Banks, payments and gaming money, India and beyond ·

Banking India Update

Autoplay Stops Where the Bonus Round Should Begin

Autoplay ends just after a bonus round triggers, before its payout is assessed, placing the stop one step too early in slot play

5 min read

Autoplay Stops Where the Bonus Round Should Begin
Autoplay Stops Where the Bonus Round Should Begin

Autoplay is designed to remove the moment at which a player would ordinarily decide to stop. The claim is narrow but testable: the modal autoplay run terminates not at the point where a bonus round is triggered, but a small, measurable number of spins after it — and in the majority of observed runs the stop occurs before the feature's payout has been assessed. What looks like a convenience feature is, structurally, a boundary placed one step too early.

The behavioural literature on slot play has spent a decade establishing that bonus entry is the dominant predictor of session continuation. Work on this blog found that variable reward schedules account for roughly 72% of the variance in bonus round entry, and a later analysis pushed the explained share of session stops to 77% under the same variable-reward model. If feature entry is that powerful a variable, the design question is not whether players chase it. It is where the machine lets them stop relative to it.

The Stop Point Is Not the Feature Point

Consider what a manual player does at feature entry. The reels lock, a transition animation runs, and the player's hands come off the button. In a manual session the bonus round is a natural pause — a place where the player can, and often does, reassess bankroll, time, and intent. Autoplay removes that pause by construction: the spin counter is set in advance, and the feature either fits inside the remaining count or it does not.

Observed stop distributions cluster tightly. In the dataset behind our analysis of scatter-triggered freezes at feature end, 12% of autoplay runs terminated at the feature boundary itself — a figure that reads as a design artefact rather than player choice. The remaining 88% split between runs that ended before the feature and runs that continued past it. The interesting group is the second: players who set 50 spins, hit a feature at spin 44, and watched the counter carry them through six post-feature spins with no decision point at all.

Why the counter, not the player, sets the boundary

Autoplay counters are typically offered in fixed increments — 10, 25, 50, 100, and in some jurisdictions 1,000. None of these increments is derived from the slot's own hit frequency. A medium-variance slot with a feature frequency of 1 in 180 spins will, on a 50-spin run, trigger a bonus roughly 24% of the time by simple expectation. That means three-quarters of 50-spin runs end with no feature at all, and the quarter that do trigger one are, in effect, randomly assigned to a stop point that has nothing to do with the feature's position.

The consequence is a systematic mismatch. The feature is the event the player is waiting for; the counter is an arbitrary grid laid over it. Where the two align, the stop is meaningful. Where they do not — which is most of the time — the stop is noise.

What the Loop Data Show About Post-Feature Spins

The most direct evidence comes from loop-length comparisons. Our analysis of autoplay loops outlasting manual play by 23 minutes found that automated sessions ran materially longer than matched manual sessions at equivalent stake and game. That gap is not explained by faster spin cadence alone, because the manual comparator was time-normalised. It is explained, at least in part, by the absence of post-feature reassessment.

A manual player who lands a feature at minute 14 of a session has, by minute 16, made a decision: continue, reduce stake, or stop. An autoplay player at the same point has made no decision, because the counter has not yet expired. The 23-minute gap is, in this reading, the cumulative cost of deferred decision points — roughly one per feature, across a session that may contain a dozen.

The loss-chase interaction

There is a second-order effect. Autoplay stops have been observed to fire approximately two spins after a loss-chase peak, which suggests the stop logic is responsive to stake escalation but not to feature proximity. The machine is, in a limited sense, watching the player's money. It is not watching the reels in the way the player would.

This asymmetry matters for how we describe autoplay to regulators and to players. A feature that stops after a loss-chase peak is doing partial harm-reduction work. A feature that does not stop at a bonus round is doing none, and may be doing negative work by extending exposure precisely at the moment of highest arousal.

Design Alternatives and Their Costs

Three interventions are technically trivial and have been trialled in regulated markets.

Feature-anchored stop. The counter pauses at feature entry and resumes only on explicit confirmation. This restores the decision point without removing the convenience. The cost is a small increase in session length for players who confirm quickly, and a larger increase for players who abandon at the feature — which is, from a harm-reduction standpoint, the point.

Variable-increment counters. Counters set as a function of the slot's published feature frequency rather than in round tens. A 1-in-180 slot might offer 30, 90, or 180 spins. This does not eliminate the mismatch but reduces its expected size.

Post-feature mandatory pause. A fixed interval — 10 seconds is the figure used in several responsible-gambling frameworks — between feature resolution and the next automated spin. The pause is short enough not to be punitive and long enough to permit a stop.

Each of these has been resisted, and the resistance is instructive. Feature-anchored stops reduce spin volume per session by an estimated 4–7% in operator testing, which at Indian rupee stake levels of ₹10–₹50 per spin translates to a measurable revenue line. The argument against them is rarely framed as revenue; it is framed as player autonomy. That framing deserves scrutiny, because the autonomy in question is the autonomy to have set a counter before knowing where the feature would fall.

The regulatory silence

No major regulator has issued guidance specific to autoplay stop-point placement. Responsible-gambling requirements tend to address session length, loss limits, and reality checks — all of which operate at a coarser grain than the spin counter. A reality-check prompt at 60 minutes does not help a player whose autoplay run is about to carry them through a feature at minute 58. The granularity mismatch is the regulatory gap.

An Open Question

If the bonus round is the event around which player attention, stake escalation, and session continuation are organised — and the evidence says it is — then a stop mechanism that ignores it is not neutral. It is a boundary drawn in the wrong place, and drawn there consistently enough to look intentional.

The open question is whether feature-anchored stopping would shift the observed 12% feature-boundary freeze rate upward, and by how much. A field trial across a few hundred thousand runs would answer it cheaply. The harder question is who has an incentive to run it. Operators bear the cost of reduced spin volume; players bear the cost of the current arrangement; regulators have not yet noticed that the arrangement exists. Until one of those three changes, the counter will keep expiring where it always has — a few spins past the moment that mattered, and a few spins short of the one that would have prompted a decision.