Players are roughly three times more likely to dismiss a safer-gambling prompt if it appears during a winning streak than if it lands during a losing one. That's the headline finding from a behavioural study of 4,200 account-holders across four licensed operators, and it points at an awkward truth for anyone designing these messages: the moment a player most needs to think clearly is the moment they're least likely to read anything at all.

Why the win state kills the message

The study, run over 11 weeks in late 2024, split prompts into three delivery windows: after a net loss session, after a net win session, and at a neutral point with no recent swing. Dismissal rates came in at 9.8% for the loss window, 31.4% for the win window, and 12.1% for neutral. Median time on screen before dismissal in the win condition was 1.6 seconds.

That gap isn't really about arrogance. It's about attention. A player mid-run is tracking a number that's going up, and anything that interrupts the arithmetic gets swatted away reflexively. The prompt isn't being rejected on its content — most participants in follow-up interviews couldn't recall what it said. It was rejected because it existed.

There's a second effect worth naming. Winning states tend to compress perceived time. Players in the win condition estimated they'd been playing for a median of 40 minutes when actual session length averaged 71. The prompt arrives into a distorted timeline, which makes any message about "how long you've been playing" read as simply wrong.

What the designers got wrong

Nearly every prompt in the sample used the same structural template: a modal that paused play, a headline referencing the player's session, and two buttons. The dismiss option was consistently the visually dominant one — larger, higher contrast, positioned right.

Three fixable problems showed up in the data:

• Timing by outcome, not by behaviour. Prompts fired on net result rather than on markers like stake escalation or speed of play. A player doubling stakes after three wins triggered nothing.
• Content mismatch. Win-window prompts led with loss-framing language ("you've spent...") that contradicted what the player was experiencing, which made the message feel inaccurate and easy to distrust.
• No persistence. 84% of dismissed prompts never reappeared in the session, so the single dismissal effectively closed the loop.

The counter-evidence on hard interruption

Before concluding that softer prompts are the answer, it's worth noting the opposite result. A 2023 trial on mandatory session breaks found that forced interruptions — no dismiss button at all — cut average session length by 22% but pushed 14% of affected players to open an account with a second operator within the month.

So the choice isn't between gentle and firm. It's between a prompt that gets read and a prompt that gets routed around. Dismissal and migration are both forms of failure; they just fail in different places.

Where this leaves operators

The practical implication is that win-state prompts need to be built differently, not just timed differently. Some operators are testing non-blocking delivery — a persistent on-screen strip that doesn't pause the game — and outcome-neutral language that never references money. Early numbers suggest dismissal falls to around 18%, still nearly double the loss-window figure but a meaningful improvement.

The harder question is what a prompt is actually for. If the goal is a documented touchpoint for compliance, the current design works fine: it fires, it's logged, it's dismissed. If the goal is to change behaviour, then a message that 31% of players swat away in under two seconds is closer to theatre than intervention. Which one a regulator is willing to accept is the argument nobody has settled yet.

If you're checking your own play patterns rather than studying them, most jurisdictions now offer deposit and time limits you can set in advance — the kind that don't need to be read mid-win to work.