The training hall opened at fifth bell.

  Kael was there at fourth bell and fifty seconds, standing outside the locked door with his notebook under his arm.

  The campus was still dark. Dew on the stone steps. The rule field running its overnight maintenance cycle—inscriptions dim along the corridor walls, pressure low, the whole system breathing at its nighttime minimum.

  He preferred this hour.

  Less foot traffic meant fewer variables. Fewer variables meant cleaner data.

  He'd spent three years practicing in borrowed spaces and storage rooms that hadn't been built for what he was trying to do. Always the first to arrive. Always the last to leave. Always working in the margins of places designed for someone else—shaping his schedule around the gaps in other people's routines, running tests with whatever ambient field conditions the space happened to carry, documenting results in a notebook that had no context outside his own head.

  What he'd accumulated over those three years was evidence that something was different about how his system worked. Not proof. Not even close to proof. The storage rooms had no instruments. The borrowed halls had instruments he wasn't authorized to use.

  The difference here was that this hall had platform inscriptions sensitive enough to register what happened at maximum load.

  That was what he needed.

  ?

  The hall manager arrived at fifth bell exactly.

  Short, solidly built. Somewhere in his sixties. The kind of hands that came from physical work rather than administrative duty—calloused, competent, the hands of someone who had maintained equipment for years and had learned every joint and groove of the machines he worked with. He had unlocked this door every morning for years and had long since stopped needing to think about it.

  He looked at Kael once—gray robe, Block Four assignment, standing alone before dawn—and didn't ask questions.

  "Sign in," he said, and held the door open.

  Kael signed in.

  He was the first name on the page.

  ?

  The hall had twelve practice platforms arranged in two rows down the center. At this hour, empty.

  Each platform sat in a shallow depression in the stone floor, enough to contain field discharge without bleeding into adjacent stations. The inscriptions around each one were recent work, the channeling grooves still sharp—a maintenance upgrade within the past few years, he estimated, based on the quality of the stonework and the consistency of the field response visible even at low ambient saturation. Better equipment than the testing halls.

  He didn't take the nearest platform.

  He walked the room first—a deliberate circuit at twenty percent load, reading the ambient field response at each station. The variations were small but consistent. Platform three, at the far end, where two channeling lines crossed beneath the floor: slightly higher saturation, field response more uniform across the surface.

  The kind of detail that only mattered if what you were measuring was very precise, or if what you were doing was very far from normal.

  He had something different to do.

  He stepped onto platform three, set his notebook on the bench at the side, and began.

  ?

  The standard framework said this:

  Output equals Capacity times Stability, divided by one hundred.

  Capacity was how much Rule energy a practitioner could hold. Stability was the transmission coefficient—the percentage that successfully converted to output. A practitioner with Capacity 80 and Stability 75% produced 60 units of effective output. The unconverted 25% dissipated.

  It became heat. Interference. The faint instability around someone working near their limits.

  The energy left the system.

  This was so foundational to the framework that the textbooks didn't state it as an assumption.

  They stated it as physics.

  Kael had never felt the energy leave.

  He'd first noticed it at twelve, practicing alone in a storage room because his family couldn't afford hall access. He'd pushed past what his Capacity was supposed to hold and felt, instead of the expected collapse, something arrive—a stabilizing presence adjacent to his Capacity, distinct from it, that absorbed the overflow without being asked and held it without apparent effort.

  Like discovering a room in a building you'd lived in for years. Behind a door you'd never noticed.

  He hadn't told anyone. At twelve, he hadn't had the language for it. By thirteen, he'd had the language but understood, with the clear-eyed pragmatism that came from growing up in the city's outer district, that announcing a deviation from the framework in institutional settings was not how you got access to better resources. It was how you got your results questioned and your file flagged.

  He'd spent three years quietly mapping its edges instead.

  Today he was going to find out if it had a ceiling.

  ?

  He started at sixty percent.

  Standard output. The platform's inscriptions held a steady low blue—the baseline emission of a practitioner working well within their limits, unremarkable. He held for thirty seconds, observing the field distribution around the platform, confirming the consistency he already expected.

  Released.

  Seventy percent.

  Pressure behind the sternum now—the particular sensation of the Rule pressing against the inside of his Capacity, like water rising in a vessel. Still clean. The inscriptions held their color without variation.

  He released and made a note.

  Not about what he felt.

  About what he didn't feel: any of the friction the textbooks described as normal at this load. Most practitioners reported warmth at seventy percent, a slight heat that built at the skin's surface as unconverted energy found its way out. He felt none of it.

  The author's tale has been misappropriated; report any instances of this story on Amazon.

  He never had.

  He had simply never had the equipment to confirm it wasn't normal until now.

  Eighty percent.

  Held for sixty seconds. Careful inventory of his system: Capacity walls engaged and steady. Affinity supply consistent. Field interaction with the platform clean. No leakage at the edges. The inscriptions burned a slightly brighter blue, appropriate to the load.

  Released. Two lines written. Reloaded.

  Eighty-five.

  This was where Initiate-level practitioners began showing instability—the flicker at the edges of output when someone pushed toward their limit, the slight irregularity that crept in when the vessel wall started flexing and the Rule found micro-fractures in the practitioner's control.

  Not failure. Just the normal physics of a system approaching its operating boundary.

  He loaded to eighty-five and held.

  The platform's inscriptions didn't flicker.

  Ninety seconds. Three times his usual test duration at this load, long enough that any latent instability would have manifested.

  Nothing.

  The inscriptions maintained their color exactly.

  He paused, wrote a single line, and loaded again.

  Ninety.

  Ninety-three.

  Ninety-six.

  The pressure had changed quality somewhere around ninety-two. Not pain. Closer to the sensation of standing at the edge of something high—a sharpening of attention, the body understanding that the margin for error had become very small. His Capacity was at the limit of what his Affinity could cleanly supply. The vessel wall was flexing hard.

  And then, at the point where every framework he'd ever read said he should be hemorrhaging energy in all directions—

  The second floor opened.

  It arrived without drama.

  A stabilizing presence, adjacent to his Capacity but not contained within it, that took the overflow and held it. Not dampening. Not redirecting.

  Simply holding.

  The way a second foundation catches what the first can't support—so quietly the building above never feels the weight transfer.

  He loaded to ninety-eight.

  To one hundred.

  The pressure at one hundred percent was total. Every wall his Capacity had was engaged. The Rule filled every available space and the second presence held the rest, and for one long moment the entire system was at maximum load.

  Nothing was breaking.

  The platform inscriptions burned blue-white.

  He held it.

  Ten seconds.

  Twenty.

  Forty.

  Sixty.

  He released.

  ?

  The hall was silent.

  He became aware of them as he stepped off the platform.

  Three students standing at the entrance to the platform area. They had come in for early training and stopped moving at some point before ninety percent and hadn't moved since.

  He didn't know how long they'd been there.

  The one on the left had a notebook gripped in both hands, knuckles white. The one on the right had gone still in the particular way people went still when their processing had been interrupted—not watching, not reacting, just waiting for the world to catch up to what they'd just seen.

  The one in the middle had let his notebook fall to the floor.

  He hadn't picked it up.

  All three were looking at platform three. At the inscriptions, still fractionally brighter than they should have been, the blue-white fading back toward normal in slow pulses. The stone itself still processing what it had been asked to hold.

  The silence in the room had weight.

  Not the absence of sound—the presence of something that had just stopped.

  None of them looked at Kael.

  None of them said anything.

  Kael walked to the bench and opened his notebook.

  He wrote for four minutes. Not impressions—data. Load percentages, transition points, the behavior of the second presence at each threshold. The precise moment it had engaged: ninety-six, not ninety-five, which was one percent higher than his last test. The platform inscriptions had gone white rather than bright blue, which meant the field interaction at maximum load had crossed some threshold the standard color scale wasn't designed to represent.

  He noted that three students had witnessed the entire sequence and had not spoken.

  He underlined that.

  Witnesses changed things. The training hall was not a controlled environment—it had records, a manager with eleven years of baseline data, students who would tell other students. He had wanted to find the ceiling. He had found something else: that what he was doing was visible, and visibility had consequences he hadn't fully accounted for.

  He was going to need the intermediate catalogue faster than he'd planned.

  ?

  "That was one hundred percent."

  The hall manager was standing at the edge of the platform area. Arms folded. His expression had stopped being neutral somewhere around ninety-three.

  "Sixty seconds sustained," he said. "No discharge."

  Kael kept writing. "Is there a rule against it?"

  "There's no rule because it doesn't happen."

  The man's voice was flat and precise—the voice of someone choosing words carefully because the situation required it.

  "The inscriptions went white. Platform three hasn't gone white since the testing standards upgraded four years ago." A pause. "I've managed this hall for eleven years. I've seen students approach their theoretical ceiling maybe forty times. They all discharged. The platform always showed it—emissions going ragged, the inscriptions cycling instead of holding. That's what the ceiling looks like."

  He stopped.

  Kael looked up from his notebook.

  "What you just did doesn't look like anything I've seen."

  The three students at the entrance hadn't moved. The one on the right had finally taken a breath. The one in the middle still hadn't picked up his notebook.

  "Your Stability," the manager said. Not a question.

  "The instrument read one hundred percent."

  "Yes."

  "I wrote it off." He said it without apology—matter-of-fact, the way you stated a fact that had turned out to be wrong and the only useful thing was to acknowledge it. "Everyone did. Instrument error. Three readings, same result, still easier to call it a glitch than—"

  He stopped again.

  Looked at the platform for a long moment.

  "I was wrong."

  The words came out quieter than the ones before them, directed less at Kael than at some internal accounting he was settling. Eleven years of baseline and comparison. Whatever column he'd put instrument errors in was going to need to be reclassified.

  He unfolded his arms. Turned toward his desk. Sat down.

  He pulled a piece of paper from the lower drawer. Uncapped his pen.

  Kael watched him begin to write. The handwriting was careful—deliberate, the way you wrote something that needed to be precise.

  Then he returned to his notebook.

  He had two more hours before class, and more to learn.

  ?

  When he left the hall at seventh bell, the letter was finished.

  He could see it on the manager's desk—folded, sealed, addressed in that same careful hand.

  He didn't ask who it was going to.

  He didn't need to.

  Someone had just watched him do something the framework said was impossible. Someone with eleven years of baseline data and the instinct to document it. The letter was going to a person with the authority to act on it.

  That meant attention was coming. The question was what kind, and from which direction, and whether the person at the other end of that letter was going to see what he was as an opportunity or a problem.

  He stepped out into the morning light and started walking toward the library.

  He had a formula to interrogate, two specific chapters to find, and a professor to read correctly on the first try—with no margin for misreading her.

  He was already running the numbers on how much time he had before the letter arrived.

  The Academy's standard processing time for committee correspondence was three days.

  That was the baseline.

  The real question was whether whoever received it would treat it as standard committee correspondence—or as something that required a faster response.

  He walked faster.

  He intended to be ready before he found out which one it was.

  ?

  The library opened at sixth bell.

  He had forty minutes before class.

  He used them on the three introductory Stability texts, working quickly—not because the material was easy but because he already knew most of it and was reading now not for information but for the shape of what was missing. Every model treated Stability as a regulator. Every explanation of energy loss described the same physics: dissipation, heat, the friction of imperfect transmission.

  None of them asked what would happen if the friction were zero.

  Not as a theoretical question—as a real one.

  If a practitioner's Stability was genuinely one hundred percent and stayed there under load, the formula predicted a fixed output of 38 units. Clean, stable, unremarkable. The math was satisfied.

  What the math didn't account for was where the pressure went.

  When you loaded a vessel to capacity and nothing leaked out, the pressure inside had to go somewhere. Either the vessel expanded, or something else absorbed it, or the system found an equilibrium the model hadn't predicted.

  Kael had been assuming the third option for three years because it fit his experience.

  This morning he had run the test at maximum load in a controlled environment with field-sensitive inscriptions and witnesses.

  The inscriptions had gone white.

  That was not equilibrium.

  That was the system doing something the model said it couldn't.

  He returned the last introductory text and requested intermediate access.

  The librarian told him it required instructor approval.

  He thanked her, wrote down the form number, and went to class.

  He had already decided what he was going to say to Aldea, and how, and in what order.

  The offer would be simple: the observation from the platform, the full version, the part he hadn't said out loud to anyone. A theoretical gap that pointed toward something real, documented in a controlled environment with an eleven-year baseline as backdrop.

  People who studied things gave access to other people who studied the same things.

  Her pause in class had told him what kind of person she was.

  He just needed to confirm it.