Why alt accounts defeat bans

A ban removes an account. It does not remove a person, and on any game where accounts are free the distinction is the whole problem.

This is not a gap in enforcement that better tooling will close. It is a consequence of what a ban actually is, and servers that understand that respond to it very differently from servers that do not.

What a ban is enforcing against

An identifier. A Steam ID, a licence, a UUID — something the server can recognise on connection and refuse.

Every layer past that is an approximation. IP bans catch a household and miss a phone tether. Hardware identifiers catch a machine and miss a reinstall. Nothing available to a community server identifies a person, because nothing available to anyone identifies a person, and the platforms that try are doing it with far more data than a server has.

So the honest framing is that a ban raises the cost of returning. It never sets that cost to infinity, and the detection techniques worth running are about raising it further rather than about closing the door.

Why the cost matters more than the wall

The useful question is not whether somebody can return. It is how much trouble it is.

On a game with a paid account, a ban costs money to evade, and the population willing to spend it is small and mostly consists of people who were being deliberate. On a free-to-play title the cost is four minutes, and the same person is back before the admin has finished writing the ban reason.

That difference decides what moderation on a server can realistically be. On the first kind of game, bans work. On the second, bans are a delay, and a server that treats them as a solution will spend its life re-banning the same person.

What actually works on free accounts

Making the account worth something. A server where a new connection has no rank, no stats, no build permissions and no voice access has manufactured a cost that the platform does not supply. The evader returns as a nobody, and being a nobody is most of what they were trying to escape.

Slowing the door rather than locking it. Newcomers who cannot do damage in their first hour make evasion pointless, because the thing the evader came back to do is not available to them. This is the same mechanism as an onboarding path that earns access, applied to a problem it was not designed for.

Not announcing the ban. A publicly announced ban is a challenge. A quiet one is frequently just accepted, and a meaningful share of people who would have evaded do not bother when nobody is watching to see whether they managed it.

The trap of escalating identification

The instinct is to demand more: VPN blocks, account age minimums, linked Discord, sometimes verification.

Each of these costs legitimate players more than it costs an evader, because the evader is motivated and the newcomer is not. Blocking VPNs in particular removes a category of ordinary players (people on restrictive networks, people in countries with bad routing, people who simply always have one on) to inconvenience somebody who will use a phone instead.

The ratio is the thing to watch. A measure that stops one evader and turns away twenty newcomers has made the server worse at the thing it was protecting.

What to accept

That some people will come back, and that the server will usually know.

An evader is recognisable within a week, because the behaviour that got them banned is the behaviour they have. The practical response is to ban again, without drama, and treat it as maintenance rather than as a failure, which is a much more sustainable position than the one where every return is evidence that the moderation is not working.

The alternative is a server that hardens itself out of existence, which is a more thorough way of losing the argument.