Industries / Media, arts and recreation / Escape rooms
Props and electronics
Props and electronics are the objects players handle and the hidden wiring that notices what they do and opens the next lock or door.
4
Milestones achieved, from 2015 to 2019.
1
To watch for: announced, not yet confirmed outside.
2
Open deliverables. Nobody has shipped them.
The road
AchievedWatch forOpen
Achieved
2015
Padlocks, paper and black light
About 70% of the games described in Nicholson's 2015 survey were wholly physical, with no computer: 86% of those from Asia and 58% of those from Europe. He described the usual tasks as paper puzzles, counting items in pictures, entering codes on padlocks and shining a black light, an ultraviolet lamp that shows hidden ink. Where a room had a computer it was mostly a puzzle in itself, and European rooms, often smaller, used computers to add layers they had no space for.
Credit
survey respondents; Scott Nicholson
2015
Hidden boards run the room
Eight Arduino Mega boards, small programmable circuit boards popular with hobbyists, ran Palace Games' Great Houdini room from a closet behind one wall, wired to lights, infrared sensors and magnetic locks through thousands of feet of cable. Alden had written more than 5,000 lines of code for about 300 input and output pins. To switch mains-powered lights and motors from the boards' low-voltage outputs, his electrician, Andrew Florek, had the boards turn the outlets of a power strip on and off.
Credit
Christopher Alden, Palace Games; Andrew Florek
2015
A screen for hints
On 12 April 2015 Escape Room Master released software that puts a countdown timer and a game master's text, audio, photo and video clues on a television inside the room, sent from a laptop by cable or a Google Chromecast. The company priced a set at about $150 for a computer, $120 for a TV and $10 to $35 for the link, and recommended one game master per room. Nicholson's survey the same year found hints given by buzzer, knock, walkie-talkie, microphone and computer.
Credit
Escape Room Master
2019
A floor that keeps score
In January 2019 Megan and Adam Schmidt, who already ran an escape room called The Real Escape, opened Activate Games on Portage Avenue in Winnipeg, Canada. Its 11 rooms set teams physical and mental tasks with no controllers; in one, the floor is a lit grid that serves as the game board, and players log in with a wristband. "The problem with solving a puzzle is once it's solved, it's solved forever," Adam Schmidt said. "We wanted to make an active, replayable facility." Room Escape Artist, which files such venues as challenge arcades, counted Activate's US sites rising from 9 to 28 in 2024 and to 48 in 2025.
LessonA puzzle, once solved, cannot be sold to the same team again, so Activate built games the same players come back to.
Credit
Megan and Adam Schmidt, Activate
Watch for
Cameras that check the reset
Drawbridge, which sells room-control software, describes a Reset Scan feature that photographs a room about 10 seconds after staff press a reset button and compares chosen areas with reference pictures of a correctly reset room. Each area gets a similarity score, and the default pass mark is 0.7.
Achieved whenan outside test reports how often the check catches a wrongly reset room and how often it raises a false alarm.
Claimed by
Drawbridge, in a help article published in March 2026
Open
01
Test the reset check
An independent test of how well camera checks and staff checklists catch a wrongly reset room. Rooms are left with one planted error, such as a lock left open, a key out of place or a drawer not shut, and each method is scored.
The briefCredit · unclaimed02
A fault log for props
A shared, anonymous log of prop and tech failures from many venues: which part failed (a magnetic lock, an RFID reader, which reads tagged objects by radio, a reed switch, a cable or a power supply), after how many games, and how long play stopped.
The briefCredit · unclaimed
Open 01 · the brief
Test the reset check
The first to publish is credited on this line, and the deliverable moves from Open to Achieved.
What to build
An independent test of how well camera checks and staff checklists catch a wrongly reset room. Rooms are left with one planted error, such as a lock left open, a key out of place or a drawer not shut, and each method is scored.
Why it is still open
Drawbridge's Reset Scan compares camera images with reference pictures against a similarity score that defaults to 0.7, but its help pages give no error rates, and Ken Ferguson's guide to handling breakages and mis-set rooms, on his site The Logic Escapes Me, gives advice, not rates.
Built with what exists
The room's existing security cameras; image comparison in OpenCV; a multimodal model asked yes-or-no questions about each frame (see the Note "A thousand questions, one pass"); a small open vision model fine-tuned on labelled photos of the room (see the Note "One GPU to fine-tune"); the venue's paper reset checklist as the baseline.
First step
Photograph one room correctly reset 30 times in different light, then 30 times with one planted error each.
Done looks like
200 planted errors across at least five rooms, with the catch rate and false-alarm rate of each method and their 95% confidence intervals, published with the photos.
Open 02 · the brief
A fault log for props
The first to publish is credited on this line, and the deliverable moves from Open to Achieved.
What to build
A shared, anonymous log of prop and tech failures from many venues: which part failed (a magnetic lock, an RFID reader, which reads tagged objects by radio, a reed switch, a cable or a power supply), after how many games, and how long play stopped.
Why it is still open
The one room Make described in 2015 ran about 300 input and output pins through thousands of feet of cable. Ferguson writes that every room will fail at some point, "perhaps yours will come one in a hundred games as opposed to one in ten", without data to say which, and Room Escape Artist warns that older games take wear from players. No published failure rates were found.
Built with what exists
Fault notes typed or spoken by game masters into a shared form, with speech-to-text such as ElevenLabs Scribe v3 or OpenAI's Whisper for the spoken ones (see the Note "Fewer word errors"); control-system event logs to time each stoppage; Python for the tables.
First step
Agree a ten-item list of part types with three venues and log every fault for one month.
Done looks like
Six months of logs from ten venues, giving failures per 1,000 games for each kind of part and the median minutes lost, published.