Why CS2 Has So Many Cheaters: Where VAC Bans Went

The community is arguing about the anti-cheat again: third party trackers show that VAC bans in CS2 have all but vanished, down to single digits where there used to be thousands. Meanwhile Valve locked close to a hundred thousand accounts in August alone. There is no contradiction here. The bans did not disappear, they simply stopped being called VAC. Here is what changed and why cheaters in your games are not going anywhere.

What the numbers show

Take the two August waves from the vac-ban.com index:

• August 6: Valve hit two case farming operations in one night. The index logged 30,630 accounts, and exactly one of them got a VAC ban. The rest were game bans. Frozen inventories were valued at $69,361;
• August 24 to 25: a wave starting at 21:00 UTC. 67,220 accounts with a new game ban and not one VAC among them. Inventories frozen: $132,958.

That is where the "VAC banned a single person" claim comes from. The number is real, but it describes one specific wave rather than every ban in the game, and it signals a change of tool rather than a broken anti-cheat.

Game ban versus VAC ban

The difference is simple and constantly confused:

• a VAC ban comes from automation that caught prohibited software in memory. It is permanent, flags the profile publicly and blocks access to secured servers;
• a game ban comes from Valve itself for breaking the rules of a specific game. That covers not just cheating but farming, botting and drop manipulation. Also permanent, but issued by a decision rather than a signature match.

Trackers count these categories separately. When Valve moves enforcement into game bans, the VAC counter empties out even though more accounts are being locked than before.

Why Valve moved away from VAC

Since May 2026 the heavy lifting in matches is done by VACnet, a system that analyses player behaviour instead of scanning memory for a cheat signature. In August Valve took the next step and opened an invite only portal where selected reviewers label match footage to train VACnet. Human review is back in the loop, not as complaint moderation but as a training set for the model.

The logic holds up. A signature based anti-cheat fights specific cheat builds and loses every time a new one ships. Behavioural analysis catches the play rather than the program: impossible flicks, inhuman reaction times, shots through smoke at an invisible target.

Why nothing changed in your matches

This is the part that usually goes unmentioned. The August waves did not target cheaters in matchmaking. Both hit farms: accounts grinding cases and drops through bots.

One detail from the August 24 to 25 breakdown says it all: eight out of ten banned accounts had been registered within the last twenty months. Those are not profiles of real players with years of history, they are fresh shells created for farming.

Hence the feeling that nothing improved. Valve is cleaning up the game economy, farms, bots and drop manipulation, while the cheater in your Premier game sits in a separate queue of work. The two wars run in parallel and are not directly connected.

How much can you trust these numbers

None of the figures above are official Valve statistics, because the studio does not publish ban data. Trackers like vac-ban.com build estimates on their own index of Steam accounts, 183.5 million tracked IDs in this case, and say so plainly: "tracker estimates, not official Valve totals".

That means two things. The order of magnitude is right: tens of thousands of locked accounts and a shift toward game bans show up across several independent indexes. The exact numbers, though, are always a lower bound, since the index keeps counting accounts it reaches later. The August 6 breakdown puts it directly: this is "a floor, not a ceiling".

There is a separate trap in stats services that count bans among their own users. That sample says nothing about the game as a whole, because someone who connects a tracker to their account is by definition not an average player.

What a player can actually do

Your leverage is limited but real:

• reports work as a training signal. Labelled matches are what feed VACnet, and without a stream of reports there is nothing to train the model on;
• do not expect an instant ban. Behavioural detection accumulates evidence, so punishment arrives in waves rather than the same evening;
• check whether the cheat invented itself. A share of "cheater" complaints turns out to be lag and desync. If you are shooting at a target and missing, read our packet loss guide, and if the picture stutters, the FPS breakdown.

One clarification: the fake VAC ban exploit we covered has nothing to do with these statistics. It cancels a match without banning anyone.

Frequently Asked Questions

Why does CS2 have so many cheaters if Valve bans tens of thousands?

Because the August waves hit case farms and bots rather than cheaters in matchmaking. Eight out of ten banned accounts were created within the last twenty months, which makes them farming shells, not the players from your games.

Did CS2 stop issuing VAC bans entirely?

As a separate category they have nearly vanished from the stats: in the August 6 wave exactly one account out of 30,630 got a VAC ban. Enforcement moved into game bans and VACnet, so the VAC counter on trackers runs empty.

What is the difference between a game ban and a VAC ban?

VAC is issued by automation for prohibited software found in memory, a game ban comes from Valve for breaking the rules of the game, including farming and botting. Both are permanent, but they are issued differently and counted separately.

Can tracker data be treated as official?

No. Valve does not publish ban statistics, and trackers build estimates on their own Steam account indexes and say so openly. The order of magnitude is right, the exact numbers are a lower bound.

Sources

• CS2 Ban History, index of ban waves and anti-cheat updates, data on the August 6 and August 24 to 25, 2026 waves and the VACnet labeling portal