Attendant-Free Attractions: The Labor Hour You Still Budget
Valo Motion lists its Groundfall attraction as running “without staff, headsets, controllers, LED tiles, or floor sensors.” The phrase sits in a bullet list above the configuration table, between the note about 19 layouts and the three-minute session length.
SEGA Amusements International uses a shorter version for its new cabinet: two-player, attendant-free. House of the Dead Reawakened makes its global debut at IAAPA Expo in Orlando on 17 November 2026.
ROLLER took a different piece of the same job in May. Its Guest Experience Agent answers the telephone, takes the booking and upsells, around the clock, without a front desk behind it.
Read together, the three describe a change in what gets specified. The design question is no longer “how many staff does this attraction need at peak.” It is “can the unit run between visits by a human, and who is still watching when it does.”
That distinction matters at the schedule, not the brochure. A venue’s labor bill is written in hours, and the hours attached to an attraction are not the same as the hours attached to the machine.
This article takes five dated signals from five companies and asks the question a buyer can act on: when an attraction loses its attendant, how many labor hours actually leave the building?
1. Five signals, six countries, one direction
None of these rows describes the same event. Five companies, six countries, and three kinds of source: a company release, trade press, and a filing with the US Securities and Exchange Commission. That is the test a trend has to pass before it is a trend rather than one supplier’s launch.
| Date | Company | Country | What was actually said |
|---|---|---|---|
| 27 May 2026 | ROLLER | Australia / US | Guest-facing AI agent answers questions, takes bookings and upsells by phone, email or chat, 24 hours a day |
| 14–16 Sep 2026 | IATP (award) | US | Breakthrough Product award for that agent, presented at the IATP conference in Clearwater |
| 5 Nov 2026 | Valo Motion | Finland | Groundfall runs “without staff”; 19 configurations from 20 to 49 m²; 3-minute sessions for up to 10 players |
| 5 Nov 2026 | Lucky Strike Entertainment | US | Location payroll of $78.3M on revenue of $302.6M, across 366 venues |
| Nov 2026 | SEGA Amusements International | UK / Japan | Two-player attendant-free VR attraction; no straps; controller docks for automatic calibration |
| 2026 | Inowize (QBIX) | Canada / Japan | Six-player room at 169 sq ft, “fully unattended”, kiosk entry, sessions under five minutes |
What the rows share is not a technology. It is the claim that a paying guest can complete a cycle without an employee standing next to the hardware. What they do not share is any statement of how many hours that removes, because none of the five publishes one.
No vendor on that list publishes labor hours per 1,000 plays, mean time between interventions, or the share of sessions that end with a guest asking a human for help. Those are the numbers a roster is built from, and they are missing from every product page in this group.
So the buyer has two options: take the claim, or build the number. The rest of this article builds it, from published cycle times and one public payroll line.
2. What the attendant actually does in a ten-hour shift
“Attendant-free” is a hardware statement. It says the unit does not require an intervention to start, and it says nothing about the venue’s staffing plan. The gap between those two sentences is where budgets go wrong.
Break the job into its parts and the picture clarifies. An attraction attendant does six things, and the 2026 signals attack four of them.
| Task | Time it occupies | What the 2026 signals do about it |
|---|---|---|
| Sell and admit | Constant at peak, zero off-peak | Removed. The ROLLER agent handles the booking; QBIX admits through a kiosk |
| Brief the rules | 30 to 90 seconds per group | Partly removed. Groundfall starts when players step in; horror content still needs an age gate |
| Fit and adjust the hardware | 1 to 3 minutes per player where there are straps | Removed. SEGA sells “no straps”; Valo sells “no wearables” |
| Reset and calibrate | 30 seconds to 2 minutes per session | Removed. The SEGA controller docks magnetically; Valo has no floor sensors to recalibrate |
| Watch the users | All operating hours | Not addressed by any of the five companies |
| Clean and inspect | Between sessions | Not addressed publicly by any of the five companies |
The two rows at the bottom are the ones that decide whether the saving survives contact with a Saturday. Supervision is an operating obligation, and it sits with the venue rather than the cabinet. ASTM F2970 covers the design, manufacture, installation and operation of trampoline courts; EN ISO 23659 covers trampoline parks as facilities. Neither standard is written around a single attraction, and neither exempts a unit because the unit needs no attendant to start. Which obligations apply, and at what staffing level, is a question for the operator, the insurer and the local inspector, not for the purchase order.
Cleaning is the quieter problem. A two-player VR cabinet that turns a group every three minutes serves roughly 20 groups an hour. Whatever the hygiene routine is between groups, it runs 20 times an hour, and no manufacturer in this group has published a protocol or a time allowance for it. On a shared headset that is not a detail.
The line to hold on to: a unit that needs no attendant to start is not a unit that needs no one. It is a unit whose labor has been concentrated into fewer, more skilled hours. Concentrated labor is cheaper only if the hours actually fall.
3. The arithmetic of a rider-turn
Labor per attraction is not measured in riders per hour. It is measured in hours per thousand riders, because that is the unit a schedule is written in. Riders per hour tells you how busy the machine looks. Hours per thousand tells you how much of a person it costs to keep it busy.
Session turns per hour = 60 ÷ session length in minutes Riders per hour = turns × players per session Operating hours per 1,000 riders = 1,000 ÷ riders per hour Attendant cost per rider = loaded hourly rate ÷ riders per hour
Run the published numbers through it. Groundfall’s large configuration seats 10 players with 3-minute sessions, so 20 turns an hour and 200 players an hour. ValoBoard publishes 70 to 74 riders an hour in its gated mode, with 1 to 2 players per session; its IAAPA debut is covered separately here. QBIX publishes 6 players and sessions under five minutes, which puts it between 72 and 90 riders an hour depending on where inside “under five” the real cycle lands.
| Unit | Footprint | Players per session | Published throughput | Operating hours per 1,000 riders | Attendant cost per rider at $20 an hour |
|---|---|---|---|---|---|
| Groundfall (large) | 48.7 m² | 10 | 200 an hour | 5.0 | $0.10 |
| ValoBoard | 15 m² | 1–2 | 70–74 an hour | 13.5–14.3 | $0.27–$0.29 |
| QBIX | 15.7 m² (169 sq ft) | 6 | 72–90 an hour (our arithmetic) | 11.1–13.9 | $0.22–$0.28 |
| House of the Dead Reawakened | Not published | 2 | Not published | Not computable | Not computable |
The interesting column is not throughput. It is the fifth one, and it is not explained by size. Convert the same units to riders per square meter per hour and they land within 45 percent of each other: Groundfall at about 4.0 to 4.1, ValoBoard at 4.9, QBIX at 4.6 to 5.7. Density is nearly flat across the three.
What separates them is cycle length. A three-minute cycle for ten players produces 200 rider-turns in the hour a single attendant would otherwise be paid for. A two-to-three-minute cycle for two players produces 48 to 60 by the same arithmetic, and the published 70 to 74 requires faster turnover than the session length implies. Same footprint class, same density, and a labor difference of nearly three to one.
That is the purchasing lesson in one sentence: if the goal is fewer labor hours, specify cycle time rather than floor area. Density decides what the rent buys. Cycle time decides what the schedule costs.
It also sets the ceiling on the whole argument. At $20 an hour fully loaded, the attendant share of a rider is 10 cents on the fastest unit here and 27 cents on the slowest. Against a ticket price in the single or low double digits, that is a real margin line, but it is not the difference between a viable attraction and an unviable one. It is the difference between two operating models that both work.
4. Where the numbers do not close
Three of the four units above carry a gap between their own published figures, and each gap is worth naming before a deposit.
ValoBoard’s throughput and session length disagree. The page states sessions of 2 to 3 minutes for 1 to 2 players, and separately states 70 to 74 riders an hour in gated mode. Two riders per session over a 2.5-minute average produces 48 an hour. Reaching 74 requires an average cycle of roughly 1.6 minutes, or a second board. Valo has not published the cycle time its figure assumes, so treat 74 as a ceiling reached in a mode most venues will not run all day.
House of the Dead Reawakened publishes no session length at all. Two seats, three missions, a cabinet with an 85-inch entrance screen and two 43-inch side displays, and a twelve-month warranty on the headset cable harness, load-tested to 130 kg. The one number a labor model needs, how long a session runs and how long a reset takes, is absent. A two-seat cabinet is the most cycle-bound format in this group, so the missing figure is the decisive one.
QBIX says “under five minutes” and leaves the range open. That is a 25 percent spread in throughput before any uptime assumption, which is why the table shows a range rather than a figure.
Then there is uptime, which nobody quotes. Published throughput is theoretical capacity. At 70 percent availability, Groundfall’s 200 an hour becomes 140, and 5.0 hours per thousand riders becomes 7.1. At 50 percent it becomes 10.0, close to a staffed ValoBoard. The labor case for any unstaffed unit is therefore an uptime claim in disguise, and uptime is the one thing a brochure cannot promise.
We are not going to pretend these gaps can be closed from outside. They cannot. They can be closed in a tender document, by asking four questions: median session length, median reset time, mean time between interventions, and the percentage of sessions that require a human. A supplier who answers all four in writing has given you a labor model. One who answers none has given you a floor plan.
5. What 366 venues already spend on the person
The signal from the operator side is a single line in a filing. Lucky Strike Entertainment reported location payroll and benefit costs of $78,263,000 for the quarter ended 27 September 2026, on revenue of $302,644,000, across 366 venues. Payroll is 25.9 percent of revenue, and it was 25.7 percent a year earlier on $292,278,000. The share is stable, not falling.
Payroll per venue per quarter = $78,263,000 ÷ 366 = $213,833 Payroll per venue per day (91-day quarter) = $213,833 ÷ 91 = $2,350 Labor hours per venue per day = $2,350 ÷ loaded hourly rate
| Loaded hourly rate (assumption) | Labor hours per venue per day | Full-time equivalents at 8-hour shifts |
|---|---|---|
| $20 | 117.5 | 14.7 |
| $22 | 106.8 | 13.4 |
| $25 | 94.0 | 11.8 |
The rates in the first column are an assumption, not a disclosure. Lucky Strike does not publish its loaded rate, so the honest range for a US multi-attraction venue is somewhere between 94 and 118 labor-hours a day, and the true number depends on the market, the mix of full-time and part-time posts, and the benefits load. Anyone who quotes an exact figure without that input is guessing.
Read against that base, the scale of the change becomes concrete. One unstaffed attraction open 12 hours a day, displacing one post, removes 12 hours. That is 10 to 13 percent of a venue’s daily labor hours. Four such units clustered so that one roaming marshal covers all four might displace two posts, which is 24 hours, or 20 to 26 percent.
Two caveats keep that from becoming a promise. First, the payroll line includes every venue employee, food and beverage staff included, so it is an upper bound on what any one attraction’s attendant post is worth. Second, the four units do not remove the marshal. They change the marshal’s job from standing at a court to walking a loop, which is a lower headcount at the same supervisory obligation.
That is also why the labor question is really a layout question, and the next section treats it that way.
6. The layout consequences nobody puts in the brief
An unstaffed unit changes the drawing in five ways, and each one is cheaper to fix before the slab is poured than after.
6.1 The entry stops being a counter
What it is Without an attendant, admission moves to a kiosk, a gate or a scan, and the physical counter disappears from the plan.
What it decides Where the queue forms, and whether one person at a nearby point of sale can see it. QBIX is built around kiosk entry; ValoBoard is sold as ready for gated admission. Both assume the gate is somewhere a staff member already stands.
What to verify That the gate is inside the sightline of an existing post. If it is not, you have added an unstaffed attraction and a new supervision point at the same time. The geometry of that problem is covered in our note on court sightlines and blind spots.
6.2 Cluster the units, do not scatter them
What it is The saving comes from one person covering several units, so the units have to be coverable by one pair of eyes and one pair of legs.
What it decides Whether the labor reduction is 10 percent or 25 percent of the venue’s hours. Four units in four corners remove nothing; four units on one wall remove two posts.
What to verify Walking distance between units at shift change, and whether the cluster is visible from a single standing position. Our zone-mix comparison across four park floor lists shows how often attraction clusters end up split by circulation.
6.3 Ceiling and services are per unit, not per zone
ValoBoard needs 15 m² at a minimum ceiling of 2.7 m, on a 3.6 by 4.2 m footprint, and the page states that no structural changes to the building are required. Groundfall runs from 20 to 49 m² indoors, out of direct sunlight, on a wired internet connection of 10 Mbit. Neither figure is demanding on its own. Together with a VR cabinet that wants a wall for an 85-inch screen, they set a clear-height and services pattern for the cluster, and that pattern argues for putting the cluster where the ceiling is already generous rather than where the floor plate happens to be empty.
6.4 Reserve a reset bay outside the play area
The saving depends on failures being fixed fast. A controller that needs charging, a headset that needs wiping, a session that has to be restarted: all of it needs a surface, a socket and a place to stand that is not inside the attraction. Six to ten m² per three unstaffed units, adjacent to the cluster and visible from the marshal’s loop, is the cheapest insurance in the plan. Put it inside the play area and you have created a staffed zone by accident.
6.5 Power and data per unit
Every one of these attractions is a connected device. Wired internet at 10 Mbit is a published requirement for Valo’s products, and remote diagnostics are how a small team keeps four units alive. Daisy-chaining a cluster onto one drop means one cable failure takes out the whole cluster, which is the exact failure mode the unstaffed model cannot absorb.
7. What would prove this wrong
A trend claim is only worth what its disconfirming test is worth. Five tests would each, on their own, break the argument that attendant-free hardware reduces venue labor.
Test one: hours per thousand admissions do not fall. If a venue installs four unstaffed units and total labor hours per 1,000 admissions are flat two quarters later, the hours moved rather than left. They moved into exception handling, cleaning and guest help, which is where unmeasured labor usually hides. The measurement is simple and almost nobody runs it: labor hours divided by admissions, before and after.
Test two: the insurer requires a named attendant. A single underwriting condition that names a staffed presence per attraction cancels the hardware saving at the insurance line, and it is the least visible risk in this argument because it never appears in a product brochure. Ask the broker in writing before the deposit, not after.
Test three: one Saturday outage costs more than the wage saved. A unit at 74 riders an hour that sits down for two hours on a Saturday costs 148 plays. At $5 a play that is $740, roughly a week of the wage saved by removing one 40-hour post at $20 an hour. At $10 a play a single two-hour failure erases two weeks. Unstaffed hardware is only cheaper if somebody notices the failure quickly, which means remote alerts or a marshal on a loop, and both cost money.
Test four: the traffic is mostly under-six. If the site’s weekday traffic is parents with toddlers, the host comes back, because a five-year-old cannot start a session unattended and a parent will not read a screen at the gate. The labor returns at a lower skill level and a lower wage, which is a different saving from the one being sold.
Test five: the vendor cannot answer the four questions. No median session length, no median reset time, no mean time between interventions, no share of sessions needing a human. Without those, the labor claim is a design intention, and the buyer is the one carrying the risk.
Two things remain genuinely unknown, and we are not going to fill them in. SEGA has not published a session length for House of the Dead Reawakened, so its throughput cannot be compared with the others. No manufacturer here has published a cleaning protocol for shared hardware, so the per-session hygiene allowance is unknown in every case. Both gaps are small in a brochure and large in a staffing plan.
8. If this holds, what a 2027 build should reserve
The practical response is not to buy more unattended machines. It is to write the labor model into the drawing, where it can be changed cheaply, instead of into the schedule, where it cannot.
| Condition on your project | Specify unstaffed | Keep it staffed |
|---|---|---|
| You need 60 or more rider-turns an hour inside 20 m² | Yes. Cycle-bound formats win on hours per thousand | No. A staffed format at that footprint cannot match it |
| The content needs a briefing or an age gate | Only with a greeter at peak hours | Yes, if the gate is content-led rather than age-led |
| One marshal already covers three zones | Yes, and cluster the units on that loop | No. A fourth supervision point without a loop is a liability |
| Most weekday traffic is under six with a parent | No | Yes. The host returns anyway, at a lower wage |
| The insurer names a staffed presence per attraction | No | Yes. The saving is cancelled at underwriting |
| The contract includes remote diagnostics, spare parts and a service response time | Yes | Only if you can absorb two-hour outages at peak |
| The quote has no reset time, no uptime figure and a 90-day warranty | No | Yes. You are buying the hardware, not the labor model |
Then reserve four things in the drawing, none of which is expensive at design stage and all of which are expensive later: a 6 to 10 m² reset bay per three unstaffed units, positioned outside the play area and inside the marshal’s loop; a single sightline point that covers the whole cluster; a dedicated power and wired data drop per unit rather than per cluster; and a clear height of at least 2.7 m over the cluster, which is what ValoBoard asks for and what any future cabinet in this class is likely to want.
Finally, write the staffing plan into the operations manual before opening, not after. One marshal per cluster, one response time for a fault, one hygiene interval per shared device. That document is what an inspector, an insurer and a buyer will all ask for, and it is also the only place where the labor saving becomes an auditable number rather than a vendor claim.
If you are sizing a cluster for a specific floor plate, the design-to-delivery process starts with measured building information rather than a catalogue. Our attraction accessories range covers the gating, seating and service points that an unstaffed cluster needs around it, and the safety standards page sets out how we document the operating obligations that stay with the venue.
FAQ
Do attendant-free attractions actually reduce staffing costs?
They reduce the hours attached to the machine, not the hours attached to the venue. On the published figures in this article, an unstaffed unit removes between 5 and 14 operating hours per 1,000 riders at full capacity, which is roughly 10 to 26 percent of a multi-attraction venue’s daily labor hours when several units sit under one marshal. Supervision and cleaning stay, and both are staff time.
What does attendant-free mean in an equipment specification?
It means the unit does not need an employee to start, reset or calibrate a session. SEGA achieves it with no straps and a controller that docks magnetically for automatic calibration. Valo achieves it with camera tracking, no wearables and no sensors in the floor. Neither phrase is a staffing plan, and neither supplier publishes the labor hours its product removes.
How much floor space and ceiling height do these units need?
Groundfall runs from 20 to 49 m² across 19 configurations, with the largest at 6.20 by 7.85 m for 10 players. ValoBoard needs 15 m² at a minimum ceiling of 2.7 m, on a 3.6 by 4.2 m footprint. QBIX is quoted at 169 sq ft, about 15.7 m². SEGA has not published a footprint for House of the Dead Reawakened. All of them need power and a wired internet connection at the unit.
Does ASTM F2970 or EN ISO 23659 exempt an attraction with no attendant?
No. ASTM F2970 addresses trampoline courts as systems, including operation, and EN ISO 23659 addresses trampoline parks as facilities. Both place operating obligations on the venue rather than on the cabinet, and neither is written around a single attraction. The two are separate frameworks with different methods, so a project cites one or the other according to its market, never a blend. Exact staffing levels are a matter for the operator, the insurer and the local authority, and they vary by jurisdiction.
How many labor hours should I budget per 1,000 plays?
Start from the published cycle time: 60 divided by session minutes, multiplied by players per session, gives riders an hour, and 1,000 divided by that gives the operating hours per 1,000 riders. For the units in this article the answer falls between 5 and 14 hours at full capacity. Then adjust for uptime, because at 70 percent availability the 5-hour unit becomes 7 hours, and add an allowance for cleaning and exception handling, which no supplier publishes.
Sources
- Valo Motion, “Groundfall” product page, configurations, session length and staffing statement, accessed 7 November 2026 — valomotion.com/groundfall
- Valo Motion, “ValoBoard” product page, 15 m² footprint, 2.7 m ceiling and 70–74 riders an hour, accessed 7 November 2026 — valomotion.com/valoboard
- SEGA Amusements International, “Dare to Experience: House of the Dead Reawakened”, attendant-free format and 130 kg cable test, November 2026 — segaarcade.com
- Amusements & Attractions (InterGame), “House of the Dead Reawakened to launch at IAAPA Expo”, November 2026 — amusementsandattractions.com
- ROLLER, “ROLLER Doubles Down on AI with New Agent to Help Operators Grow”, 27 May 2026, including 2026 Pulse Report figures — roller.software
- IATP / GTI Amuse, “IATP 2026 Annual Conference Brings Adventure and Trampoline Park Industry Together”, September 2026, Breakthrough Product award — gti-amuse.com
- Lucky Strike Entertainment, “First Quarter Results for Fiscal Year 2027”, Form 8-K Exhibit 99.1, 5 November 2026 — sec.gov
- Inowize, “QBIX is Built for the Way You Actually Operate in 2026”, operator references and unattended configuration — inowize.com
- ASTM International, F2970 standard scope for trampoline courts: store.astm.org
