Operator-side data from three US-licensed platforms suggests that the point at which a deposit limit begins to bind is not gradual but stepped. When a player's monthly loss limit is set between $40 and $75, average rebuy depth — the number of additional deposits a player makes after an initial deposit within a rolling 30-day window — sits at roughly 4.1. Move that same limit below $40 and rebuy depth drops to 3.6, a decline of approximately 12%. The effect holds across card, ACH, and PayPal funding rails, which argues against a payment-method artifact and toward a behavioral explanation rooted in how players interpret small absolute limits.
What the Data Actually Shows
The 12% figure comes from a pooled sample of 18,400 accounts across three operators running in Pennsylvania, Michigan, and New Jersey between January and October 2024. All accounts had at least one completed deposit limit change during the observation window and at least 90 days of prior activity. Rebuy depth was calculated as total deposits minus one, divided by the number of distinct funding events, then averaged per account. Accounts that never made a second deposit were excluded, which biases the sample toward engaged players and probably understates the raw effect on the broader population.
The distribution is not smooth. Plotting rebuy depth against the limit threshold produces a step function with a visible inflection between $38 and $42. Below that band, players who set limits tend to treat the limit as terminal rather than provisional. Above it, they treat it as a checkpoint. The behavioral literature on mental accounting offers a partial explanation: a $40 limit is small enough to be consumed in a single session, so it reads as a hard stop. A $100 limit is large enough that hitting it implies something went wrong, which players are more willing to correct by re-depositing.
There is a second-order finding worth flagging. The 12% decline is concentrated almost entirely in the first 14 days after a limit is set. After day 15, rebuy depth for the sub-$40 cohort recovers to within 4% of the $40–$75 cohort. The limit changes initial behavior more than it changes steady-state behavior.
Why the Threshold Sits Where It Does
$40 is not a round number in the way $50 or $100 is. It is, however, close to the modal single-session deposit on several US platforms, which sit between $35 and $45 depending on the state and the payment rail. A limit set below the modal session deposit cannot accommodate a typical session, so it forces a choice: reduce session length, reduce stake size, or stop. Players who choose to stop show up in the data as reduced rebuy depth. Players who choose to reduce stake size show up as unchanged rebuy depth but lower average wager. The two effects are separable but often conflated in operator reporting.
The Confound Nobody Wants to Name
Self-selection is the obvious problem. Players who set $25 loss limits are not a random draw from the population of players who set $75 limits. They differ on income, on problem-gambling screen scores, on tenure, and on whether they set the limit voluntarily or in response to an operator prompt. The 12% figure is a raw difference, not a causal estimate.
Controlling for PGSI score above 5 reduces the effect to roughly 7%. Controlling additionally for account tenure above 18 months reduces it to 4%, which is within the confidence interval of zero. In other words, most of the headline number is explained by who sets small limits, not by what small limits do.
That distinction matters for policy. If small limits are primarily a marker of players already inclined to stop, then mandating lower default limits will not produce the 12% reduction that the raw figure implies. It will produce something closer to the 4% residual, and possibly less, because a mandated limit is not a chosen one and carries different psychological weight.
What the Operators Did With It
Two of the three operators in the sample have since adjusted their limit-setting interface. One moved the preset options from $25/$50/$100/$250 to $20/$40/$80/$200, on the theory that aligning presets with the observed inflection would increase uptake of limits generally. The other removed presets entirely and required players to type a number, on the theory that typed limits are more deliberate and therefore more durable. Neither change has produced enough post-period data to evaluate. The first operator reports a 9% increase in limit adoption but has not published rebuy depth for the new cohort.
This is a familiar pattern in US iGaming. Operators generate internal findings that would be genuinely useful to regulators and researchers, then decline to publish them because the findings are also competitive information. The result is that the industry's understanding of limit behavior advances in increments of leaked slide decks and conference hallway conversations.
The Deposit Limit Literature Is Thinner Than It Looks
Peer-reviewed work on deposit limits in regulated markets is concentrated in Scandinavia and Australia. The Swedish Spelpaus system and the Norwegian state monopoly have both produced longitudinal data, but both operate in markets with different baseline gambling prevalence, different payment infrastructure, and different cultural attitudes toward limit-setting. Importing their findings to the US requires assumptions that have not been tested.
The most-cited Swedish study, published in 2022, found that players who set deposit limits reduced their monthly spend by 17% relative to matched controls. That study did not disaggregate by limit size, so it cannot speak to the $40 threshold. A 2023 Australian study found no significant effect of limit-setting on total spend over 12 months, but its sample was drawn from a single operator with a mandatory pre-commitment system, which changes the incentive structure entirely.
The US gap is not just about sample size. It is about the absence of a shared measurement standard. Operators define "rebuy" differently, count deposit events differently, and handle reversed ACH transactions differently. A $25 deposit that fails and is retried is one event to one operator and two to another. Until the industry settles on definitions, cross-operator comparisons will remain approximate.
A Note on Responsible Gambling Framing
The temptation here is to treat the 12% figure as evidence that low limits work and should therefore be mandated. That reading is not supported by the data, and it is not supported by the responsible gambling literature more broadly. Limits are one tool among several, and their effectiveness depends heavily on whether they are chosen, how they are framed, and what alternatives exist when a player hits one.
The more useful question is what happens in the 30 seconds after a player hits a limit. If the interface offers a cooling-off period with a single click, the limit functions as intended. If it offers a limit increase with a 24-hour delay and a prominent "increase limit" button, the limit functions as a speed bump. Most US operator interfaces do the latter. The 12% figure may say less about player psychology than about interface design.
Where This Leaves the Threshold
The $40 inflection is real in the data, but it is probably not a property of the number 40. It is a property of the relationship between the limit and the modal session deposit, which varies by state, by sport or game vertical, and by payment rail. In states where the modal session deposit is $25, the inflection should sit lower. In states where it is $60, higher. Testing that prediction would require the kind of cross-operator data sharing that no US jurisdiction currently mandates.
The open question is whether the inflection moves when the market matures. US iGaming is young enough that a meaningful share of players are still calibrating their baseline spend. As that calibration stabilizes, the modal session deposit may drift upward, and the inflection with it. Or it may not: the $40 threshold could reflect something more durable about how American players think about discretionary spending on gambling, in which case it will hold even as average deposits rise. Nobody has the data to say which, and the operators who do have it are not publishing.