Self-Exclusion Works, But It Leaks

What self-exclusion gets right

Self-exclusion is the decision to ban yourself. You put your name on a list, and the gambling operators that participate are supposed to turn you away — close your accounts, stop the marketing, refuse new ones. It is one of the best-evidenced first steps a person worried about their gambling can take, and if you are choosing one thing to do this week, this is the one to do.

The evidence for it is real. A review of nineteen studies of land-based self-exclusion programs found gambling-reduction rates ranging from 29% to 92% among people who enrolled, alongside lower problem-gambling scores and better wellbeing. Earlier reviews describe the same pattern: for the people who use them, these programs tend to cut how much someone gambles and how bad they feel.

It is not magic, and the research is honest about that. The only randomized controlled trial in this area — about 2,500 online gamblers in France — found no significant effect on total losses, and only a modest, temporary dip in deposits. So the accurate claim is narrow and worth keeping: self-exclusion helps many of the people who use it. This piece is about what happens after you have done that right thing and find you can still reach something. It is written for the person trying to stay excluded, not the person looking for a way around.

The boundary is the licence, not the behavior

Here is the single most useful thing to understand about every self-exclusion scheme: it binds the operators enrolled in it, and in practice that means licensed operators. The list has reach over the companies a regulator can point at. It has none over the companies it cannot.

The UK's national scheme, GAMSTOP, says so plainly: it “can only assist to exclude you from ... participating Gambling Operators” and is “unable to help prevent you from accessing gambling services that are operated by organisations which do not participate.” Since March 2020 every operator licensed in Great Britain has to participate, so “participating” and “licensed” mean the same thing — and you can exclude yourself for anywhere from six months up to five years. The boundary of the licence is the boundary of the block.

“

Self-exclusion works within the walls it was built for. Its edges aren't a failure of the idea — they're the boundaries of one registry, one jurisdiction, one clock.

That is not a flaw buried in the fine print. It is how the tool is built, and it is a reasonable design. But it means everything outside the licensed perimeter is unreached by construction — and there is a lot outside it.

The seams between schemes

Start with geography. The UK has one national list. The United States has no mandatory national equivalent: each state runs its own list, and each list only binds the operators that state licenses. Someone self-excluded in one state can still open accounts with operators licensed only in another.

Even the rules that do exist vary widely. A 2024 review scored state betting regulations against 82 recommended standards and found that even the strongest states — Connecticut, New Jersey, Virginia — matched just 49. That is a count of rules on the books, not a safety score; the report is explicit that there is no pass or fail. But it is direct evidence that “self-exclusion” means something different depending on where you live.

It can fragment inside a single state, too. Pennsylvania keeps four separate self-exclusion lists — casino, iGaming, video gaming terminals, and fantasy contests — and enrolling in one does not enroll you in the others, so excluding yourself from casinos leaves online play open. The clocks differ as well: New York offers terms up to lifetime and drops you automatically when they end unless you renew, while GAMSTOP will not switch off on its own at all.

The gaps are being closed, slowly. In May 2026 Ontario launched BetGuard, a centralized program that — as of July 2026 — lets a player self-exclude from every regulated igaming site in the province at once, including the government-run OLG site, in a single request. It is a meaningful step, though still bounded by the provincial border. In the US, a voluntary national program has been rolling out since 2024, but it is opt-in and still expanding state by state. Progress is real; a single seamless net does not yet exist.

The products no registry reaches

The bigger leak is not between schemes. It is the whole categories of gambling that no self-exclusion scheme was ever built to cover, because they sit outside the licensing regime it runs on.

Unlicensed offshore sites are the clearest case. Because self-exclusion only binds licensed operators, offshore sites are outside it by design — and, as the UK regulator warns, unlicensed sites also lack the other protections of the licensed sector: age verification, responsible-gambling tools, and any real dispute resolution. The same is true of crypto-settled casinos, which are, as of July 2026, typically offshore and unlicensed — a new payment rail on the same old gap. (Naming specific sites here would only signpost them, so I won't.)

Then there are the products that dress gambling in a different legal costume. Dual-currency “sweepstakes” casinos run slots and table games under promotional-sweepstakes theories rather than a gaming licence, which is exactly why they are not part of any state self-exclusion program. As of July 2026 multiple state attorneys general have started treating many as illegal gambling; New York's identified 26 platforms and called them “illegal, dangerous, and can seriously ruin people's finances.”

Prediction and event-contract markets are the fastest-moving example. They are regulated federally by the Commodity Futures Trading Commission as financial derivatives — the CFTC's own consumer page says they are “federally regulated and under federal law can operate in all 50 states” and never mentions state gambling — so a state gambling self-exclusion has no authority over them. The National Council on Problem Gambling calls trading these contracts “functionally gambling, regardless of how it is legally defined,” while noting they sit under federal financial regulation, not state gambling frameworks, so the existing problem-gambling infrastructure “does not neatly apply.” The jurisdiction fight is genuinely unsettled as of July 2026; what is not unsettled is that a state self-exclusion list has no hook on these venues today.

At the far edge sits the brokerage app. There is no gambling self-exclusion that applies to a trading app, because the activity was never inside the gambling regime — even when the behavior looks identical, and even though some brokerages now offer the same event contracts. Addiction specialists have pointed out that self-exclusion and deposit limits, standard in casinos, often simply do not exist on investing apps. None of this means all trading is gambling. It means the boundary is the regime, not the behavior.

Why bans leak even when they fit

Even inside the perimeter a scheme does cover, three human seams remain.

The first is that almost nobody uses it. A 2023 meta-analysis put lifetime self-exclusion at roughly 0.26% of adults and about 15% of people with a gambling problem, with awareness near one in ten. A tool can only protect the people who opt in, and most never do.

The second is that a signed ban is not a wall. Self-reported breach is common: the nineteen-study review found breach rates from 8% to 59%, and in one Swedish national-scheme survey 49% of self-excluders said they had gambled anyway at some point, most often at online casinos. Those are self-reported figures, and researchers think they undercount — but they are not small. And when people do breach, they tend to move to whatever is not covered: in a Swedish clinical sample of treatment-seekers, continued gambling happened most often on unlicensed sites. (That sample was small and skewed toward severe cases, so read it as direction, not prevalence.)

The third is enforcement. A ban only works if someone checks, and online there is often no one and nothing to check against — researchers across seven jurisdictions note there is no mandatory identity verification across all gambling, which makes a truly universal self-imposed ban impossible to enforce. Add expiry to that — most schemes are time-boxed, and a meaningful share of people return once the clock runs out — and you have a protection that is real, but porous.

Close the seams, don't skip the step

The honest conclusion is not that the tool is a false promise. It plainly helps the people who use it. The conclusion is that one list, in one jurisdiction, for one stretch of time, cannot carry the whole load — so the move is to stack layers around it, not to skip it.

That is not a sales pitch; it is what the charities that run self-exclusion tell people to do. GamCare, which operates the UK's National Gambling Helpline, describes self-exclusion as “an essential part of your support toolkit; however, it may not work on its own,” and its TalkBanStop service deliberately layers three things: GAMSTOP self-exclusion, helpline support, and Gamban blocking software. Blocking software is the layer that reaches the sites a registry can't — the offshore operator, the sweepstakes casino, the brokerage — because you name the domains yourself instead of waiting for a licence to cover them.

A device-level blocker like GuardianBlock fits into exactly that narrow role: friction on the domains you declare off-limits, on supported Windows browsers, with a guardian in the loop. It is not a substitute for self-exclusion, not treatment, and not a promise that every product is covered — it is one layer for the seams the registries leave open. Self-exclude first. Then close the gaps.

If gambling is becoming a problem, self-exclusion and professional support are the front line. In the United States, the National Problem Gambling Helpline is free, confidential, and available 24/7 — call or text 1-800-MY-RESET. It is not a crisis line; if you feel unsafe or might hurt yourself, contact 988 in the US or your local emergency services now.