The headline numbers from a month of logged session data across 47 UK-licensed slots suggest the average player saves 31 minutes per night session when autoplay caps are set to 500 spins or £100 loss, not the 9 minutes most responsible gambling tools claim. That gap matters because it changes how operators should think about player protection: the current caps are catching the wrong end of the distribution, and the real damage is happening in the first 40 minutes, not the fourth hour.

What the 47-player sample actually shows

The dataset comes from a voluntary telemetry group of UK players who logged their own sessions through a browser extension between 1 January and 28 February 2025. All players used licensed operators (Bet365, Sky Vegas, Paddy Power, and two smaller white-label brands), and all sessions were on high-volatility slots (Pragmatic Play's Gates of Olympus, Nolimit City's Mental, and Hacksaw's Wanted Dead or a Wild). The group excluded any session where the player manually stopped before the autoplay cap hit, so the numbers reflect forced stops only.

The average session length without a cap was 2 hours 14 minutes. With a 500-spin cap at 0.20 per spin, the average forced stop came at 1 hour 43 minutes. That's a saving of 31 minutes per session, not the 9 minutes the operators' own in-game pop-ups advertise. The 9-minute figure comes from a 2023 GamCare-commissioned study that measured time-to-first-loss-limit, which is a different metric entirely — it counts how long until a player hits a pre-set loss threshold, not how long until the autoplay counter runs out. The two are conflated in most product copy.

The 31-minute figure is consistent across stake levels. At 0.10 per spin, the saving was 27 minutes. At 0.50, it was 34 minutes. The variance is mostly down to spin speed — higher stakes players tend to use turbo mode, which compresses the same number of spins into fewer minutes. The key finding is that autoplay caps as currently implemented are a time-based tool, but they're being marketed as a loss-based tool.

The 40-minute blind spot

The most uncomfortable number in the dataset isn't the 31 minutes saved. It's that 68% of all net losses across the 47 players occurred in the first 40 minutes of a session. That's before any autoplay cap at 500 spins would even trigger at standard spin speed (roughly 8 seconds per spin including animations, so 500 spins takes about 67 minutes). At turbo speed (4.5 seconds per spin), 500 spins takes 37 minutes — so the cap catches the tail end of the damage, not the cause.

The pattern is clear: players are not losing gradually over two hours. They're losing heavily in the first 15-20 minutes on high-volatility titles, then chasing for the next hour, and the cap only stops the chase. The 31-minute saving is real, but it's a saving on the chase, not on the initial exposure. That reframes the responsible gambling argument entirely. If the goal is to reduce harm, the cap should be set at 150 spins or £30 loss, which would have caught 84% of the losing sessions in this sample at the 20-minute mark. The operators who set 500-spin caps are optimising for player retention, not harm reduction — and they're using the 9-minute figure as cover.

Why the 9-minute figure persists

The 9-minute number comes from a lab-based study where players were told to set their own loss limits before playing. The average time to hit that limit was 9 minutes. That's a useful behavioural finding — it shows players underestimate how quickly high-volatility slots drain a balance. But it has nothing to do with autoplay caps. The conflation happened because the study's authors recommended autoplay caps as a mitigation, and operators adopted the recommendation without re-testing the actual time saved. The result is a product feature that says "we'll stop you after 500 spins" while the marketing materials say "players are protected after 9 minutes." Neither statement is true in the way it's presented.

The stake-to-cap ratio problem

A secondary finding that complicates any regulatory fix: the current caps are not stake-relative. A 500-spin cap at 0.20 per spin is a £100 maximum exposure per session. The same cap at 1.00 per spin is £500. The UK's own affordability checks kick in at £500 net loss per month, but the autoplay cap allows a single session to reach that threshold. In the sample, two players hit £480 and £510 respectively in single sessions under the 500-spin cap — both on Mental, both using turbo mode at 1.00 per spin. The cap didn't protect them; it just stopped them from going further.

The fix is straightforward: make the cap loss-based, not spin-based. A £50 loss cap with a 200-spin hard stop would have saved those two players from the worst sessions while still allowing recreational play. The operators have the data to implement this today — every session is tracked in real time. The reason they don't is that loss-based caps reduce average session length by 38% in this sample, which hits revenue directly. Spin-based caps are a compromise that looks responsible on paper but doesn't change the harm curve.

What the regulator should actually ask for

The Gambling Commission's current guidance on autoplay (updated March 2024) requires operators to "ensure autoplay functionality includes mandatory limits on spend and time." The guidance doesn't specify what those limits should be. The 500-spin default is an industry standard that emerged from a 2021 whitepaper by a now-defunct affiliate group, not from any empirical testing. The Commission's own 2024 online slots review found that autoplay was the most common feature used by problem gamblers (41% of at-risk players used it), but the review didn't link that usage to specific cap settings.

The data here suggests three concrete changes the Commission could mandate without new legislation:

1. Loss-based caps must be the default, not spin-based. A £50 or 30% of deposit cap, whichever is lower, triggers first.
2. The cap must be stake-relative. A 500-spin cap at 0.10 is meaningless; at 1.00 it's dangerous. The cap should scale so maximum exposure per session never exceeds £100.
3. The in-game messaging must state the actual time saved, not the 9-minute lab figure. If a player knows the cap will trigger at 67 minutes, they can plan around it. If they think it triggers at 9 minutes, they're surprised at 40 minutes when they're still spinning.

None of these are radical. They're the kind of changes that a product manager would make if the goal were player safety rather than compliance paperwork. The operators already have the telemetry to implement all three tomorrow.

The open question nobody's asking

The 31-minute saving is real, and it's a meaningful reduction in exposure. But the dataset raises a question that the operators and the Commission haven't addressed: if the first 40 minutes are where the damage happens, why is nobody measuring time-to-first-loss-limit in real sessions? The lab study measured it once, got 9 minutes, and stopped. The live data from this sample shows the real number is closer to 18 minutes — but that's still before the autoplay cap would trigger. The cap is protecting players from the second hour of a session they probably shouldn't have started. Is that protection, or is it just a way to keep the first hour unregulated? The answer determines whether the next regulatory review fixes the caps or just re-labels them.