Max Bet Warnings Cut Rebuy Spikes 31% Only Above €5 Stakes

A session-level cap on maximum bet size reduced the frequency of rapid rebuys by 31% among players staking above €5 per spin, according to operator-side data from 14 Croatian-facing brands covering roughly 2.8 million sessions between January and June 2024. Below that €5 threshold, the same intervention moved rebuy frequency by less than 4% — a figure within the noise band of the study's own confidence interval. The finding matters because it suggests the popular "max bet warning" model, which most Croatian operators deploy as a blanket prompt, is doing almost nothing for the players who generate the bulk of low-stake, high-frequency deposit churn.

What the Data Actually Shows

The dataset came from a voluntary data-sharing arrangement between operators running under Croatian licence and two independent analytics vendors. Sessions were tagged by average stake per round, deposit velocity in the 30 minutes following a loss, and whether a max bet prompt had fired. The 31% figure applies specifically to the cohort averaging €5 or more per spin — about 11% of total sessions but 34% of gross gaming revenue in the sample.

That revenue concentration is the reason this number is politically awkward. The players most responsive to the intervention are also the ones operators have the least commercial incentive to interrupt.

Break the cohort down further and the picture gets messier:

Average stake band	Rebuy reduction	Share of sessions	Share of GGR
Under €1	2.1%	61%	19%
€1–€4.99	3.8%	28%	47%
€5–€9.99	27.4%	7%	21%
€10+	38.9%	4%	13%

The jump between the €1–€4.99 band and the €5–€9.99 band is not gradual. It looks like a step function, and that's the most interesting thing in the whole dataset. Either the €5 line corresponds to something real in player psychology — a self-imposed threshold where a warning reads as a genuine interruption rather than background noise — or it's an artifact of how the prompt was designed.

Why the Step Function Probably Isn't Random

Three explanations hold up better than the others.

First, stake size correlates with intentionality. A player spinning at €0.50 is often in a passive, extended session — the kind of play where a pop-up is dismissed reflexively. A player at €10 is making a deliberate, higher-consequence decision each round, and a warning lands differently when the next spin costs more than a coffee.

Second, the €5 band overlaps heavily with players who have already self-classified as higher-risk through their own behaviour: larger deposits, faster deposit sequences, fewer session breaks. These players have more to lose from a single rebuy, so the marginal deterrent value of a prompt is higher.

Third, and this is the one operators rarely say out loud: below €5, the rebuy behaviour isn't really about the bet at all. It's about deposit friction — how quickly a player can top up after a loss. A max bet warning does nothing to address that. The intervention that would move the needle there is a deposit-velocity limit, not a bet-size prompt.

The Croatian Regulatory Context

Croatia's gambling framework, administered through the Ministry of Finance and the local regulator, has tightened considerably since the 2020 amendments to the Law on Games of Chance. Mandatory responsible gambling tools now include session time displays, deposit limits, and self-exclusion registers. Max bet warnings, however, remain a matter of operator discretion rather than a statutory requirement.

That discretion has produced inconsistency. Walk through the account settings of the ten largest Croatian-facing operators and you'll find max bet prompts triggered at wildly different points — some at €2, some at €10, some only after a loss sequence of three or more rounds. There is no shared standard, and the data above suggests the threshold matters more than the existence of the prompt.

The €5 line is not currently enshrined anywhere in Croatian regulation. It emerged from the data, not from policy. That's worth flagging because it means any operator adopting it is making a commercial choice, not a compliance one.

The Problem With Blanket Prompts

Most operators implement max bet warnings as a universal setting — it fires for everyone above a nominal floor, usually €1 or €2. The logic is defensible on its face: warn broadly, let players self-select. But the data shows this approach dilutes the intervention. When a prompt fires dozens of times per session for low-stake players, those players learn to dismiss it without reading. That dismissal behaviour then carries over into the sessions where the prompt actually matters.

There's a behavioural term for this — warning fatigue — and it's well documented in other domains, from medication labels to aviation alerts. The iGaming industry has been slow to apply it. A prompt that fires 40 times a session is not a warning. It's wallpaper.

The counter-argument is that segmenting prompts by stake looks like targeting, and targeting looks like a compliance risk. If you only warn high-stake players, you're effectively admitting that low-stake play is harmless — which is not true, and which the €1–€4.99 band's 47% share of GGR should immediately complicate.

What the 31% Figure Doesn't Tell You

The study measured rebuy frequency in the 30 minutes after a prompt fired. It did not measure:

• Whether players who reduced rebuys simply moved to a different operator
• Whether session length increased to compensate for lower rebuy frequency
• Whether the reduction persisted beyond the first week of exposure
• Net revenue impact, which is the number every operator will actually ask about

That last omission is the one that determines whether any of this gets implemented. A 31% reduction in rebuy spikes among high-stake players sounds like a meaningful harm-reduction win. It also sounds like a revenue line item. The study's authors were careful not to publish net revenue figures, and the operators who shared data did so under terms that kept GGR impact out of the public version.

This is not a conspiracy. It's how industry-academic data-sharing works. But it means anyone citing the 31% figure as proof that max bet warnings "work" is overstating what the data supports. The warnings change behaviour in a specific cohort. Whether that behaviour change is net positive for the player, net negative for the operator, or roughly neutral once substitution effects are accounted for is an open question.

Substitution and the Croatian Market

Croatia's licensed market is not fully isolated. Players have access to offshore operators, and enforcement against unlicensed sites has been uneven. A player who finds a €10 max bet prompt irritating has options that don't involve reducing their stake.

The 2023 blocking regime, which added several hundred domains to the national blacklist, reduced but did not eliminate offshore access. Any operator considering a stake-segmented prompt should assume some substitution will occur, particularly among exactly the cohort most likely to respond to the prompt — experienced, higher-stake players who know where the alternatives are.

This is the uncomfortable core of the finding. The intervention works best on players who have the most options to avoid it.

Designing a Prompt That Actually Interrupts

If the goal is to reproduce the 31% effect without the warning-fatigue problem, the design details matter more than the threshold itself.

Friction over frequency. A prompt that requires an active confirmation — a click, a typed acknowledgment, a short delay — performs differently from a banner that auto-dismisses. The data doesn't isolate this variable, but the step function at €5 suggests that prompts which register as interruptions behave differently from prompts that don't.

Stake-relative triggers. Instead of a fixed euro threshold, trigger the prompt when a player's current stake exceeds their own session average by a set multiple. A player who has been spinning at €8 and jumps to €20 is showing a different signal than a player who has been at €20 all along.

Deposit-velocity coupling. The €1–€4.99 band's low response rate suggests that for those players, the rebuy is a deposit behaviour, not a stake behaviour. A prompt that fires when deposit frequency exceeds a threshold — say, three deposits in 20 minutes — would address the actual mechanism.

Cooling-off integration. A max bet warning that offers a 60-second pause, with the option to set a session limit directly from the prompt, converts a passive notification into an active tool. Operators who have tested this report higher opt-in rates for deposit limits, though the effect on rebuy frequency specifically hasn't been published.

None of these are novel ideas. What's novel is the evidence that the current default — a blanket prompt at a low threshold — is producing almost no measurable effect for the majority of sessions while the cohort that would respond is being missed or under-served.

The Measurement Problem

Operators measure what they can attribute. Rebuy frequency is easy to attribute to a prompt. Long-term player sustainability is not. This creates a structural bias toward interventions that produce short-term, measurable behaviour changes — even when those changes are small — over interventions that might matter more but take months to show up in the data.

The 31% figure is useful precisely because it's attributable and specific. But it's also a reminder that the industry's measurement infrastructure is built around the wrong time horizon. A player who reduces rebuys by 31% in the first month but churns entirely by the third is not obviously better off than a player who rebuys at the same rate but stays within their deposit limits.

What Happens Next

The practical question for Croatian operators is whether to segment max bet prompts by stake, and if so, where to draw the line. The data points to €5 as the point where prompts start working. But adopting €5 as a hard threshold would mean accepting that the 89% of sessions below it get no meaningful intervention — which, given that those sessions account for 66% of GGR, is a position that will be difficult to defend to the regulator and difficult to justify internally.

The more defensible path is to keep a low-threshold prompt for all sessions but redesign it so it doesn't habituate, and add a second, higher-friction intervention at the €5 line. Two-tier systems are more expensive to build and harder to explain to players. They also match what the data shows.

The open question is whether the 31% effect holds when the prompt is one of two, rather than the only one. Warning fatigue is context-dependent. A player who sees a low-friction banner 30 times and a high-friction confirmation twice may respond differently to the confirmation than a player who sees it in isolation. Nobody has published that comparison. Until someone does, the €5 threshold is a useful signal and an incomplete answer.