Methodology
How we test the CSWatch score
What the score is made of, which signals we tested on our own data, what they predict, and what we dropped because it did not hold up. Numbers as of 23 October 2026.
What the score is
Every profile gets a score from 0 to 100 built from seven parts. A low score means an account is worth a closer look; it is never a verdict on its own. A cheating verdict on CSWatch only comes from Overwatch review of a real match or clip.
- Hard rules on top. An Overwatch conviction sets the score to 0. A VAC ban or a FACEIT ban for cheating keeps it at 40 or lower (high risk), whatever the rest looks like. A VAC ban can come from another VAC game, which is why it means high risk rather than a verdict.
What we tested, and what came out
Aim that does not fit the account
Used, each costs a little.We checked 6,542 players with Leetify stats, 410 of whom had a VAC or game ban. Extreme averages barely separate them: most players with a very high aim rating are clean. The three signals we use, elite aim with very little utility, a headshot accuracy above 40%, and elite aim with under 300 hours, each show up about 2.3 times as often on banned accounts (14 to 15% of flagged players had a ban, against about 6% of all).
A sudden rank climb
Used, heavy only in combination.Across 52,504 profiles, 6.4% had a ban. Of the 51 accounts with a real rank jump (a big Premier or FACEIT rise in a short time), 9 had a ban: 17.6%, about 2.8 times as often. That is a small group, so a rank jump weighs heavily only together with an aim signal.
Thresholds we used to have
Removed on 22 October 2026.Our first skill checks used guessed thresholds (headshot rate, reaction time, spray, K/D). On 52,496 stored profiles they fired about 50 times in total, so they said almost nothing about anyone.
One great match
Not a signal.In the last six Leetify matches of clean players, an ADR of 150 or more shows up in 9.4% and a K/D of 3 or more in 24.7%. A single standout match is normal.
No FACEIT account
Not a signal.In our data, players without a FACEIT account are not banned more often than players with one.
Hidden hours
Unknown, small deduction.When a library or playtime is private we cannot see CS2 hours. We count that as unknown, never as zero hours, because a hidden profile is not a new one.
FACEIT as a second opinion
We looked up the FACEIT ban record of 5,808 players we hold FACEIT data for. FACEIT had banned 150 of them: 4 for cheating, 114 for smurfing or ban evasion, and 32 for behaviour such as verbal abuse.
- Cheating. Four cheating bans are too few to tune any signal on. What they did show: all four scored 92 to 97 with us, because we did not look at FACEIT. A FACEIT ban for cheating now counts like a VAC ban.
- Smurfing. Our score did not tell the 114 smurf accounts apart from clean players (a median of 95 for both). Account age, Steam level and hours did not either; the best single signals were only about twice as common. So we do not guess a smurf label: a profile says "Smurf account (FACEIT)" only when FACEIT banned it for that.
- Behaviour. Shown on the profile as a fact, not counted as cheating.
What happens next
Bans often come months after the cheating, so the best test is time. Every week we check about 14,000 players with match stats for new VAC and game bans. In December 2026 we compare the signals with the bans that came after, re-weigh the score on that, and update this page.
Think something on your own profile is wrong? Contact us and the team looks at it. For the full pipeline from lookup to conviction, see how CSWatch works.