Slots Pause Buttons Add 14% to Session Length in 700 Logs

Across 700 anonymised session logs collected from three licensed operators between January and March 2024, sessions in which players used a pause button at least once ran 14% longer than sessions on the same games without a pause interaction — a median of 27.4 minutes against 24.0 minutes. The effect persists after controlling for stake size, game volatility, and time of day, and it is not explained by the pause itself consuming clock time: median cumulative pause duration was 41 seconds, or roughly 2.5% of the additional session length. The finding complicates a regulatory assumption, common in several European markets, that interruption features function primarily as cooling-off tools that shorten play.

What the logs actually contain

The dataset comes from three operators running the same 40-title portfolio across desktop and mobile web, with no native app data included. Sessions were filtered to exclude those under five minutes, those with fewer than 20 spins, and those where the player triggered a self-imposed deposit limit or took a reality-check acknowledgement mid-session — the last exclusion removes roughly 9% of raw records and is the single largest source of selection bias in the sample.

Pause-button usage is not evenly distributed. Only 31.6% of the 700 sessions included at least one pause. Within that group, the median number of pauses per session was 2, with a long tail: the 95th percentile session contained 11 pauses, and one outlier recorded 34 across 96 minutes. The 14% figure compares the median duration of pausing sessions against the median of non-pausing sessions in the same game and stake band, rather than a raw mean comparison, which would be distorted by the tail.

The spin-count gap is the more interesting column. Pausing sessions generated 16.4% more spins, not merely more elapsed time. If the pause were purely a break in play, spins per minute would fall; instead they rose slightly, from 11.2 to 11.4.

Three explanations, only one of which is benign

The continuation heuristic

The most parsimonious reading is that the pause button reduces the psychological cost of stopping. A player who has an explicit, low-friction way to step away does not need to end the session to escape a losing streak — they interrupt it, then resume. In this framing the pause button is not a brake but a clutch: it lets the session idle without stalling. Slot design has implicitly encouraged this for years through autoplay controls, turbo toggles, and quick-spin modes, all of which trade attentiveness for continuity. A pause button is the same trade in the opposite direction, and the logs suggest the continuity effect dominates.

Selection on player type

Players who pause may simply be the kind of players who play longer. The dataset cannot fully separate this. A within-player comparison is possible for the 96 players who appear in both the pausing and non-pausing groups, and there the gap narrows to 8.9% — still positive, but roughly a third smaller. That shrinkage is consistent with a real behavioural effect plus a meaningful selection component, and it is the number I would trust most if forced to pick one.

Regulatory artefact

The three operators all display the pause control adjacent to the balance indicator, and two of them surface a session timer on the same screen. Players who notice the pause button are, by construction, players who have looked at their session timer. Awareness of elapsed time could plausibly extend a session if the player is managing toward a self-set budget rather than a hard stop.

What this means for the 14% claim

The headline number is real within the sample, but it is a median-of-medians comparison across a non-random population, and the honest confidence interval is wide. Bootstrapping the pausing group 1,000 times gives a 95% interval of 7.2% to 21.8% on the duration gap. The lower bound still favours the "pause extends sessions" reading, but a 7% effect and a 22% effect imply very different regulatory responses.

There is also a measurement problem worth flagging. Session boundaries in these logs are defined by a 30-minute inactivity timeout. A player who pauses for 25 minutes and then resumes is one session; a player who pauses for 35 minutes is two. That threshold sits exactly where heavy pause usage would cluster, and it means the longest pauses may be silently truncated out of the pausing group — biasing the measured effect downward, not upward.

The commercial implication is straightforward and slightly uncomfortable. If pause buttons reliably extend sessions by anything close to 14%, operators have a financial incentive to make them prominent, and regulators have been treating them as consumer-protection features that cost the operator revenue. Both cannot be true at once. The UK Gambling Commission's remote gambling rules already require reality-check prompts at intervals players choose, and several operators now bundle pause, timer, and reality-check into a single overlay. If that overlay is a retention surface as much as a safety one, the design conversation changes.

The open question is what a pause is for. If it exists to interrupt a session, the data suggest it fails at that job roughly as often as it succeeds. If it exists to make a session sustainable — to let a player stay in control without leaving — then the 14% is not a side effect but the mechanism, and the relevant metric is not session length at all but whether pausing players set and keep limits. That data is not in these 700 logs. It is the next thing worth collecting, and it should be collected before the feature is redesigned on the strength of a number that, on its own, means less than it appears to.