Player Behavior Knowledge Base > Leaverbuster


1. About


Leaverbuster is an automated system designed to identify and penalize frequent leavers and AFKers [193, 206, 358], but not punish for occasional issues [206, 609].

It’s active in all matchmade queues, while custom games are devoid of it [394].


2. Features


The first implementation of Leaverbuster was generally less restrictive than the current iteration, due to Riot erring on the side of caution while punishing players for leaves and AFKs, but player feedback convinced them to set the system a tad harsher, due to players dodging AFK penalties by switching to smurfs [193, 194]. The result of this was the implementation of low priority queue.

The upgrade occurred upon the old Tribunal being taken down for revamp [188], with players getting an one-time notification the first time they leave a match [192, 206]. This pop up requires them to enter an “I agree” message; it’s meant to be both a warning for displaying such behaviour, and a request to discontinue it. This system is based on the psychological concept of “Verbal Commitment” which states that an individual is up to 30% more likely to fulfill a task after verbally and explicitly committing to do so [510, 592], and for Riot’s paramount goal of aiding players in reforming. If players understand what gets them punished with low priority, they’re more likely to discontinue such behaviour [218], also due to social pressure and positive example from their peers [219].


Upon continuing the action of leaving games or AFKing after getting the pop-up [206], players get hit by low priority queue penalties for a set of five games, and the time spent in low priority (which is the time before they get into the actual queue) can be 5, 10 or 20 minutes, according to frequency of offenses [195]. Being in a premade doesn’t help to dodge this penalty; premade mates are notified, and subjected to, the same low priority time as the player being originally punished with such penalty [195]. Penalties are harsher and faster in case of leaving or AFKing during ranked games [180], the reason being that such offenses impact matches more severely [192].


The low priority timer decreases based on quantity of games played without leaving or AFKing [135, 136], which is often signaled with an informative pop up appearing for the reformed player, stating that their “account status has improved” [409].

Low priority queue is not the only punishment; continued and insistent abuse can lead to bans [195]. Actions such as walking around every few minutes, or occasionally clicking into the game to avoid the AFK popup will not aid in bypassing penalties [172, 413].


Further upgrades to Leaverbuster were introduced with the implementation of the /remake command, which allows players to reset a game if one or more players didn’t connect at all. The offending players get punished, while everyone else gets no penalty [134, 303].


Another feature of Leaverbuster is showing the name of disconnected players in the post-game lobby as being red, but it doesn’t occur frequently to prevent players from finding out which AFK behaviour is detected by Leaverbuster [172, 412].


3. Miscellaneous


  • Riot’s example has pushed Valve into implementing features similar to low priority queue and Leaverbuster penalties for Dota 2 [329].
  • Leaverbuster reduced leavers/AFKs at Level 30 by about 19% in NA and about 52% at Levels 1-5 [428, 467], which seems to indicate that most offenses in low levels are due to lack of awareness about the toxicity of this behaviour [428]. However, a slight rise generally occurs when a ranked season is coming to an end [467].
  • Leaverbuster revamp was one of the reasons Riot pushed back the rework on honours [506].
  • The development name for the /remake function was “early abandon feature” [544].
  • 1.5% - 4% of all games start 4v5 [597].
  • Excessive leaving can potentially result in a permanent ban [658]


4. Sources

172 ; Page 318
180 ; Page 315
303 ; Page 19
329 ; Page 57
394 ; Page 291
409 ; Page 265
412 ; Page 269
413 ; Page 269
428 ; Page 255
467 ; Page 117
506 ; Page 228
510 ; Page 233
544 ; Page 63, 121
592 ; Page 169
597 ; Page 182

Created by Coxis on 16-01-2017. Last edited: 28-06-2017