’s focusing on extremely clear-cut cases first, but League has a long history of things that looked like trolling until they won a meta.

We’ve seen this movie: Singed support roaming, Smite top laners, funnel comps, weird jungle paths that ignore early camps. Half the strategies that ended up on Worlds stages would’ve been reported as griefing in Season 3 solo queue.

Riot’s answer, so far, is that the system tracks intent and outcomes over time. It’s less about one strange game and more about patterns:

• Do you consistently deal near-zero damage while dying repeatedly?
• Are your “creative plays” ever associated with wins, CS, or objective pressure?
• Is your ability spam harming allies more than enemies across dozens of matches?

That’s why off-meta players aren’t automatically doomed the first time they pick something weird. But if you’re spamming troll picks to lose fast or using mechanics solely to screw your own team, the system is designed to catch you, even if you never type a word in chat.

The scarier experiment is the automatic game termination idea Riot floated: a system that can unilaterally end games it considers hopelessly compromised. Community feedback has been icy on that one, and Riot has already walked it back to “under review, not yet implemented.” Players are willing to trust an algorithm to spot outliers; they’re far less willing to let it decide when the game is over.

And then there’s chat. Riot plans additional toxicity sanctions in patch 26.5, with a specific focus on reducing false positives — a hint that their current filters still nuke borderline cases too often. Expect more nuanced penalties: partial chat restrictions, longer lockouts for repeat offenders, and maybe tighter integration between voice, text, and post-game reports.

The tightrope here is obvious. If Riot overcorrects, League stops feeling like a pressure valve and starts feeling like a monitored classroom. If it undercorrects, nothing changes and all of this reads like another annual “we’re serious this time” dev blog.

Why Riot is doing this now

On paper, Riot has cared about “player behavior” for more than a decade. In practice, enforcement has lagged way behind the size of the problem. So why flip the switch in 2025-2026 and start banning ten times as many people?

Part of it is simple economics. Riot has been quietly gutting or downsizing projects that don’t immediately pull their weight — from cancelled experiments to the recent brutal cuts on its new fighter 2XKO. When a company pivots that hard to profitability, suddenly the flagship game’s player retention looks like sacred ground.

League is a mature live service. You’re not doubling player numbers with a new champion or a pretty battle pass. But you can stop bleeding players who are sick of losing 20% of their games to trolls and AFKs. Cleaning up ranked is a growth strategy dressed up as player advocacy.

There’s also the esports angle. The same month Riot is pushing tougher systems inside the client, some of its biggest stars are lawyering up outside it. FANABLE, the agency representing Faker and Gumayusi, is preparing civil and criminal action against fans engaged in targeted harassment — from death threats to funeral flowers. When your competitive scene’s poster child has to get the courts involved, “we muted him” doesn’t cut it anymore.

League’s culture isn’t just a community problem; it’s a brand risk. Publishers, sponsors, and regulators are all paying closer attention to how platforms handle abuse. Riot needs to be able to point to hard numbers — 10x more bans, 99.95% detection accuracy, faster sanctions — the next time someone asks why their game’s chat logs look like they were scraped from a 4chan archive.

In that context, tying griefing, smurfing, and cheating into one integrity stack makes perfect sense. Every time a player feels like Riot actually had their back when a game went to hell, that’s one less reason to uninstall and wander off to a rival live service.

This could actually change ranked culture — if it holds

If Riot delivers what it’s promising, ranked League gets a new social contract.

For most players, the upside is obvious:

• Fewer games hard-ruined by a single bad actor.
• More LP protection when it does happen.
• A way out of hostage situations via instant end votes.
• Less smurf-fueled chaos as account-linking and LP refunds make fresh accounts less attractive.

High-elo solo queue in particular could feel less like a content farm for mental breakdown clips and more like an actual competitive ladder. When the system expects better from top players — and punishes them faster when they cross the line — it nudges the whole culture away from performative inting and towards playing to win.

But there are real questions that only live data will answer:

• How transparent will Riot be about false positives and overturned bans?
• Will the “griefer queue” just become a smurf playground with long queue times?
• Can the instant-end vote avoid becoming a stealth dodge tool in high MMR lobbies?
• Will creative, high-risk strategies quietly die because players are scared of triggering the detector?

If I had Riot’s behavior lead in front of me, the question would be simple: What’s your plan for the 0.05%? Not the obvious trolls — the edge cases, the weird games, the off-meta picks, the players who really are trying and tripping every wire anyway. That’s where trust in this system will live or die.

What to watch next

• Patch 26.5 notes: The exact wording on new chat penalties and any changes to appeal processes will show how confident Riot is in its filters.
• First post-mortem stats: If Riot publishes numbers on bans, false positives, and LP refunds after a few months, we’ll know whether that 99.95% claim holds up.
• Rollout of instant-end votes: Watch how often the feature triggers and whether high-elo streams start using (or abusing) it to manage bad games.
• Queue health in “griefer island”: If separate queues quietly disappear after a season, you’ll know they caused more matchmaking problems than they solved.
• Esports and pro feedback: When pros and big streamers start publicly trusting or trashing the system, ranked culture will follow.

TL;DR

Riot has massively upgraded League of Legends’ automated systems for spotting inting, griefing, and toxicity, leading to around ten times more daily bans and plans for harsher chat sanctions. That means more LP protection, a future instant-end vote for ruined games, and even a separate queue to contain chronic offenders, all wired into the same backbone as anti-smurf and anti-cheat tools. The real test will be whether Riot can keep edge cases and false positives under control — without killing the creative chaos that makes League worth playing in the first place.