How CSWatch Works

A player profile enters our system whenever it's loaded: someone looks them up by Steam profile, custom URL or SteamID64, checks a whole match or lobby that includes them, or uses the Discord bot. Every profile we load gets a pass through the scoring layer, and accounts that score below 70 are flagged automatically.

Why this exists: a non-trivial fraction of cheaters never get reported because their victims don't know they were cheated. Automatic flags surface suspicious accounts before anyone files a report.

Each profile is scored across seven categories: ban record (VAC, game, community and trade bans, and how recent they are), community reports on CSWatch, account legitimacy (age, CS2 hours, game library, profile setup), skill consistency (performance that doesn't fit the account's experience), rank trajectory (sudden rank jumps), friend network and profile signals. Together they make the 0 to 100 trust score on the profile page. How the score is weighted.

Why this exists: scoring is fast, works on every lookup and shows at a glance which accounts deserve a closer look. It never convicts anyone on its own.

A report needs something to watch: the match share code (we check the reported player was in that match) or a link to the reporter's own clip. We save the match demo, and the team can render it into a video from the suspect's point of view with X-ray, showing every round where they got a kill. Renders queued for reports from Champion and Legend members go first.

Why this exists: without the match or a clip, a report is just an opinion. Evidence lets a reviewer see the suspect's actual in-game behaviour. We don't convict anyone on stats alone.

The case enters the review queue. Overwatch reviewers (players who meet the CS2 hours and skill bar, approved automatically when they apply) watch the evidence and vote cheater or insufficient evidence, optionally with a short note. The suspect's name stays hidden while they review, so the verdict is about the gameplay.

Why this exists: human judgment is irreplaceable for the cases that matter: borderline cheats, novel cheating techniques, and decisions where context matters. Votes are weighted by each reviewer's track record: from half a vote up to two votes, depending on how often their verdicts match the final outcome. The team can remove reviewers who vote carelessly.

A case is decided once it has at least 3 votes and one side holds more than 66% of the weighted vote. A cheater verdict becomes a public conviction (once the match demo is saved); an insufficient-evidence verdict closes the case without touching the suspect's record. Until one side passes 66%, the case stays open. The CSWatch team can also decide a case directly when the evidence is clear, and every such decision records who made it.

Why these thresholds exist: three votes prevent single-reviewer mistakes, weighted votes account for reviewer reliability, and the 66% bar keeps close calls from producing convictions. False-positive minimisation matters because public convictions stick.

Convicted players are marked on their profile page and listed on the leaderboard, and a conviction sets their trust score to 0. Scores are recalculated from current data as new bans, reports and matches come in.

Why this exists: the whole point. Without a public, accountable record, the entire pipeline has no consequence. Convictions exist to inform queue decisions and create deterrence at the margin.