A 22% drop in average session length. That's the headline figure from three operators who switched off autoplay at midnight local time rather than letting it run indefinitely, and it's held steady across a 90-day window ending in March. The same operators reported a 9% decline in total spins per active player, but almost no change in deposit frequency. People didn't stop playing. They just stopped playing as long.

What actually changed at midnight

The intervention was narrow: autoplay could still be set, but any session running past 00:00 local time had its spin counter reset to zero. No forced logout, no deposit block. A player who wanted to keep spinning could simply re-enable it — one tap.

That's what makes the 22% number interesting. This wasn't a hard stop. It was friction, and friction alone moved behaviour.

Operators running the test split their traffic into a control group (unchanged) and a treatment group. The treatment group's median session length fell from 41 minutes to 32 minutes. The 95th percentile — the long tail of grinders — fell harder, from 3h 10m to 2h 18m. That tail is where a disproportionate share of net gaming revenue usually sits, which is why the revenue line moved too: down 6.4% in the treatment cohort over the quarter.

Why autoplay is the lever, not the game

Autoplay removes the moment of decision. Every spin you press manually is a small checkpoint — a chance to notice the balance, check the clock, get a drink. Autoplay deletes those checkpoints and replaces them with a metronome.

This is well understood in behavioural research. Default settings and reduced friction shape consumption more reliably than warnings or self-set limits. A 2019 study of Swedish online gamblers found that players using autoplay features had session lengths roughly 40% longer than those who didn't, controlling for stake size. The midnight reset works because it reintroduces a checkpoint at the exact hour when fatigue and loss-chasing tend to peak.

The stake-size wrinkle

One operator's data showed something less tidy: players who re-enabled autoplay after midnight raised their average stake by 7%. The plausible reading is that the interruption broke the flow state, and some players responded by escalating rather than stopping. That's a real caveat for anyone treating the 22% figure as a clean win.

What the regulators are watching

Several European regulators have already floated autoplay restrictions, and the UK Gambling Commission banned autoplay on slots in 2021 — one of the few markets where this isn't hypothetical. Malta and a handful of others still permit it with limits (typically 25–100 spins, mandatory loss displays).

The operators in this test are watching for a specific signal: whether the session-length drop persists past 90 days, or whether players simply shift their start times earlier to bank the same uninterrupted run. Early data on that is mixed. About a third of the affected players moved their first spin of the evening forward by 40 minutes or more, which suggests some are optimising around the rule rather than reducing play.

The question nobody's answered

If a one-tap re-enable cuts sessions by 22%, what does a two-tap re-enable do? Or a 30-second cooldown? The dose-response curve here is the whole argument — and it cuts both ways. Regulators pushing for hard autoplay bans should ask whether the marginal gain over a midnight reset justifies the loss of a feature that, for some players, is just convenience.

For operators, the more uncomfortable question is what the 6.4% revenue hit buys them. If it buys a lower regulatory risk profile and fewer 3am sessions from players who shouldn't be there, it may be cheap. If players route around it within a month, they've paid for nothing.