Han Weiwei's instruments caught it before he'd been in the building three minutes.
He was crossing the fourteenth floor toward the research lab when the automated alert sounded from inside — a tone he hadn't heard before, higher-pitched than the standard calibration alert, with a repeating cadence that meant the system had flagged something for immediate review.
He entered the lab. She was already at the primary console, the spatial awareness running at full resolution, her hands moving across the instrument interface with the speed of someone who had spent four years learning what every alert tone meant.
"Stop," she said without looking up. "Stand where you are."
He stopped. Two meters inside the door.
"The spectral array just registered a new emission band from your position," she said. "480 to 520 nanometer range. It wasn't there last session."
She pulled up the spectral data on the primary monitor. A graph — wavelength on the horizontal axis, intensity on the vertical. The graph showed his field signature as the instruments read it: the standard probability field output she'd been measuring for seven weeks, the familiar curve in the probability manipulation range.
And a new peak. A clean, sharp emission line at 497 nanometers that hadn't existed in any previous session's data.
"When did this start?" she said.
"Yesterday," he said. "The field crossed a density threshold."
She looked at him. Then back at the data. The researcher's discipline overriding the personal — she wanted to ask what threshold, but the data was in front of her and the data came first.
"Don't move," she said. "I'm running the full characterization suite."
The full characterization suite took forty minutes. He stood in the same spot for all of it, the instruments scanning the probability field from seven different angles at four different spectral resolutions. She didn't speak during the scan except to ask him twice to adjust his qi circulation — once to raise it to full output, once to drop it to resting baseline.
The difference between full output and resting baseline was significant. At full output, the 497-nanometer emission line spiked to three times its resting intensity. The secondary peaks in the 480-520 range widened. The field's spectral profile at full output was dense enough to measure from across the room with standard-grade instruments.
She compiled the data into a characterization report while he watched.
"Probability manipulation," she said. "Direct type. The spectral profile is consistent with Xu Mingzhi's theoretical predictions for a probability-warping field operating at sustained elevated output." She looked at the data. "This is the first time this ability type has been instrumentally verified. The theoretical models predicted it was possible. No one has ever documented it."
She turned to face him. The spatial awareness at full resolution, tracking every variable in the room, tracking him at the priority level she'd been running since the protocol signing.
"The 497-nanometer line," she said. "It's a specific emission wavelength. It corresponds to the probability field's effect on atmospheric photons — the field literally bends the probability distribution of photon interactions in the surrounding atmosphere, which produces the spectral line. At the density you're currently running, the effect is measurable. At lower densities, it was below the detection threshold."
"Below the threshold until yesterday," he said.
"What changed yesterday?"
He considered how much to say. The LP threshold was system-specific information. The 500,000 mark and its secondary radiation effect were mechanics of the Luck System. But the observable result — the field's increased density — was a physical fact she could measure regardless of the cause.
"Cumulative development," he said. "The field's density has been increasing gradually. Yesterday it crossed the detection threshold."
"Gradually," she repeated. "The emission line appeared clean. No gradual onset in the data — the instruments would have caught a sub-threshold signal building over multiple sessions if the increase was truly gradual. This looks like a discrete transition. A step function, not a ramp."
She was right. The 500,000 LP threshold was a step function. The secondary radiation activated at a specific point, not as a gradual emergence. Her instruments were accurate enough to tell the difference.
"The development has discrete stages," he said. "The field strengthens in increments rather than continuously."
She stared at him for three seconds. Then she pulled up a new window on the console — the theoretical model she'd been building with Xu Mingzhi's framework.
"Discrete cultivation stages," she said. "That's consistent with the applied-mind cultivation model. Classical cultivation theory describes advancement as stage-based, not continuous. Each stage represents a qualitative shift in the practitioner's field output." She pointed at a section of the model. "Xu Mingzhi's convergence framework predicts that the compensatory protocol would replicate classical stage-based cultivation progression using the system's own metric structure."
She was describing LP thresholds without knowing the term. The theoretical model predicted discrete stages. The Luck Aura's level system was the implementation.
"The formal classification," she said. "This data changes the characterization."
"From unclassified to what?"
"Probability Manipulation — Direct Type, Stage-Based Progression, Novel Classification." She pulled up the Bureau's classification database interface. "The Bureau's ability classification system has a category structure. Your field doesn't fit any existing category. It requires a new one."
"A new classification category," he said.
"The first new ability type classification since the Bureau's database was established four years ago," she said. "Every other ability type documented since the Awakening Event has fit within existing categories — physical enhancement, psychic, spatial, energy, elemental. Probability manipulation was a theoretical category with no documented instances."
She looked at the classification interface. The empty field where the new category would be entered.
"If I file the formal classification," she said, "the data goes into the Bureau's database. The classification is accessible to anyone with research clearance. Your field type changes from 'unclassified' to a named, characterized ability type with spectral documentation."
"Who has research clearance?" he said.
"Bureau research staff. Affiliated researchers — which now includes you and the joint charter participants. And anyone granted clearance through formal request to the division head."
"Which is you."
"Which is me," she said. "I control access requests. But the classification's existence in the database is visible to anyone with basic Bureau database access — they can see that a new category was created, and for whom, even if they can't access the spectral data itself."
He understood the problem. The classification's existence would be visible. Anyone with Bureau database access — which included every registered awakened practitioner in the city — could look up Chen Haoran and see: Probability Manipulation, Direct Type, Novel Classification.
Lin Zhengyue would see it. Wei Changming would see it. Fang Junhao would see it.
Lin Meiyao would see it.
"The anonymity cost," he said.
"Yes," she said. "You lose the 'unclassified' notation. Unclassified is invisible — nobody investigates a field type that doesn't have a name. A named classification draws attention. Especially a novel one."
She paused. Looked at the spectral data, then at the classification interface, then at him.
"The research value is significant," she said. "The formal classification establishes the probability manipulation category in the Bureau's database. That category is evidence — the first instrumentally documented ability type that matches the theoretical predictions for a deliberately engineered capability set. The Awakening Event paper needs this data point."
"The paper is eighteen months to two years from publication," he said.
"The classification enters the database now," she said. "The paper uses it when published. But the database entry is immediate."
```
[SIDE TASK: HAN WEIWEI'S CLASSIFICATION DECISION — SHE WANTS TO FILE THE FORMAL CLASSIFICATION BECAUSE THE RESEARCH REQUIRES IT. SHE'S ALSO HESITATING BECAUSE SHE UNDERSTANDS THE OPERATIONAL COST TO YOU. THE HESITATION IS GENUINE — SHE'S WEIGHING HER RESEARCH AGENDA AGAINST YOUR STRATEGIC POSITION. IF YOU MAKE THE DECISION FOR HER — EITHER DIRECTION — SHE DOESN'T HAVE TO HOLD THE TENSION BETWEEN THE TWO. REWARD: 3,800 LP + [HAN WEIWEI: TRUST DEEPENED — DECISION SHARED]]
```
He read the side task. The system generating a task that was about giving her the decision rather than making it himself. Or making it for her. The reward was the same either way — the LP came from resolving the tension, not from the specific choice.
He looked at the spectral data on the monitor. The 497-nanometer emission line. The clean, sharp peak that proved the existence of a deliberately engineered probability manipulation ability type.
Evidence. For the paper that would explain the Awakening Event's design to the world. For the research that the architect had identified as part of Phase 3's operational requirements.
And the cost: his field type, named and documented, visible to anyone who looked.
"File it," he said.
She looked at him.
"The classification," he said. "File the formal classification. Probability Manipulation, Direct Type, Novel Classification."
"The anonymity—"
"Was always temporary," he said. "The field density will keep increasing. The spectral signature will become detectable by portable instruments at the next threshold. The classification happens now or it happens later, but it happens."
She was quiet for a moment. The spatial awareness tracking, the researcher processing, the person behind both of them registering that he'd just given her what she needed for the research without her having to ask for it as a cost.
"The classification filing process," she said. "I'll draft it today. It goes through the Bureau's classification review board — three members, all in my division. Review takes five business days."
"Five business days puts the classification in the database next week."
"Yes."
Next week. After the Lin Zhengyue preliminary consultation on Thursday. The classification wouldn't be visible before the meeting.
He'd considered the timing before answering. The "file it" wasn't impulsive. The five-day review meant the anonymity held through Thursday, and the classification entered the database after the preliminary consultation but before the quarterly review. By the time Lin Zhengyue saw the classification, she'd have already met with him and formed an impression based on the consultation itself rather than on his classified ability type.
The timing worked. Not because the probability field had arranged it — because the Bureau's standard review timeline happened to align with his operational schedule.
Or because the probability field had nudged the Bureau's standard review timeline years ago, when the review board's procedural guidelines were drafted. Level 4 was ambient. It didn't distinguish between large outcomes and small ones.
`[SIDE TASK: COMPLETE. +3,800 LP. HAN WEIWEI: TRUST DEEPENED — DECISION SHARED]`
"The spectral data," she said. "I want to run a longitudinal comparison. Weekly measurements of the 497-nanometer line's intensity. The stage-based progression should produce predictable intensity increases at each stage transition."
"You want to track the development curve."
"I want to document the first probability manipulation field's progression in real time," she said. "This has never been done. The data set will be unique in the global research literature."
"The weekly sessions we're already running," he said.
"Yes," she said. "With an additional fifteen minutes for the spectral characterization suite. The data collection integrates into the existing research session without disrupting the collaboration schedule."
She was already structuring the methodology. The researcher taking over from the person who'd been weighing costs and benefits thirty seconds ago. The transition was the fastest context switch she did — from the human assessment to the scientific framework.
He watched her draft the characterization protocol on the console. The precision applied to each measurement parameter, the spectral resolution settings, the calibration requirements. Four years of research experience compressed into fifteen minutes of protocol design.
"The Xu Mingzhi collaboration," she said without looking up. "His theoretical model predicts the stage-based progression. My instruments can now verify the prediction in real time. The joint research charter — when it's filed — has this data set as its first collaborative deliverable."
"The convergence of theory and observation," he said.
"Yes." She finished the protocol and turned. "The convergence is what makes the paper publishable. Theory alone is speculation. Observation alone is data. Together, they're evidence."
She looked at the spectral data one more time. The 497-nanometer line, sharp and clean, the first instrumentally documented proof that probability manipulation existed as a discrete ability type.
"You're the evidence," she said. Not with the weight of a declaration. With the clinical observation of a researcher noting a fact about her data set. "The field, the spectral signature, the stage-based progression. You're the first documented case."
"I know," he said.
"The paper will cite you," she said. "As the primary observational subject. Anonymized in the publication, but the Bureau's classification database will link the data to your name."
"I know," he said again.
She saved the characterization protocol. Closed the console. The instruments settling into their background monitoring cycle, the spectral array now tuned to track the 497-nanometer line continuously whenever he was in the building.
"The classification review board meets Monday," she said. "I'll submit the filing this afternoon. Five business days puts the database entry at the following Monday."
After Thursday. After the preliminary consultation with Lin Zhengyue.
He thought about this woman who had spent four years building a case that the Awakening Event was designed, who had just received the first physical proof, who was processing the scientific excitement and the strategic implications simultaneously and choosing to let the scientific excitement lead because that was who she was.
He picked up the research session's data folder from the console. The morning's first measurements, the calibration data, the spectral characterization results.
"The research session," he said. "We still have two hours scheduled."
She looked at the clock. Then at the instruments. Then at him, with the expression of someone who had been briefly distracted by a discovery and was now returning to the work that the discovery was part of.
"Yes," she said. "The cultivation school cross-reference. Where we left off."
They worked for two hours. The spectral array tracked the 497-nanometer line in the background, measuring something that had never been measured before, while they continued the research that would eventually explain why it existed.
The classification filing sat in her drafts folder, waiting for the afternoon submission.
In five business days, Chen Haoran's ability type would have a name.
