Abusive behaviour in online games is persistent and studios have tried many approaches. Some work considerably better than others.
What does not work well alone
Reporting systems with slow or invisible outcomes. If a player reports someone and nothing observable happens, they stop reporting.
Permanent bans as the only sanction, which are easy to evade with a new account in free games.
What has evidence
Immediate feedback. Systems that tell a player their message was flagged, before it sends, reduce the behaviour measurably. Many people reconsider when interrupted.
Telling reporters what happened. Closing the loop increases reporting rates substantially, which improves the data everything else depends on.
Graduated sanctions – short mutes escalating to longer ones – which change behaviour more than a single severe penalty that is evaded.
Restricting communication by default for new accounts, which reduces the value of creating a fresh one to continue.
The design dimension
Game structure affects behaviour. Systems that make individual players blameable for a loss produce more abuse than systems that do not. Voice chat produces more than text. Competitive ranking produces more than casual play.
Studios that treat this as a design question rather than a moderation question get better results.
What players can do
Mute early rather than engaging. Report with the specifics the system asks for. And recognise that responding is what the behaviour is seeking.
The honest position
No platform has solved this. The ones doing best have reduced it substantially through many small structural changes rather than one policy.