Your League of Legends Playstyle Could Expose Account Boosting

A proposed detection system looks for sudden changes in how an account plays—not where or how it logs in.

Published: August 16, 2026
Estimated reading time: 4 minutes

A League of Legends booster may have the correct password, use the normal game client and pass every conventional security check.

What the booster probably cannot do is play exactly like the account owner.

Researchers at Korea University have proposed detecting account sharing and rank boosting through what they call a behavioral fingerprint. Instead of searching for cheat software or suspicious login locations, the method measures how an account normally plays and looks for meaningful changes.

Clever? Yes. An automatic “booster detector”? Not even close.

Building a Fingerprint From Match Data

Everything comes from information already recorded by the game server. It doesn’t need mouse movements, keystroke timing or monitoring software installed on the player’s computer.

The fingerprint combines 12 gameplay features, including champion preferences, position distribution, early-game farming, vision control, kill participation and damage share.

Several measurements were adjusted by position. That matters because a Support player naturally produces different farming and vision statistics from a Jungler or Bottom laner. Without that adjustment, changing roles could look more suspicious than it really is.

The basic premise is straightforward: if a different person suddenly takes control, the account’s gameplay should change with them.

Twenty Old Matches, Five New Ones and 15 Suspicious Accounts

The detector establishes an account’s normal behavior from 20 previous matches and compares that baseline with the five most recent matches. A larger difference produces a higher anomaly score.

The study examined 100 accounts from ranked solo-queue matches on the Korean League of Legends server. Each account had at least 35 recorded matches.

Fingerprints from the same account were generally much closer together than fingerprints from different accounts. The largest reported average within-account distance was 3.293, while the smallest average distance between different accounts was 4.489.

Using an anomaly threshold of 4.62, the system classified 15 accounts as suspicious.

But 15 suspicious accounts do not equal 15 boosters. No verified labels showed which accounts had actually been shared or boosted. The detector found behavioral changes, but it could not determine what caused them.

The 80 Percent Result Is Not Proof

To test whether the fingerprint could spot an intentional shift, ten legitimate account owners changed their normal playstyle for at least five matches. Eight crossed the anomaly threshold.

An 80 percent detection rate for instructed behavioral changes is not an 80 percent success rate against real boosters. The original owners were still playing. An experienced booster might produce a smaller or more carefully disguised change.

Legitimate behavior could also trigger an alert. A player might change roles, learn new champions, return after a long break or adopt a different strategy.

Long-term misuse creates the opposite problem. Because the baseline comes from the preceding 20 matches, a booster who controls an account long enough may eventually become part of its “normal” behavior. The method is better suited to detecting abrupt changes than persistent account sharing.

The authors call it a triage tool. Its score can flag an account for review; it cannot prove who played or justify a ban by itself.

That distinction matters.

This doesn’t show that boosting can now be detected reliably. It shows that players leave consistent patterns in ordinary match data—and that another person may have trouble reproducing them.

A booster may have the password.

They may not have the playstyle.