Deposit presets cut £5 stake slips by 28% in the 22:00–23:00 window, according to a six-month comparison of two operator cohorts handling a combined 41 million monthly sessions. The effect was largest among players aged 25–34 and among those who had previously set a deposit limit. It was smallest — close to zero — among high-stake sports bettors who routinely type four-figure amounts.

The mechanism is unglamorous. When a cashier offers £10/£20/£50/£100 buttons, the number a player intends to stake is anchored before they reach the keyboard. At 10pm, after a drink and with a phone in one hand, that anchor does more work than any responsible-gambling banner. The 28% figure is not a measure of harm prevented. It is a measure of how much of the £5 slip volume was never really a decision in the first place.

What the 28% actually measures

The comparison ran from September 2024 to February 2025 across a UK-licensed operator's sportsbook and casino cashier. Cohort A (roughly 1.9 million monthly actives) used a free-text stake field with a quick-pick row of previous stakes. Cohort B (2.2 million) used fixed preset buttons with a "custom amount" link that opened a secondary screen. Both cohorts saw identical deposit limits, reality checks and time-on-site prompts.

"Stake slip" here means a wager that was placed after at least one rejected or corrected entry — a mistyped amount, a decimal in the wrong place, a stake placed and immediately voided, or a stake materially different from the player's modal stake for that product. In the 22:00–23:00 hour, Cohort B logged 28.4% fewer of these than Cohort A, adjusted for product mix and day of week. Across the full day the gap narrowed to 11.2%.

That divergence matters. The preset effect is not a general behavioural nudge that operates uniformly. It is concentrated in the hours when attention is thinnest and when, at least in this dataset, the £5 stake is most likely to be an artefact rather than an intention.

Three findings stood out:

• The £5 cluster. In Cohort A, £5 stakes spiked 34% between 21:30 and 23:30 relative to the 14:00–16:00 baseline, with a disproportionate share arriving via corrected entries. In Cohort B the spike was 9%. The preset row started at £10, so £5 required the custom screen — one extra tap that filtered out the casual entry.
• Age gradient. The 28% headline was 39% among 25–34s and 14% among over-55s. Younger players corrected entries more often in free text, which is consistent with faster typing and less checking rather than with greater impulsivity.
• Limit-setters behaved differently. Players with a pre-existing deposit limit showed a 31% reduction, slightly above the cohort average. That runs against the assumption that limit-setters are already "protected" and therefore immune to interface effects.

The custom-amount escape hatch

Removing the typed field entirely would have been the cleaner experiment, and no operator will run it. The "custom amount" link was used by 6.8% of sessions in Cohort B but accounted for 22% of total staked value. Any cashier redesign that ignores this will push high-value players to competitors within a fortnight. The presets work precisely because they are a default, not a constraint.

Why 10pm is the interesting hour

The 22:00–23:00 window carries a well-documented pattern in UK remote gambling data: session starts rise, average session length rises, and in-play football markets on midweek European fixtures pull in casual bettors who did not plan to bet that evening. It is also the hour when the GamCare helpline historically sees its highest call volume — a correlation worth noting without over-reading, since call volume reflects when people are awake and alone as much as when they are struggling.

The practical point is that interface defaults do their heaviest lifting when cognitive load is high. A player at 2pm comparing Championship odds has the attention to type 7.50 correctly. The same player at 10.40pm, second screen open, is a different decision-maker. Presets are cheap insurance against that version of the player — and unlike a deposit limit, they cost nothing in friction for the player who genuinely wants to stake £25.

There is a less comfortable reading. If 28% of £5 slips at 10pm were corrections rather than intentions, some proportion of those corrected entries would previously have become £50 stakes with a misplaced decimal. Presets reduce both the accidental small stake and the accidental large one. Operators tracking only average stake will see a fall and may misread it as cannibalisation.

The operational case, and where it breaks

Compliance teams have an obvious interest here. Mistyped stakes generate disputes, and disputes generate complaints to ADR providers and, at the margin, to the Commission. A cashier that reduces corrected entries by 28% in the peak evening hour reduces a category of customer-service contact that is expensive and rarely resolved to anyone's satisfaction.

But presets are not a harm-reduction tool in any formal sense, and treating them as one would be a mistake. They do not restrict spend, they do not surface time or loss information, and they do not intervene. A player who wants to stake £500 at 10pm will find the custom screen in under three seconds. The 28% figure describes an interface that removes noise, not one that changes intent.

Where the approach breaks down:

1. High-variance products. On slots, the "stake" decision is per-spin and presets of £0.10/£0.50/£1/£5 map poorly to how players actually think about session budget. The cohort data showed almost no effect on slot stake corrections, which is unsurprising.
2. Odds-implied staking. Betting £10 at 1.20 and £10 at 8.00 are entirely different propositions. Presets that fix the stake rather than the liability can nudge players toward short prices without them noticing.
3. The 6.8% who opt out. Any operator rolling this out should instrument the custom-amount path from day one. If that share climbs past roughly 12%, the preset row has become decoration.

The second point deserves more attention than it gets. Fixed-stake presets are a reasonable default for a football accumulator and a poor one for a single short-priced favourite, where the liability is the number that matters. An operator serious about this would offer liability presets on the sportsbook and stake presets on the casino, and would measure corrections separately for each.

The question the data can't answer

What the cohort comparison cannot tell you is whether fewer corrected entries means fewer regretted bets. A player who stakes £10 instead of a mistyped £5 has not necessarily had a better evening. The metric is about interface friction, and friction is not the same as harm — some friction is load-bearing, and the deposit limit sitting behind a deliberate menu is a good example.

The open question for UK operators is whether preset design should be tuned to the hour. If the 10pm effect is genuinely about cognitive load, a cashier that surfaces presets and hides the free-text field between 21:00 and 00:00 — and reverses that at 2pm — would capture most of the benefit without the 6.8% opt-out drift. Nobody has published that test yet. Given how cheap it is to run, and how much of the £5 slip volume appears to be noise rather than intent, it is a strange thing to leave unmeasured.

Players who find themselves correcting stakes more often than they'd like, at any hour, can set a deposit limit directly in the cashier or via GamStop, and talk to GamCare on 0808 8020 133 if the pattern feels less like a typing problem and more like something else.