200: Not In-Game

April 1, 2018 was the one-year anniversary of the music publication feature.

TwilightTide noted it at nine AM in the private channel: *One year.*

*Yes,* I sent.

*The article is still being cited. I saw two references last week — both in contexts I didn't expect. One in a mental health discussion context, one in something about community design. They were quoting the sentence about whether the structure holds the people who depend on it.* She paused. *I didn't know that sentence would travel.*

*It travels because it's accurate.*

*Yes.* Another pause. *I'm going to be in the Iron Hills at three AM tonight. Same as last year. Zhu Yuhan confirmed yesterday.*

*I'll be there.*

I set the phone down and looked at the thesis document on the screen. Chapter 4 of nine. The chapter was on the [x] modifier accumulation period — the eighteen months of session data that preceded the class activation. The evidence was precise: session logs, modifier readings at regular intervals, Wenqing's seven sequential models building on each other. It translated well into academic notation. The formal language gave it a different quality than the raw data, but the underlying argument was the same argument Wenqing had been building since October 2015. The formal language was a translation, not an improvement. The original had been correct. The translation was for a reader who needed the citation apparatus.

The thesis was on schedule for the June 22 defense. Four chapters written, five remaining. Professor Liang's review cycle was reliable — he returned chapters within ten days and his comments were specific and consistent. The process had found its rhythm in February and was running on that rhythm now. I recognized the rhythm as the same thing Father's workshop had found in its second year: not fast, not effortful, just the cadence of something that had learned what it was.

***

Wanqing's morning message came at seven forty.

She was at the bench already — the spring seminar had started in the first week of March, and she'd begun arriving early on some days to work through the unit preparation before the seminar sessions. I'd arrived some Tuesdays to find her already there with the thermos full and the materials spread out. She'd been doing this since March. I hadn't asked about it.

*The maple is the same stage it was April 1 last year.*

I looked out the dorm window. The angle showed the maple's upper canopy — the mid-spring green, the young leaves fully open now, past the early-bud stage of March, into the full April growth. She was right. The same stage, same tree, same bench, the third consecutive April I'd observed it at this point in the cycle. The maple's pattern was as reliable as the surgeon referral network, as reliable as the Iron Hills zone at 3 AM. Things that had cycles and held them.

*Yes,* I sent.

*Third year in a row I've been at this bench in early April.* She sent it as an observation rather than a statement — no conclusion attached, just the fact noted.

*Yes,* I sent again.

*You'll be here later.*

*Afternoon.*

She replied with a period. Which was her conversational marker for "acknowledged, no further response needed" — she used it when a message needed receiving but not answering. The period had been in her usage pattern since the first year. I'd learned its precision gradually, the way you learned anything through repetition without being taught.

I looked at the April maple from the window for a moment longer.

Three years at this bench. Three Aprils.

Then I turned back to Chapter 4.

***

Wenqing's Volume 2 note arrived at ten AM.

*One year since the class activation (April 13, 2017 → April 13, 2018 is tomorrow, but the March floor data closes the year cleanly). The March Floor 20 average: 3h 09m. Three minutes below the February average of 3h 12m. The asymptotic curve is still moving.*

*What's the theoretical floor,* I sent. We'd discussed this before but not with a specific number.

*The minimum time to clear all 20 floors with a 133-member formation, assuming perfect information and perfect execution at every decision point — skill timing, Phase transitions, positioning, healer distribution, tank rotations.* He paused. *My estimate is 2h 15m. That's with zero latency, zero misreads, and full formation coherence at every moment. It's the number the formation would hit if every member was executing the optimal decision at every second of the run.*

2h 15m. The March average was 3h 09m.

*What's the gap.*

*54 minutes above the theoretical floor. That's 71% efficiency relative to the perfect-execution model.* Another pause. *Higher than I expected at one year post-activation. Significantly higher. The formations I've studied in the public records — guilds operating for three to four years at their peak — are typically at 65 to 75% efficiency against their class configurations. You're at 71% at year one. The class integration is what's driving it.*

The best formations in the public record, after three or four years, were operating at 65-75% efficiency. The Heaven-Severing class was getting the guild to 71% in one year. The class was what it had been designed to be — not the Berserker path's ceiling but something that kept integrating past the point where a standard class plateaued. The resonance had a floor. The floor kept rising.

*Is there a projection for when the curve stops moving,* I sent.

*No. The data would need to plateau for me to estimate the plateau. It hasn't plateaued.* Wenqing paused. *I'll note when it does.*

He'd been patient with uncertainty across 847 pages of Volume 1 and however many pages Volume 2 had accumulated. He didn't force projections when the data didn't support them. He built the model structure and waited for the data to arrive. He'd been doing this since October 2015 and it had produced 847 pages of accurate documentation. The patience was not restraint — it was the working method. You documented what was real, and the real would eventually show you its shape.

I turned back to Chapter 4.

***

The afternoon bench. April 1.

Wanqing had the spring seminar materials for the third unit and the thermos. The April campus — the mid-spring quality that was different from March's quality in a specific way, the warmth settled rather than tentative, the light consistent at an angle that meant another hour of sun past where it had been in March. The campus population was in operating rhythm. Semester Month Two. The difference between Month One and Month Two was exactly the difference Wenqing described in the resonance integration: by Month Two the initial parameters had been absorbed and the operating standard had moved up.

The maple was in full young-leaf green.

I sat down. She poured.

I told her about Wenqing's note — the 3h 09m, the theoretical floor of 2h 15m, the 71% efficiency, the asymptotic curve still moving a year after the class activation.

She listened without writing anything. When I finished she was quiet for a moment.

"71%," she said.

"Yes."

"Against a perfect-execution floor."

"Yes. Which can't be achieved. But it's what the math says the formation would hit with zero latency and zero error."

She turned a page. "The workshop is running at approximately 80% of Wanqing's projected ceiling right now — not the theoretical maximum, but the ceiling she modeled based on the actual referral network parameters. 80% of what the parameters predict is achievable." She turned the page back. She wasn't comparing the numbers to draw a parallel — she was noting that she recognized the kind of reasoning. "The class floor is the same type of number. It tells you where the gap is."

"Yes."

"And the gap is still closing."

"Wenqing says yes."

She looked at the young-leaf maple. The April afternoon. The warm light on the campus paths. "June is the thesis defense."

"June 22."

"And after June—"

My phone received a notification.

I looked at it.

The notification was a direct personal message. Not the guild communication channel. Not any of the regular channels. The personal message thread — the same thread Bai Yueran had used on December 12, the one that had held eleven sentences and the question that was five words.

The message was nine words.

*I'm ready. We need to talk. Not in-game.*

I read it once. Read it again.

She'd said in February that she had something to tell me that she wasn't ready to say yet. She'd said the thinking would take longer than December. She'd said she'd come back in spring. She'd said she'd contact me when she was ready, not in-game. She'd kept each of those things precisely. The nine words were the arrival at the end of what she'd described in February. The thinking had taken seven weeks. She'd said longer than December. December to April was four months. Longer was accurate.

It was April 1.

She was ready.

Wanqing, across the bench, saw me stop mid-sentence. She looked at me with the full-situation attention — the attention she used when the context had changed and she was reading the new context.

"Bai Yueran," she said. Not a question.

"Yes."

"She's ready."

"Yes."

She was quiet. The April bench. The full young-leaf maple. The warmth in the afternoon light. The thesis in Chapter 4. The defense in June. The guild at 71% efficiency and the asymptotic curve still moving.

She set the seminar materials down. Not beside her — she set them on the bench with a finality that meant she was done with them for the moment. She'd never done that in the middle of a session before. She always kept working unless something required her full attention. The set-down meant this required it.

"Tell me," she said, "what you're going to do."

Not: *what will she say.* Not: *what do you think it means.* What are you going to do.

I thought about it. Not because I didn't know, but because she'd asked it specifically and the specific form of the question deserved a specific answer.

Bai Yueran had been moving toward something for nine months. From the first meeting in May 2017, when she'd proposed the alliance with the precision of someone who'd been watching and waiting for the moment it was bilateral. Through December, when the five-word question had been asked. Through February, when the question had grown and she'd said she wasn't ready yet. Now it was April and she was ready.

*I'm ready. We need to talk. Not in-game.*

The arc of the phrasing was its own answer. Not a question. Not a request for analysis. A statement about her own state. She'd arrived at what she wanted to say and she was saying it was time to say it.

"Meet her," I said. "Listen to what she says. Respond honestly."

Wanqing looked at me. The April light. The mid-spring maple that had been at this bench through three winters now, the same tree cycling through the same stages every year. The fourth year. The thesis and the defense in June and whatever Hangzhou looked like after June.

"Good," she said.

She held my gaze for a moment — the direct, full-attention look she used when the thing being said was the full thing and not an approximation of it. Then she looked at the maple.

Then she picked up the seminar materials and turned back to the problem set she'd been marking.

I looked at the phone. The nine words still on the screen.

I sent: *When and where.*

The reply came in less than three minutes.

*Saturday. Same café.*

*Saturday,* I sent.

*Yes,* she replied. And then, after a brief pause: *Thank you for being patient.*

I set the phone face-down on the bench.

The April maple. The thermos. Wanqing working through the problem set, her pen making the small marks that were her grading notation, the papers arranged in the order she'd review them. The sound of the April campus — the warmth settled, the birds in the campus trees, the rhythm of a semester running at its operational pace.

The arc was not ending.

It was becoming something else.