196: January

The thesis outline was due February 28.

I'd started the outline in December — the framework had been clear since October, when the thesis topic had been fixed for six months and the evidence chapters were mostly organized in my head if not yet on the page. The topic: discrete mathematics applied to formation optimization in competitive game contexts — the academic version of what Wenqing had been doing since October 2015, translated into formal methodology and citation apparatus, with the session data presented as a case study corpus rather than as a guild record.

The thesis committee had approved the topic in September. The chair, Professor Liang, had described it as "unusual but methodologically sound" — the kind of approval that kept options open without committing to enthusiasm. He hadn't played competitive games. He understood graph theory and probability models well enough that he was willing to treat the game context as an applied domain and evaluate the methodology on its own terms rather than on the prestige of the subject matter. That was all I'd needed from him. Willingness to evaluate the methodology on its own terms was the permission I required; the rest was going to come from the work itself.

The outline was seventeen sections. The first three were the formal literature review — the academic work on formation coordination, distributed system optimization, and competitive behavioral modeling. The next eight were the case studies, which were the guild session data translated into the formal notation the literature review established. The final six were the comparative analysis and conclusions, where the case study results were measured against the existing literature's predictions.

I worked on the outline in the January mornings before the game sessions and in the evenings after. The outline took shape at a rate I recognized from the thesis proposal process: fast at the beginning, when the broad structure was being established, then slower in the middle sections where the specificity of the case study presentation required decisions about what to include and what to leave in the raw data. The fast part felt like clarity. The slow part was the actual work — each case study had its own structure and the structure had to be consistent with the structure of every other case study in the same section while also being specific enough to the event to be analytically useful. The outline was not the thesis. The outline was the map. The thesis would be built from the map.

The game sessions in January had the post-championship quality — regular Floor 20 runs without the competitive preparation layer that had been present since June. The formation ran clean. The resonance was the baseline now, not a discovery; the formation had been inside it for nine months and the integration was ongoing in the way that things continued to deepen after they had already arrived.

Wenqing: *Starting Volume 2 of the archive. First entry is the CW III post-match analysis. The next twelve months are the thesis year and the guild's regular maintenance cycle. I'll document what happens.*

*What does Volume 2's scope look like.*

*Unknown at the start. Same as Volume 1 — I'll document what happens and the documentation will accumulate as the events accumulate.* He paused. *I don't build the scope in advance. I find out what the scope is by documenting.*

He'd documented October 2015 through December 2017 without knowing it would be 847 pages. Volume 2 would accumulate the same way. The scope of a thing that was documenting events as they unfolded was, by definition, the scope of the events themselves. Wenqing understood this. He wasn't looking for the shape of the next two years before the next two years had happened. He was starting January 1 with the first entry and trusting the rest to follow from what actually occurred.

***

TwilightTide's January schedule had changed.

She sent a message on January 8: *The tour schedule is lighter this year. The management team and I renegotiated the structure in December — fewer promotional events, more recording-focused time. The tour gap opens an additional block in the week that wasn't there before.*

*What does the additional block look like for the sessions.*

*The same sessions. But I'll be available during the day on some Tuesdays and Thursdays when I was previously in transit or at promotional events.* A pause. *I'll still be in the three AM window. That doesn't change. But the gap in the day schedule means I'm more available in general — not just at three AM, but at the margins.*

More available. She was reconfiguring her professional schedule and part of the reconfiguration had created space that she was describing in terms of the guild, the sessions, the general margin. The specificity of how she was describing it said something about what the guild occupied in the structure of her time — not a priority competing against other priorities, but the baseline container against which other things were measured.

*What's driving the renegotiation.*

*The article,* she said. *The music publication feature. After April, the management team received a different composition of inquiry — more focused on the craft, the production choices, the reasoning behind the work. Less focused on the event, the appearance, the visibility. The promotional event volume was calibrated to the promotional event demand. The demand composition changed.*

The feature had changed the composition of demand. The music publication's twelve thousand words had moved the professional inquiry toward the substance of the work and away from the visibility of it. The people reaching out were reaching out for different reasons. Different questions about different things — not *can you attend this event* but *how did you approach this choice*. The two kinds of inquiry were compatible in principle and incompatible in practice, because the second kind wanted time and the first kind consumed it.

*That's good,* I sent.

*Yes.* A pause. *It's what I wanted the article to accomplish. I told Wenqing, when we were working on the data presentation, that I wanted the article to shift what people were paying attention to. It took several months for the composition to shift. But it shifted.*

The article had done what she'd wanted it to do. She'd wanted it to do it when she sent the first message about working with Wenqing in January 2017, before the article had even been commissioned. The twelve months between intention and result had been twelve months of the work doing what it needed to do — the article going out in April, the inquiry composition beginning to shift by summer, the management team noticing the shift by autumn, the renegotiation happening in December. A year-long lag between the initial intention and the structural consequence. She'd been patient enough to let the lag happen.

She was rebuilding the schedule around the new demand composition. The schedule was the practical expression of what the article had accomplished.

***

Wanqing sent a message on January 15. She was in Beijing for the winter — her home between the Hangzhou seminar seasons. The spring seminar preparation started mid-February.

The message was about the workshop model. She'd received the Q4 2017 data from Mingzhu and was updating the annual model.

*The year-two actuals are 22% above the projection I set in October 2016,* she wrote. *The original three-year projection needs to be rerun with the actuals. I'll have the updated three-year model by the end of January.*

22% above projection. The workshop had exceeded her careful model by more than a fifth. The number had a specific quality: large enough to require a parameter update, small enough that the structure underneath the numbers was still intact. If the actuals had been 60% above projection, it would have meant the model's framework was wrong — the assumptions about how referral networks and craft quality interacted would have needed revision. At 22%, the framework was right and the inputs had been better than expected.

*Does the model structure hold,* I sent. *Or are you rebuilding the framework.*

*The framework holds. The parameters update. The referral network density was slightly underestimated — the surgeon cohort is more interconnected than the October 2016 data suggested. The physicians who referred in December 2015 are still referring in January 2018, and their referral rate hasn't decayed. The actual referral persistence is higher than the model's assumption.*

She was updating the parameters, not the framework. The structure she'd built in May 2016 — the model's bones, the logic of how referral networks and craft quality and workshop capacity interacted — was still correct. The inputs had been better than projected. Father had built better relationships with the surgeon referral network than the conservative October 2016 parameters had anticipated. The quality had held and the evaluators had kept coming. The same thing Father had described at the kitchen table on December 28 — the March dyes, the two-year referrals, the quality that made the return warranted. Wanqing's model had captured the mechanism and slightly underestimated the magnitude.

*Three-year projection with the updated parameters,* I sent. *What does it show.*

*Not done yet. I'll send it when it's ready.* Then, after a pause: *How's the thesis.*

*Outline is due February 28. On track.*

*Good.* A pause. *I'll be in Hangzhou February 7. The spring seminar starts March 5.*

February 7 was one week before Bai Yueran's conference dates.

She'd told me in December that she usually came back a week early. She hadn't made the connection explicit. She didn't need to.

*Yes,* I sent.

*Same bench,* she said. *February 7, assuming the weather holds.*

The February bench. The maple in its February bare stage — the specific bare that was different from December's bare and November's bare. The February bare was the waiting-for quality, the still period before the first buds. Third year at this bench in this stage of the maple's cycle. I'd been at this bench in three different Februaries: the first when the guild was three months old and the Pioneer's Path was still being navigated, the second when the Floor 20 sessions were running and the class was accumulating, this one where the class was complete and the bench was the thing it had become over two and a half years of being itself.

*I'll be there,* I said.

***

The January guild sessions had a different texture than the competitive sessions — Floor 20 run as maintenance rather than as preparation for something incoming. The formation's maturity showed in the clean runs: the resonance intervals regular, the healers' response anticipatory, the Phase transitions reading faster than the analysis predicted. There was no opponent, no bracket, no window to prepare against. The formation was running for its own sake, and for its own sake it ran well.

Wenqing's January 20 note: *Floor 20 clear time for the January average: 3h 22m. The April 2017 record was 3h 36m. The formation has been improving in the post-activation period without specific optimization effort — the resonance is integrating more deeply into the formation's operational baseline.*

The resonance integrating more deeply. The class had been fully active since April 2017 — nine months. The formation had been running with it for nine months. The integration was ongoing and unstructured — not the result of specific sessions designed to deepen it, but the natural consequence of continued operation inside the condition the class had created. The formation was getting better by running. The running was the deepening.

*Is there a ceiling on the integration,* I sent.

*Unknown. The transitional period data suggested a ceiling at 100% accumulation — the full designation. The full designation doesn't have the same accumulation mechanic. It's a continuous integration rather than a quantified one.* He paused. *I'm building a new model. I don't have the parameters yet. The model structure will precede the parameter data.*

He was building a new model before he had the parameters. The structure first, then the data to fill it. The same way he'd built the [x] modifier model in January 2017 before he had enough data to confirm it, and the same way Wanqing had built the workshop model in October 2015 before she had enough data to confirm it. Build the structure. Let the data fill it. Find out what the scope is by documenting.

The thesis. The three AM sessions. Wenqing's Volume 2. Wanqing coming back to Hangzhou on February 7.

The year beginning.