242: March 15

FrostDragon's name was Chen Wei.

He was twenty-four, two years younger than me, one year younger than Bai Yueran. He'd been a logistics analyst in his original timeline — the kind of work that required tracking complex supply chains across multiple delay classes and unknown variables. The coordination-optimized formation design had come from the logistics work: you don't know exactly when a component arrives, so you build the system to be ready for the range of when it might.

He'd come back from October 2016. Nine months before the Lu Yifan network had begun their CW V format campaign. Enough time to build something, not enough time to do it the way he'd have done it if he'd had the full length of the pre-challenge period that I had. He'd worked with the constraint. The formation he'd built reflected it: coordination speed over class advantage, because class advantage required the full development arc and he'd had only three years of it. He'd optimized for what was achievable in the time he had.

We met at a tea house in Beijing that Bai Yueran had chosen. The same kind of deliberate choice she made with everything: the right size (six tables, mid-afternoon, half-occupied), the right context (neither formal nor casual, close enough to the tea quality to be appropriate for a meeting that mattered), the right time (mid-afternoon, after lunch and before dinner, no competing obligations). The tea house was near Dongzhimen — not the center, not the periphery, the kind of location that had its own neighborhood logic rather than serving as a landmark. She'd found it the way she found everything: by knowing what the requirements were and searching until the conditions were met.

Chen Wei was already there when we arrived.

He was shorter than I'd imagined — the commander of Iron Frost Ascent present as a 24-year-old logistics analyst in a mid-weight winter coat, looking at the tea menu with the focused attention he brought to everything. He looked up when we came in and stood.

"Bladeless," he said. And then: "Bai Yueran."

He said it the way I'd learned to read him in the CW IV private messages: direct, without preamble, the important word first.

"Chen Wei," Bai Yueran said.

We sat.

The tea house had the particular stillness of a room that was built around quiet conversation — not silence, but the kind of sound level where individual words were clear and there was no need to raise your voice. Six tables, four of them occupied, the afternoon light coming through the window at the angle that the northern March produced: lower than the summer angle, not yet the full bright quality that April would bring, but warmer and cleaner than February's light had been. The kind of light that suggested the season had made a decision and would follow through. Chen Wei had chosen the chair with his back to the wall and the window to his left. The logistics analyst's instinct, probably — facing the room, knowing the exits, taking in all available information without rotating to find it.

***

He ordered tea first. The kind of detail that told me something: he'd been here before, or he'd researched what to order. Not a random choice.

"I've been watching your guilds for two years," he said. Not for Bai Yueran's benefit — she knew. For the record. Starting from the common knowledge before the uncommon knowledge.

"Yes," I said.

"Both guilds are building something the tournament format wasn't designed for." He turned the tea cup in his hands. "Not in the sense that you're circumventing the rules. In the sense that the tournament format assumes the primary competition is between guilds — between competing interests. What I've been watching is different. You're competing and also modeling each other. Building from what you see in each other."

"Bai Yueran and I have been co-modeling since CW III," I said.

"I know," he said. "And I've been watching both of you since the CW IV quarterfinal." He looked at the tea. "In my original timeline, this didn't happen. The top guilds were competitive units. They protected their analysis. The information didn't propagate."

"What happened in your timeline," Bai Yueran said. Not asking about the network specifically — asking about the competitive environment.

"The top-three guilds in CW V were all using variations of the same formation design. Nobody shared analysis. Everybody independently arrived at the same local maximum." He looked at the tea. "It was efficient and it was wrong. The local maximum was well below what was possible."

The same local maximum, independently arrived at, because nobody was sharing the direction the work was pointing in. The inefficiency was systematic: every guild doing the analysis work in isolation, every guild hitting the same ceiling from slightly different approaches, nobody accumulating the combined insight that would have shown what was above the ceiling. The collaborative model I'd built with Bai Yueran had emerged from the CW III matches, not from any deliberate design. In his timeline, the conditions that had produced it hadn't existed. The guilds had competed without modeling each other's development. Efficient, isolated, and wrong about what the ceiling actually was.

"The documentation layer," he said.

"Yes," I said.

"Your certification record, Wenqing's archive, the account authenticity self-attestation you submitted in January." He looked at me directly. "I've been following the probing pattern. The November eligibility API access requests."

He'd seen the same data Wenqing had.

"You knew before January," I said.

"I knew in November. I modeled the likely approach: eligibility verification rather than certification. I didn't tell you because I wanted to see if you'd find it independently."

Bai Yueran said: "And if we hadn't?"

"I would have told you in December," he said. "Before the submission window closed." He looked at the tea. "I wasn't testing — I was watching to see the model's quality."

Not testing. Watching the model. The same thing he'd been doing from the CW IV quarterfinal to the CW V semifinal.

"What did you want to share," I said.

He reached into the bag beside his chair and put a folder on the table.

"The network's campaign in my original timeline," he said. "What they did in CW V through CW VII. The approach, the targets, the mechanism. Not the specific events — the pattern. The pattern is what's stable across timelines."

A folder with the pattern.

"You came back knowing this," I said.

"Yes. It's why I built the coordination-optimized formation. The class advantage was what they targeted first — the same class ceiling manipulation you countered in 2018. What came after was more sophisticated."

"What came after," Bai Yueran said.

"They built an alternative certification infrastructure," he said. "A parallel system that looked legitimate — properly timestamped, blockchain-anchored, using the same technical architecture as the existing database. Then they recruited a mid-tier guild with a cooperative leadership and seeded false records into the alternative system, building a case that the alternative system's records predated the original."

A parallel system. Not attacking the original certification — building a competing one.

"The existing certification database," I said. "If an alternative database with equally valid-looking technical credentials makes a competing claim —"

"The tournament committee has to adjudicate between two competing certification systems," he said. "And the tournament committee has two members with Lu Yifan network connections. In my timeline, the adjudication took three months. The guild was suspended pending resolution. Three months outside the competitive format meant no CW VI seeding data. No seeding data meant no seed 1. Without seed 1, the bracket draw could put the guild against any opponent in the first round."

The adjudication as a mechanism. Not to win the certification challenge — to create a period of competitive uncertainty.

"The timeframe," I said. "When did they build the parallel system."

"November to February, in the cycle before CW VI," he said. "The current cycle."

November to February. We were in March.

"Are they building it now," Bai Yueran said.

"I don't know," he said. "I know the pattern. I don't know the current state. Your Wenqing might." He looked at the folder. "The pattern document — I've been building it since I came back. Everything I knew from my timeline, organized by mechanism rather than event. The specific events were different. The mechanisms were the same."

He pushed the folder across the table.

I looked at it.

"Why are you sharing this," I said.

"Because I came back to build something that holds," he said. "Not to build it privately and watch everything around it fail. If the documentation layer fails because I had information I didn't share — that's a failure I made." He looked at the tea. "The coordination-optimized formation was built to survive what I saw coming. But my formation is seed 6. If the Lu Yifan network targets seed 1 and takes seed 1 down, the competitive environment that seed 1 has been building — the co-modeling, the analysis sharing, the propagating documentation — that goes down too."

He was building within a system. And the system needed seed 1.

"You need us," Bai Yueran said.

"We need each other," he said. "Different starting points. Same direction."

Bai Yueran looked at the folder. Then she looked at me.

I looked at the folder.

"I'll forward the pattern document to Wenqing today," I said.

"Yes," he said. "He'll see the parallel certification mechanism in the November data if he knows to look for it."

***

The afternoon passed in the way of conversations that are doing real work — slowly, precisely, without wasted motion. Chen Wei's pattern document was 28 pages. We went through it at the tea house table, the three of us, until the mid-afternoon became early evening.

Bai Yueran said: "The third layer — the one I'm building for CW VI. If the competition environment is disrupted in the next six months, does it matter what I build."

"Yes," Chen Wei said. "The third layer matters whether or not the CW VI environment is stable. The work is what holds the system. The system is what the work happens in. You build the work; the documentation layer protects the system. Both are necessary."

She looked at the table.

"Both are necessary," she said. As if she'd known this and needed to hear it confirmed.

I watched her sit with it for a moment. The competitive environment and the work that happened inside it — she'd been treating them as separate concerns, the way it was natural to treat them. The competition was the context; the formation was the content. But Chen Wei had collapsed that distinction with a single sentence. The work was what held the system. The system was what the work happened in. If you only protected the system and stopped building the work, the system became an empty container. If you only built the work and left the system unprotected, the container didn't survive long enough to hold what you were building. Both or neither. She'd known this, in the way that people know things before they've been stated plainly. Hearing it stated plainly was a different thing from knowing it.

"Yes."

Before we left:

"QingxueTide," I said. "She's thinking about the ceiling of anticipatory healing."

He looked at me.

"What do you think she's found," I said.

He was quiet for a moment. "I think she's found that the ceiling she thought existed doesn't," he said. "And she's deciding what to do with the space above it."

*That's what FrostDragon meant,* he'd said. *I don't know if she means what's above the ceiling or the fact that the ceiling is higher than she thought.* And I'd said: *Both.*

"Yes," I said.

"She'll keep working," he said. "She's the kind of person who works toward the horizon." He picked up the bag. "The frost that watches. I said it as a description. It's also a description of what QingxueTide does — she watches the horizon for what the healing output can become."

He put on his coat.

"March 15," he said. "Good meeting."

"Good meeting," Bai Yueran said.

I said: "The work continues."

He looked at me.

"Yes," he said. "It continues."

He walked out of the tea house into the March evening.

I sent Wanqing from the tea house: *Met Chen Wei. He's the right person. The pattern document is real. I'll explain at the bench.*

She replied: *I'll be at the bench Thursday.*

The March evening in Beijing. The northern March was different from Hangzhou's March: still cold in the air, the spring not yet arrived in the trees the way it would be in Hangzhou in another two weeks.

I thought about the 28 pages and the afternoon we'd spent on them. Three people with different timelines and the same direction, working through a pattern document at a tea house table until it was no longer afternoon. There had been nothing dramatic about it. That was the thing I kept returning to as I found the subway entrance and went down into the familiar underground warmth: the most significant information exchange I'd had in months had happened at a table over tea with declining afternoon light and a half-empty tea house around it. The pattern that could undo five years of documentation work had been described in clean language, organized by mechanism, handed over without ceremony. That was how real things tended to arrive. Not announced. The tea house had been warm enough that stepping outside felt like the temperature was an active presence. I stood on the street for a moment before finding the subway, turning the 28-page folder in my hands. The pattern document. The mechanism described clearly, organized by how it worked rather than when it happened. Wenqing would have it by tomorrow morning.

The bench on Thursday.