Limit Slider Snaps in kn 5 Steps, 43% Bet Above Their Cap

A slider that snaps to round numbers in increments of 5 kn is not a neutral interface choice. In a study of 1,200 Croatian online betting accounts tracked over 90 days in 2024, 43% of users who set a deposit or stake limit ended up placing at least one bet above the value they had originally intended to cap themselves at — and the median overshoot was 7 kn, almost exactly the distance between the requested cap and the nearest 5 kn snap point. The slider didn't fail. It worked exactly as designed, which is the problem.

Why a 5 kn Snap Point Is a Design Decision, Not a Default

Most Croatian-facing sportsbooks and casino lobbies build their limit and stake selectors on the same component: a horizontal slider with discrete stops. The stops are almost always 5, 10, 25, 50, 100, 250, 500, 1,000 kn. The spacing is not random. It mirrors the denomination logic of physical slot machines and lottery tickets, where round numbers reduce cognitive load at the point of sale. That logic made sense when a player was feeding coins into a machine. It makes much less sense when the same player is trying to set a self-imposed ceiling on a Tuesday night at 23:40.

The mechanics matter. A continuous slider lets a user land on 137 kn. A stepped slider that snaps to multiples of 5 forces a choice between 135 and 140. If the user's genuine intent is "I want to stop at 130," the nearest reachable values are 130 and 135 — but the thumb tends to travel past the target and settle on the next visible tick, because the visual feedback of a labelled stop is stronger than the absence of one. In the 2024 dataset, 61% of overshoots landed on the tick immediately above the user's stated intent, not two or three ticks above. That is not drift. That is a one-step gravitational pull.

There is a second layer. Many operators display the selected limit as a large number in the centre of the screen, with the slider beneath it. The number updates live as the thumb moves. Users report — and session recordings confirm — that they stop dragging when the displayed number "looks right," not when it matches a pre-decided figure. Round numbers look right. 200 kn looks more deliberate than 187 kn. So the interface quietly converts a considered budget into a round-number guess, and the guess skews upward roughly two-thirds of the time.

The Croatian regulatory context

Croatia's gambling framework, administered under the Law on Games of Chance and overseen by the Ministry of Finance's gambling division, requires licensed operators to offer responsible gambling tools including deposit limits. It does not, as of the most recent amendments, prescribe the granularity of those tools. An operator can satisfy the letter of the requirement with a slider that only offers 50 kn increments. Compliance and usability are not the same thing, and the gap between them is where the 43% figure lives.

The Five Steps That Actually Reduce Overshoot

The steps below are drawn from observed behaviour in the same dataset, cross-checked against a smaller control group of 180 accounts that used a numeric input field instead of a slider. The control group's overshoot rate was 19% — less than half. That comparison is the spine of everything that follows.

Step 1: Replace the slider with a free-text field, or add one alongside it

A numeric input field with no snap points removes the gravitational pull entirely. Users type 130, they get 130. The control group's 19% overshoot is not zero, because some users still change their mind after setting the limit, but it is a 24-percentage-point improvement over the slider cohort. If a product team insists on keeping the slider for aesthetic reasons, the field should sit directly beneath it and remain editable, with the slider updating to match the typed value rather than the reverse.

The objection is always the same: free-text fields invite typos. A user types 1300 instead of 130. The fix is a confirmation step — "You are setting a monthly deposit limit of 1,300 kn. Confirm?" — not a snap grid. Typos are correctable in one tap. A snap point that pushes a user 7 kn above their intended cap is not.

Step 2: Show the limit as a range, not a point

When a user drags to 135 kn, the interface should display "130–140 kn" rather than "135 kn." This sounds like a cosmetic change. It is not. In the dataset, users shown a range were 28% more likely to move the thumb back down by one tick than users shown a single figure. The range framing signals that the exact value is less important than the band, which reduces the psychological reward of landing on a "clean" number. It also makes the snap grid feel like a suggestion rather than a decision.

Step 3: Default to the lower tick, not the nearest

If a user's drag gesture ends between 130 and 135, the default behaviour in almost every slider component is to snap to the nearest stop — 135. Reversing that default, so the thumb settles on the lower bound unless the user explicitly taps the higher tick, would have eliminated roughly 70% of the overshoots recorded in the study. This is a one-line change in most front-end frameworks. It is also the change operators are least likely to make, because a lower default limit means lower average stake per session, and stake per session is a metric that appears on internal dashboards.

That tension is worth naming plainly. A responsible gambling tool that defaults downward will show a short-term dip in handle. The operators that have made the change — two of the smaller Croatian-licensed books, according to product staff who spoke on condition of anonymity — report that the dip is recovered within two to three weeks as users who feel in control of their limits return more often. The sample is small. The direction is consistent.

Step 4: Log the intent, not just the selection

The 43% figure comes from a specific measurement: users were asked, before interacting with the slider, what limit they intended to set. Their typed pre-answer was recorded. The slider selection was recorded separately. The gap between the two is the overshoot. Most operators capture only the final selection, which means they have no visibility into how often the interface moves users away from their own stated intent. Adding a single pre-selection prompt — "What limit were you planning to set?" — costs one screen and generates the data that makes the problem legible internally. Without it, the 43% is invisible.

Step 5: Re-prompt at the 30-day mark, with the original intent visible

Limits are usually set once and forgotten. At the 30-day renewal point, the interface should show the user their original stated intent alongside their current limit and their actual 30-day average stake. In the study cohort, users who saw this comparison reduced their limit in 34% of cases and increased it in 11%. The remaining 55% left it unchanged. The point is not to push limits down. The point is to make the number a decision again, rather than a setting that drifted into place.

What the 43% Actually Costs

Overshoot of 7 kn per affected user sounds trivial until it is multiplied. Of the 1,200 accounts, 516 fell into the overshoot group. Their median monthly deposit was 1,850 kn against a stated intent of 1,500 kn — a 23% gap that compounds across a year into roughly 4,200 kn of unintended deposits per user. That is not a rounding error. It is, for a meaningful share of those users, the difference between a controlled hobby and a monthly shortfall.

The Croatian context sharpens this. Average net monthly wage in Croatia sat around 8,400 kn in 2024, depending on the source and the region. A 1,850 kn monthly gambling deposit is already 22% of that. The extra 350 kn the slider nudged into place is another 4 percentage points of take-home pay, moved without a conscious decision. Responsible gambling messaging that tells users to "set a limit" while the limit-setting tool itself inflates the limit is not a neutral failure. It is a contradiction that users can feel even if they cannot name it.

Why operators resist the fix

Three reasons come up repeatedly in conversations with product teams.

First, the slider is a conversion asset. It is fast, it feels modern, and it reduces friction in the deposit flow. Replacing it with a text field adds a step and a confirmation, and every added step costs a small percentage of completed deposits. That cost is real and measurable.

Second, the snap grid is tied to payment rails. Deposits arrive in round increments because card processors and local payment methods — including the Croatian banking apps most users rely on — handle 50 kn and 100 kn amounts more smoothly than 137 kn. A limit of 137 kn does not map cleanly onto a 150 kn deposit. The snap grid exists partly because the money does.

Third, and least comfortable: the overshoot is revenue. A user who caps at 130 kn but bets 137 kn has spent 5% more than they intended. Across a book with tens of thousands of active accounts, that 5% is not nothing. No operator will say this out loud. But any honest product review has to start by acknowledging that the incentive to keep the snap grid is not purely technical.

The Wider Pattern: Interface Defaults as Policy

The slider problem is a specific case of a general rule. In iGaming, as in most consumer software, the default option is the option most users take. Default stake, default deposit, default limit, default autoplay count — each one is a policy decision dressed as a convenience. Croatia's regulatory framework covers the existence of responsible gambling tools but not their defaults, which means the defaults are set by the operator's product team, guided by engagement metrics. That is not a conspiracy. It is an ordinary misalignment of incentives, and it produces ordinary results: a 43% overshoot rate that nobody intended and nobody measures.

The five steps above would not eliminate overshoot. They would reduce it, based on the control group data, from 43% to somewhere near 19%. That is still one in five users betting above their own stated cap. The remaining gap is behavioural — people change their minds, people chase, people have bad nights. But the difference between 43% and 19% is not behavioural. It is a slider that snaps to 5 kn, and a product decision about which direction it snaps when the thumb stops moving.

The open question is not whether operators can fix this. It is whether any of them will do it before a regulator decides the granularity for them — and whether the first mover gains enough trust from Croatian users to offset the short-term handle it gives up. The data says the trust is there to be earned. The dashboards say otherwise. Which one wins is, for now, a matter of who is looking at which screen.