154: February

The spring semester started January 20 and ran through the usual schedule.

Floors 12, 13, and 14 cleared in January and February — each one a variation on the pattern that the Black Castle's design seemed to favor: a new mechanic that could be decoded from observable behavior if you watched long enough, a formation solution that required something the guild had been building over the previous months, and a kill notification that updated the server record and generated the expected forum traffic. The mechanics themselves were each distinct enough to require fresh analysis but each one's solution drew from skills the formation had been building since March. Floor 13's mirror-aggro system required precisely timed simultaneous positioning from two sub-formations; Floor 14's crystallization debuff required the kind of sustained coordination under accumulating pressure that the Floor 8 mode-shift work had built. TwilightTide's briefings noted the connection without dwelling on it — she included a single sentence in each analysis that traced the current requirement back to the floor that had established it. It was the kind of thing you only wrote if you'd been watching the pattern build.

The forum traffic had shifted since September — the threads about Severing Light's formation mechanics were still there, substantial and ongoing, but the volume on TwilightTide threads had expanded significantly since the controlled announcement. The gaming press, the entertainment press, and the intersection of both had all written something in the previous four months.

The intersection coverage was the most interesting to read analytically — pieces trying to hold both frames simultaneously, the artist and the analyst, the public figure and the guild member. Some of them got it right. Some of them got the formation analysis approximately right but missed the context. A few of them got both right, and those were the ones that Wenqing filed in the public relations folder under a separate category labeled: *Accurate coverage (rare).*

TwilightTide's in-game performance was public record. The kill notifications, the formation analysis threads on the server forum, the Phase breakdowns that Wenqing had started publishing in the guild's public-facing records after the Floor 10 clear. Anyone who wanted to understand what she'd done in the Black Castle could read the public record. The coverage didn't always understand what they were reading. It didn't have to. The people who understood were in the guild channel.

***

Wanqing was in Hangzhou the first weekend of February.

She'd come up for a seminar — one of the things her advisor ran on the Hangzhou-Suzhou rotation — and came to the east courtyard bench afterward with the February version of her problem set. The maple was at bare branches again. A full cycle from last February's first-leaf, completed. I'd been at HZUT for twenty months now.

She had the thermos. The same dark green sweater, or possibly a different dark green sweater that occupied the same role in the seasonal wardrobe. The February campus had its characteristic stripped-down quality — the tree branches bare, the foot traffic purposeful rather than wandering, the cold efficient at discouraging outdoor lingering. The maple was bare in the way it was bare every February. I'd sat at this bench through a full year now. The maple had given me its full cycle.

"The Pioneer's Path," she said.

"Cycle 91 as of last week." The three AM sessions had continued through the fall and winter. The rate had slowed in December — the Suzhou break had interrupted the schedule — but had resumed at the previous frequency in January. "Eighty cycles to the next Heritage node."

"Eighty more."

"Wenqing calculated: at five sessions per week average, the 80 cycles complete in mid-spring." I looked at the bare maple. "April, possibly May."

"And what comes after cycle 160."

"Beigong Yan says the Heritage Fragments are at cycles 50, 80, and 160. The fourth fragment — the final one before the class transition — is at cycle 160."

"Class transition."

"The end of the Pioneer's Path chain. Beigong Yan has been guiding me through the Sword Sovereign's history. The fragments are the memories of what he built and what he chose. The end of the chain is the class itself — whatever he left behind." I looked at the maple. "The Heaven-Severing Blade Sovereign class is somewhere in the Floating Cloud Sect's hidden vault, past cycle 160. That's what the quest chain is building toward."

She looked at the bare maple. "You've been on this path for eleven months."

"Since March last year."

"March to April this year is thirteen months."

"Yes."

She turned a page of the problem set without reading it. "You're not in a hurry."

"No. The path is the right length. The cycles build something real — I can feel the difference between cycle 30 and cycle 91 in the combat sessions. The precision is higher. The game's combat AI adapts to me faster than it used to because my response patterns are less predictable."

"What's different."

"The combat reads differently. I see more options per second than I did at cycle 30. It's like — at cycle 30, I was seeing the correct options. At cycle 91, I'm seeing the correct options and the two options on either side of the correct ones, which is different information." I paused. "It's subtle. Wenqing tracks it in the session data."

"He tracks your improvement."

"He tracks everything. I'm in the dataset."

She smiled — the corner version, the one that appeared when something amused her without her deciding to be amused. "You'd hate that if anyone else did it."

"Wenqing does it correctly."

"I know." She went back to the problem set. The February campus around us — the bare maple, the east courtyard's cold stone, the low February sun at the angle that put the bench in partial shadow. "The transplant coordination team."

March 24 was the eight-month near-end of the window. Twenty-four days.

"February 28 today," I said. "Twenty-four days."

"Do you think it comes in the window."

"I don't know. In the old timeline it came at month fourteen." I looked at the maple. "The documentation is cleaner in this timeline. The fund was complete earlier. Doctor Yan's bloodwork results in November were the best profile they'd had since the diagnosis." I paused. "But the match timeline depends on donor availability. Not documentation. Not fund status."

"The match timeline depends on the donor," she said. "Not the documentation."

"Yes."

"So you don't know."

"No."

She was quiet. The bare maple's branches had the February quality of being exactly what they were — no leaf cover, the full architecture visible, the structure of the thing rather than its seasonal presentation.

I'd looked at the maple enough times now to know it in all its registers. Full summer leaf — the dense canopy, the particular way the light broke through it in patches, the sound it made in a light wind different from the sound in a heavy one. The October thinning, the leaves going color and then falling in the sequence that started at the outermost branches and worked inward, so that the last leaves to go were the ones closest to the trunk. The February version, which was this: bare architecture, the full branch structure clear, nothing concealed. The maple didn't hide what it was in winter. It just showed the frame. I'd been coming to this bench through enough seasons now that I'd stopped thinking of the maple as the same tree in different states and had started thinking of it as one thing that contained all the states simultaneously — the leaf and the bare branch and the first bud all present as potential at any given moment, the current state just the one the season had selected.

"I'm going to be in Suzhou," she said. "When it comes. When the call comes."

"I know."

"I want to be there."

"I know," I said. "I know you do."

She went back to the problem set. I looked at the maple.

Twenty-four days to the near end of the window. The donor was somewhere in China — Zhejiang Province, maybe, or somewhere else, a person whose existence was entirely independent of everything I'd been building for fifteen months. The fund and the documentation and the formation precision and the Pioneer's Path cycles had no bearing on when that person's biology became available.

Some things you controlled. Some things you waited for.

I was good at waiting. I'd spent five years of a previous lifetime learning to wait for the right moment, the correct information, the thing that couldn't be forced. The waiting was different now — more informed, more situated in a specific outcome — but the discipline was the same.

***

TwilightTide sent a guild-channel message in mid-February that wasn't formation analysis.

It was a link to an interview she'd given to a gaming publication — one of the ones that had been covering the Black Castle progress and the CW I result with the kind of ongoing attention that indicated they'd decided Severing Light was a story worth following. The interview was about the game, primarily: the mechanics, the formation analysis work, the three AM sessions. The journalist had been careful — the questions stayed in the domain she'd announced as open in the September announcement, rather than reaching toward the things she'd announced as closed.

The Wenqing-approved coverage category.

At the end of the interview the journalist had asked: *Was there a moment when you decided this was worth it? The secrecy, the schedule, the six months of protected identity?*

She'd answered: *There were a lot of moments. The one I think about most was when I was at three AM in the Iron Hills, running solo protocols because I couldn't sleep, and someone logged in and started running parallel cycles without saying anything. They didn't ask why I was there. They just ran their cycles. After a few weeks there were three of us, all at three AM for our own reasons, running adjacent sessions in the same zone. No one asked why anyone else was there. That felt like something worth staying for.*

I read the answer and then sat with it for a moment. The Iron Hills at three AM, the three of us in adjacent positions — TwilightTide, Zhu Yuhan, and I, each running our own thing in the same zone without requiring the others to account for their presence. It had become structural, that arrangement, so structural that I'd stopped noticing it as a thing that had been chosen and had started experiencing it as a thing that simply was.

What struck me most about the answer was what she'd chosen not to say. The journalist had asked about worth — about the decision. She hadn't answered with the CW I result or the floor records or the server rankings. She'd answered with the three AM image. Which was either the truest account or the most strategic one, or both simultaneously, which in her case was often the arrangement.

She'd written about it in an interview. She'd answered a journalist's question about worth with that specific image.

I sent her a brief message through the private channel: *Good interview.*

She replied: *Thank you. They were careful with the questions.*

*The journalist was.*

*I notice careful people,* she said. *Professional habit.*

She went back to the Floor 14 analysis. I closed the channel and looked at the February campus outside the lab window — the specific campus February, cold and coursework-heavy, the semester at its most demanding stretch before the spring break.

Three of them in the Iron Hills at three AM, each for their own reasons, each not asking why the others were there.

That felt like something worth staying for.

I understood exactly what she meant.

The interview had been published on a gaming site with a readership in the tens of thousands — small relative to her announcement's eleven million, but the right tens of thousands. The people who'd been following the Black Castle progress and the CW I result and the floor analysis threads on the server forum. The people for whom "three AM in the Iron Hills" was a specific recognizable geography rather than a general description of dedication.

She'd answered the journalist's question about worth with that specific image. Not the CW I result, not the Floor 10 clear, not the formation analysis or the server records. The image of three people running parallel sessions in the same zone without asking about each other's presence.

She said the last thing — the three AM image, the three people running parallel sessions for their own reasons, no one asking why — and the interview ended and I sat with it for a moment longer than I'd intended. It was a precise description of the thing it was a description of, which was not a common quality in answers given under interview conditions. She'd have known that. She would have known it was precise when she said it, and said it anyway, which was the kind of accuracy that required a certain confidence in the audience — the belief that at least some of the people reading would understand what she'd done with the answer. The ten of them in the guild channel who'd been in the Iron Hills at three AM. A few journalists who read carefully enough. That might have been sufficient for her.

I stayed on the bench until the February cold made it impractical, then went back to the computer lab and worked on the semester problem set until nine PM. The problem set was algorithms — the kind of problem that required sustained precision rather than inspiration, the kind that rewarded exactly the thing the Pioneer's Path cycles were building.

Twenty-one days to March 24.

I set the three AM alarm and closed the laptop.