Containment Engineering Standard T-11, with Cthulhu Eternal statistics
Failure report: Harker Deep Shelter
Harker’s original containment cell had exactly one wall between the specimen and the control room: 1.4 metres of irregularly poured magnetite concrete, with no door, duct, pipe, cable, or continuous reinforcement crossing the boundary. Cameras watched both faces. Food, instruments, and sedated test animals came in through a remote transfer system that never opened both shutters at once. The specimen tried to localize eleven times in six weeks, and every attempt failed. The review board called that an unacceptable margin.
So they installed a second shell inside the first, thinner but independently founded, and between the two shells the engineers left a thirty-centimetre service cavity so neither structure touched the other. The cavity was evacuated, instrumented, and lit. The final report described the design as two barriers in series and therefore safer than either barrier alone. At 02:13 on the first night of operation, the outer cameras showed the specimen standing against the old wall. The inner cameras showed an empty cell. At 02:14, a blue condensation appeared on the far face of the old wall, inside the service cavity, its outline too shallow to be an animal and too organized to be moisture, and the cavity instruments logged one pressure impulse while recording no gas. At 02:15 the duty engineer brought up the high-rate observation array. The condensation vanished. So did the specimen, from the outer cameras. Both walls stayed intact. At 02:16, something started breathing behind the control-room technicians. The second shell had supplied the intermediate state that made the passage easier.
The rule
A Tunneller crosses a wall when its exterior body stops being the stable expression of its angular state and a new body nucleates on the far side. No tiny animal worms through the concrete, and if you drilled into the barrier you would never find it halfway across. This distinction earns its keep in play.
- Thickness and material matter, but not as simple distance travelled.
- A cavity inside a barrier can help it rather than hinder it.
- A camera may suppress localization, accelerate it, or just document the result.
- Cracks, cables, ducts, and repeated architectural layers change the available configurations.
- Once it has fully localized in a room, it has a finite body that can be injured and trapped.
The wall stays structurally intact while the creature is outside the wall decays into the creature is inside it.
What quantum tunnelling is not
Popular explanations love to say a particle “borrows energy” to cross an impossible barrier, and applied to a monster that turns into a ghost animal briefly ignoring matter. Neither picture is any use here. Quantum tunnelling is not a licence to violate conservation of energy. A state can have nonzero amplitude across a classically forbidden region even when no classical trajectory carries it through, and in a simple one-dimensional problem the wavefunction inside a barrier decays about as
That same familiar formula is why ordinary macroscopic animals do not tunnel through walls: the exponent grows with mass, with barrier height, and with distance. For a warm, many-particle body interacting nonstop with its environment, coherent centre-of-mass tunnelling is not merely unlikely; the naive probability is suppressed so catastrophically that calling it “rare” undersells it. Their passage relies on a different kind of tunnelling entirely, not on some improbable centre-of-mass hop by an ordinary animal.
A body is not always the coordinate that tunnels
Macroscopic quantum tunnelling is real when the tunnelling degree of freedom is a collective coordinate: one variable summarizing the organized state of a much larger system. In a Josephson junction the relevant coordinate can be the phase difference across a macroscopic quantum state, and in superconducting devices, flux or phase configurations holding the coordinated behaviour of enormous numbers of microscopic constituents can escape a metastable state by tunnelling. The apparatus does not shove every electron through one miniature passage like a classical crowd. A collective configuration changes.
Field theory gives an even better model for the Tunneller. A field can sit in a metastable state, a local minimum that is not the true lowest-energy configuration, and decay through a special solution found in imaginary, or Euclidean, time. This bounce, or instanton, describes the nucleation of a region of the new phase. The leading decay rate has the form
![\Gamma\simeq A e^{-B},\qquad B=\frac{S_E[\phi_b]-S_E[\phi_f]}{\hbar}](https://s0.wp.com/latex.php?latex=%5CGamma%5Csimeq+A+e%5E%7B-B%7D%2C%5Cqquad+B%3D%5Cfrac%7BS_E%5B%5Cphi_b%5D-S_E%5B%5Cphi_f%5D%7D%7B%5Chbar%7D&bg=ffffff&fg=000&s=0&c=20201002)
where 
The Tindalosian axiom
A Tunneller’s ordinary body is the stable curved-space phase of an angular localization field. The collective coordinate deciding where that phase condenses stays protected in angular space, and it can tunnel between metastable configurations without an action proportional to the mass of the completed body.
Say 
not
The completed body holds ordinary mass and obeys ordinary conservation laws. During localization, energy and momentum are traded with the wall, the air, the electromagnetic environment, and the angular sector, and that exchange throws off the characteristic pressure impulse, thermal bloom, and blue residue. The creature does not conjure a free body’s worth of matter from nothing; it changes where its projected phase is stable and settles the energetic account through couplings ordinary matter does not have. This axiom is the single fictional step. Everything after it is an attempted consequence of it.
Configuration space, not floor plan
Containment engineers make their worst mistakes by drawing a building and asking how the creature could move through it. The map that matters is a landscape of possible field configurations. Two rooms side by side can be extremely far apart in configuration space if their boundary conditions differ, and two physically distant chambers can be close if they support nearly identical localization modes and share a coupling. That is why a Tunneller may fail over and over against a rough concrete wall and then localize two metres into a carefully manufactured cell whose dimensions mirror its current enclosure: the destination is easy to nucleate because the two body-configurations resemble each other. Useful containment therefore leans on detuning. Make opposite faces geometrically dissimilar. Avoid repeated cell dimensions. Break continuous material layers. Vary density, texture, temperature, and electromagnetic response irregularly. Remove cavities that could hold an intermediate localization. Keep pipes, reinforcement, sensor cables, and air columns from coupling both sides. A massive perfect shell can perform worse than a thinner heterogeneous one. The question was never just how much matter lies between the states. It is whether the structure hands the creature a low-action route connecting them.
The thing in the barrier
Failed localization attempts leave evanescent signs. These are not pieces of an animal trapped in the wall. They are exponentially decaying correlations between the current body and a destination state that never became stable. Common ones:
- Blue film spreading across the destination face with no source.
- A pressure pulse too brief to be displaced air.
- A warm outline whose deepest point stays right at the surface.
- Scratching picked up by contact microphones but absent from the room.
- A transient rise in wall mass balanced by cooling somewhere else.
- Parallel impressions on both faces with no fracture connecting them.
- A smell of wet limestone, ozone, or opened blood, only on the destination side.
The outline can include a limb, a mouth, or a sensory organ, and that does not mean one appendage has pushed through the barrier. A subcritical bubble of the creature’s projected phase has formed on the far side, and inside that temporary region a fragment of anatomy can become completely real, act once, and then collapse. There is still nothing inside the wall. Investigators who grasp this can use failed attempts to measure the route without ever opening the barrier. Investigators who do not tend to drill toward the sound.
Resonant containment failure
A second barrier usually cuts transmission, but a resonant intermediate region can flip that advantage. In ordinary resonant tunnelling, two barriers enclose a region that supports a state at the right energy, and instead of one severely suppressed transition the system couples efficiently through a matched intermediate state. The exact physics depends on coherence, loss, geometry, and the nature of the system; “two walls are five times worse” is not a universal law. For Tunnellers, an empty service cavity, a double hull, a wall void, an airlock, or a nested shell can support a metastable projected body,
and if 
Watching the wall
The quantum Zeno effect is another idea shorthand has mangled. “A watched pot never boils” does not mean consciousness freezes quantum systems, and ordinary CCTV does not automatically prevent tunnelling. Repeated or continuous measurement can inhibit a transition when the apparatus couples strongly and appropriately to the state being tested, but measurement also changes the system’s coupling to its environment and broadens or reshapes its energy response, so under other timing, bandwidth, and bath conditions observation can accelerate the decay: the anti-Zeno effect. Three regimes matter for containment.
Passive observation. Cameras record visible surfaces after photons have scattered off them. If they do not couple to the localization coordinate, they give evidence but almost no back-action. The Tunneller is not impressed by being on television.
Zeno lock. A calibrated array repeatedly distinguishes the outside configuration from every measured destination state faster than the transition can develop. It has to monitor the actual order parameter, the matched pressure, phase, field, mass-distribution, and boundary probes established during failed attempts, and while it stays within its tested bandwidth, localization along that route is suppressed.
Anti-Zeno drive. A badly timed array, a noisy backup, an intermittent pulse sequence, or a detector with the wrong spectral response can strengthen the coupling to accessible final states. It hands the transition more useful opportunities instead of interrupting it. Which is why the power transfer at Harker matters more than the outage did. The main array had been calibrated as a Zeno lock. The backup array sampled at a tenth the frequency and swept through the interstitial cavity’s resonance, turning observation into assistance.
Transmission Rating
At the table, compress the whole decay calculation into a Transmission Rating: the percentage chance that one meaningful Localization Attempt produces at least a partial projected body across the named barrier route. If the physical rate is 
Nobody needs to evaluate that in play. The equation just says what the rating means: a probability over one consequential attempt, not a chance per second that the GM keeps rolling until the monster inevitably arrives.
| Established route | Base Transmission Rating |
|---|---|
| Interior partition, ordinary door, vehicle skin | 30% |
| Heavy masonry or structural wall | 15% |
| Reinforced, isolated containment shell | 5% |
| Massive heterogeneous barrier with no shared penetrations | 1% |
| Calibrated Zeno lock | No attempts while maintained |
| Resonant intermediate state | Multiply final rating by 5, maximum 95% |
These are scenario values, not universal material constants. The GM sets the initial number from the fiction, records it, and lets investigation reveal it. Apply modifiers only when the physical route actually changes.
| Change | Effect |
|---|---|
| Shared pipe, cable, continuous reinforcement, or open air column | +10% |
| Destination closely duplicates the origin’s geometry | +10% |
| Active angular residue or Corner Brood in the barrier | +10% |
| Irregular layering or deliberately mismatched faces | −10% |
| Shared penetrations disconnected and isolated | −10% |
| Destination geometry changed after the route was mapped | Halve rating |
| Verified resonant cavity detuned or filled heterogeneously | Remove multiplier |
Minimum 1% unless a functioning Zeno lock is in place. A route below 1% is practically closed for the scene; do not keep rolling hoping for a dramatic critical.
Do not roll once a minute. Make a Localization Attempt only after a meaningful coupling event: the Tunneller touches, strikes, or maps the barrier; a door, shutter, pump, or transfer system changes the boundary conditions; the Protagonists activate a probe spanning both sides; a pipe, cable, duct, or continuous structural element starts carrying energy across the route; the monitoring regime changes; the destination becomes newly hospitable or geometrically matched; or the Protagonists deliberately lure the creature toward a prepared state. Announce the risk before the action when the characters could reasonably see it: “cycling that shutter will give it another localization attempt.” The threat is in choosing under pressure, not in hiding the rules.
Roll percentile against the current Transmission Rating. A failure means no stable projection forms; leave an evanescent sign that can be studied. A success above half the rating forms a subcritical Partial Localization that acts once and collapses. A success at or below half the rating is a Full Localization: the outside body disappears and a complete body nucleates at the destination. A critical is Occupied Emergence, where the stable phase nucleates where solid matter or a living body already is. If a 5% route succeeds at all, it is necessarily a full or critical localization. Weak routes fail repeatedly and then fail catastrophically; they do not produce a polite succession of paws.
Occupied Emergence
A Tunneller cannot coexist stably with matter filling the same localization volume, so when a critical attempt puts it inside machinery, a wall, or a person, the competing phases resolve violently. Everyone in the emergence volume faces a 20% Lethality attack; on survival, take the sum of the dice as damage and get thrown clear. The Tunneller takes 1D10 ignoring Armour. Machinery is destroyed, and a structural emergence may breach the barrier and open a conventional route. This is not its preferred attack. Occupied Emergence risks the creature and leaves it tangled in finite matter, so it does this when desperate, driven, or unable to find an empty compatible destination. SAN loss for witnessing it is 1/1D10, and a survivor also tests SAN for Helplessness or Violence as appropriate.
Tunneller
Angular localization predator, involuntary phase organism
Fully projected, a Tunneller is a low, long organism built around boundary contact. Its limbs are broad plates that meet floors and walls at shifting perpendiculars, no joint crossing smoothly from one orientation to another; each segment belongs to a different local surface. A dark sensory membrane runs the length of it and ripples when it maps nearby configurations. It is opaque, casts an ordinary shadow, displaces air, bleeds blue fluid, and leaves weight-bearing tracks, and its solidity is exactly what makes the transit worse: this much animal can be absent from one room and complete in another without ever occupying the distance between. Roll each specimen or use the typical column.
| Characteristic | Roll | Typical |
|---|---|---|
| STR | 4D6+6 | 20 |
| CON | 4D6+4 | 18 |
| DEX | 4D6+3 | 17 |
| INT | 3D6 | 10 |
| POW | 4D6+4 | 18 |
| HP | (STR+CON)÷2 | 19 |
| WP | = POW | 18 |
Movement: 12 metres. Skills: Alertness 70%, Athletics 60%, Navigate 80%, Stealth 70%, Unarmed Combat 60%, Unnatural 40%.
Armour: 3 points of projected tissue. While partially localized, mundane attacks cannot target the absent body; attacks on the visible anatomy suffer −20% and can do no more than 4 HP before the bubble collapses.
vs Lethal Damage: HIGHLY RESILIENT while a mapped localization route is available; a successful Lethality roll does damage equal to the Lethality rating. NORMAL once fully localized and every mapped route has been detuned, Zeno-locked, or made inaccessible.
SAN Loss: 0/1D6 to see a complete Tunneller; 1/1D8 to watch a body disappear and nucleate somewhere else.
Attacks
Finite Bite 60%, damage 1D10, Armour Piercing 3. A complete jaw closes in one stable location. On a critical the victim is pinned and loses their next action unless they pass STR×5 or Unarmed Combat.
Evanescent Strike 50%, damage 1D8, ignores worn armour. Only through Partial Localization. A temporary mouth, plate, or cutting surface condenses beside the target, attacks once, and collapses. Cover helps only if it is not part of the same destination state.
Boundary Rush 50%, Lethality 10%. It drives its complete projected body through a confined space, pinning victims against the boundaries that stabilize it. On a failed Lethality roll, victims take the sum of the dice. It cannot use this in open terrain.
Powers
Barrier Map. Staying in contact with a boundary for one turn, it learns the easiest connected destination and its approximate Transmission Rating, and instruments pick up a rising evanescent signature. Damaging it or substantially changing the boundary interrupts the map.
Localization Attempt. After a meaningful coupling event, the GM rolls against the established route. The attempt itself costs no WP; maintaining an imposed projection does.
Partial Localization. A high success forms a subcritical bubble and allows one Evanescent Strike, sensory observation, or manipulation on the destination side. It spends 2 WP. Damage to the projection transfers to the shared body before the bubble collapses.
Full Localization. A low success replaces the current body with a complete destination body. It spends 4 WP and loses its next action to projection shock unless the roll was critical. A thermal pulse, blue vapour, and pressure wave make the arrival obvious.
Tunnelling Retreat. Once per combat, a fully localized Tunneller may attempt a previously mapped route as its action, spending the WP only if the roll succeeds. If the route has been detuned since mapping, halve its rating. If every route is closed, it must stay and can be killed normally.
Metastable Physiology. At 0 HP the body collapses into blue condensate unless all mapped routes are closed. If a route remains, test its Transmission Rating once; on a full success it re-nucleates there with 6 HP and 0 WP, and on a failure it dies in curved space.
Containment procedure
Containment is about controlling the available state landscape, not finding the thickest wall.
Survey. Establish the origin state, every plausible destination, and every coupling between them. Architecture, utilities, material composition, temperature, pressure, active electronics, and angular contamination all matter. A successful Science, Engineering, Search, or Unnatural test reveals one of: the current Transmission Rating of a named route; one modifier currently raising or lowering it; whether an interstitial state is approaching resonance; what observable a Zeno lock must measure; or which physical change would detune a mapped destination. Do not hide all five behind one specialty. An electrician can spot cable coupling, a structural engineer can find repeated geometry, a physicist can read decay statistics, an occultist can recognize angular residue.
Detune. Change the destination after the creature has mapped it. Flood a cavity with irregular aggregate. Move a wall. Alter pressure and thermal conditions. Hang nonparallel baffles. Break continuous reinforcement. Rotate internal equipment. A successful operation halves the route until it is mapped again.
Isolate. Remove shared penetrations. Fibre links, wireless cameras, independent power, staggered shutters, and sacrificial sensor heads beat cables spanning the barrier. “Disconnected” equipment still couples the spaces if it shares a ground, a pipe, a rigid mount, or an air column.
Observe correctly. A Zeno lock is a discovered technology, not a box that works because somebody said “quantum.” The Protagonists first record at least two failed attempts, identify a localization-sensitive observable, and calibrate an array with a successful Science or Engineering test. Established, it stops attempts along one named route while powered and supervised. A fumble, a damaged sensor, an untested backup mode, or an improvised change does not automatically produce anti-Zeno enhancement, but the GM should foreshadow the dangerous response in the recorded decay rates or stronger evanescent signs, and if the group proceeds anyway, the route gains the resonance multiplier until the array is shut down or corrected.
Trap. The most reliable kill is to make one destination deliberately easier than all the others: prepare an empty cell with a high but measured Transmission Rating; detune or Zeno-lock every competing route; trigger one deliberate attempt; confirm Full Localization; change the destination geometry immediately to close the return route; and fight the now-finite animal. The plan is dangerous, comprehensible, and fair. It converts an almost untouchable field transition into a monster with nineteen Hit Points, and it does so by forcing an angular organism into exactly the condition it has spent the whole incident trying to escape.
Running Tunnellers fairly
Tunnellers work best as engineering antagonists, not as a licence to ignore doors. Name the route: every attempt crosses a specific established barrier between specific states. Set the number: record the Transmission Rating before rolling. Show the failed attempts, so evanescent signs let players learn rather than just endure. Warn about coupling when an action creates an attempt and the risk is knowable. Let modifications matter, so changing geometry changes the recorded number. Earn resonance from the location’s established design. Respect calibrated observation, and do not disable a functioning Zeno lock by fiat. Allow the trap, and once the creature is fully localized and detuned, do not let it invent a new route just because combat is going badly. The investigation roll discovers what is true. The tactical decision determines what to do with it. The transmission roll resolves the uncertainty. Keep those three functions separate. The math decides why the wall fails. The players decide which side the monster is on when it does.
Operations
The Improved Bunker
A captured Tunneller stayed contained for months until a safety review ordered a second shell, and activity has since dropped to zero, which the administrators are reading as success. The Protagonists find the empty service gap accumulating blue condensation in standing-animal outlines. The original design offers a 5% route. The resonant cavity raises it to 25%. Activating the automated inspection sweep adds a shared electromagnetic coupling and pushes it to 35%. The group can flood the gap with heterogeneous foam, evacuate the control level, or use the state deliberately as a trap. Headquarters would prefer more cameras.
The Backup Array
A hospital’s experimental containment ward runs a legitimate Zeno lock built around superconducting pressure and field sensors, and rolling power failures keep handing the system to an older pulsed controller. Every transfer throws stronger signatures into rooms holding sedated patients. The patients are not bait because they are conscious. Their beds, monitors, and identical rooms provide redundant destination configurations. The group has to choose between shutting down life-support equipment that couples the rooms, maintaining a failing observation regime, and moving patients through corridors the Tunneller can map.
Three Knocks
Residents of a sealed apartment hear three impacts from inside a load-bearing wall, and when they knock back the sound moves to the opposite face. A livestream audience starts coordinating response patterns, and every call-and-response actively probes and matches a destination mode, granting another Localization Attempt. The wall has no cavity, and demolition would not free a trapped victim, because there is no victim in the wall. The Protagonists have to stop the broadcast, work out which apartment geometry is serving as the destination, and decide whether to make that room less compatible or empty it and invite a complete emergence.
The Panic Room
A financier survives three impossible attacks by retreating into a custom panic room. The Tunneller always stays outside, visible on the security feeds, and the consultants conclude the room is secure. The room actually has the highest Transmission Rating in the building, because its opposing walls, isolated utilities, and repeated panels make a clean metastable state, and the creature has only delayed localizing because the financier’s security team keeps altering the exterior state. When lockdown freezes the building into a fixed configuration, the “safe” room becomes the easiest destination in it. The Protagonists can save the client by wrecking a multimillion-dollar room while armed guards read them as the threat.
The Open-Space Hunt
A rural community reports livestock turning up inside sealed grain silos with planar wounds and blue vapour on the steel walls, and the predator is never photographed crossing a field, so the investigators assume it lives beneath the farm. It lives in open woodland several kilometres away. The silos are nearly identical, high-quality destination states connected by radio telemetry and repeated maintenance cycles, and every automatic fill-level pulse gives it another attempt. Turning off the network stops new events but leaves a fully localized Tunneller inside one sealed silo with two missing workers. Opening the hatch makes a conventional exit. Detuning the silo traps it with them.
Halfway to Exist
A defense contractor has built a resonant localization chamber to lure Tunnellers into a kill box, and the mechanism works, but the fatalities stay high because the target sometimes appears in a nearby housing complex instead. The complex was built by the same contractor from the same modular plans, and its utility voids offer a lower-loss intermediate state than the official chamber. The company knows this and has classified the residents as background coupling. Destroying the research saves lives but loses the only proven containment method. Correcting it means entering the service cavities while a Tunneller is actively mapping them.
Field classification
TUNNELLER Type: angular localization organism Habitat: metastable boundaries, repeated architecture, nested shells, sealed infrastructure Diet: unknown; its attacks usually remove observers, couplings, or anchors to projected space Primary sign: effects on the destination face with no route through the barrier Movement principle: collective-coordinate tunnelling between projected field configurations Containment principle: detune destinations, remove shared couplings, suppress named transitions Critical hazard: resonant intermediate states created by redundant containment
Where the science ends
Macroscopic quantum tunnelling of collective variables is established physics. Instantons and bounce solutions are standard semiclassical tools. Metastable states can decay through barrier penetration. Resonant structures can enhance tunnelling. Repeated measurement can suppress or accelerate a transition depending on how the system, the detector, and the environment are coupled. None of it lets an animal relocate through a concrete wall. Real macroscopic quantum systems need extraordinary coherence, carefully controlled variables, and specific physical couplings, and decoherence does not politely ignore a warm biological body. The Tindalosian axiom supplies a protected angular collective coordinate and an unfamiliar reservoir that real physics does not. That axiom is impossible by current theory. The consequences have been kept as honest as the premise allows.
Technical notes
- Sidney Coleman, “The Fate of the False Vacuum: Semiclassical Theory”, Physical Review D 15 (1977). The foundational bounce treatment of false-vacuum decay.
- Curtis Callan and Sidney Coleman, “The Fate of the False Vacuum II: First Quantum Corrections”, Physical Review D 16 (1977). The fluctuation prefactor accompanying the semiclassical exponent.
- Ju H. Kim and Kyungsun Moon, “Macroscopic quantum tunneling of phase-locked fluxons in a long Josephson junction”, Physical Review B 71 (2005). Collective-coordinate tunnelling in a macroscopic superconducting system.
- Ju H. Kim and Ramesh P. Dhungana, “Effects of Resonant Cavity on Macroscopic Quantum Tunneling of Fluxon in Long Josephson Junctions”, Physical Review B 83 (2011). A real analogue for environmental structure enhancing a collective tunnelling rate.
- P. M. Harrington et al., “Quantum Zeno Effects from Measurement Controlled Qubit-Bath Interactions”, Physical Review Letters 118 (2017). Experimental access to both Zeno and anti-Zeno behaviour in a superconducting qubit.
- Masahiro Hotta and Masahide Morikawa, “Does imperfect detection of emitted photons lead to the quantum Zeno effect?”, Physical Review A 69 (2004). Why merely detecting an emitted signal need not feed back strongly enough to inhibit the source transition.
Closing note from Harker
The final design review found no crack, breach, deformation, residue path, or missing material in either containment shell. Every specification had been met. Both barriers individually performed better than required. Harker had built its wall around an intermediate state where the specimen could exist.
