241: 2020

The new decade started at the bench.

Not the December bench — the January bench, different in its first-of-year quality. The winter was the same as all the Hangzhou winters: cold but not severe, the maple bare, the campus in the between-semester quiet. The first bench of a new decade. The fifth year of the same bench, which made the decades something the bench had outlasted rather than something that interrupted it.

The campus was in its post-holiday configuration — the students returned but the full semester rhythm not yet established, the cafeteria hours irregular for another week, the library operating at reduced staff. The January light came at a low angle even at midday, and at the bench it meant the shadows of the bare maple branches lay long across the stone. Different from the December light, which had been colder and more direct. January had the particular quality of a year that had just opened: the structure of the coming months visible in outline, none of the details filled in yet.

Wanqing arrived with the power grid data printouts and the green thermos she'd started bringing in the second year of bench winters. The thermos habit had started as a practical concession to the cold and become part of what the winter bench looked like.

"What does 2020 look like," she said. Not the question she'd asked in December — that had been about the near-term structure. This was the longer view.

I looked at the January maple.

"Volume 3 of the archive continuing. Wenqing's monthly entries are running longer than Volume 2 — he's documenting the things the analysis doesn't predict, and there are more of them now. The formation finds things under match pressure that weren't in any model." I looked at the campus. "The third paper — the power grid data integrated by mid-year, submission by autumn. The workshop in Suzhou, Feng Li's apprenticeship in its second year." I looked at the bench. "CW VI registration in October. The dual-flow second layer against whatever Bai Yueran builds for third."

"You're planning against what you don't know yet," she said.

"Yes. The planning layer is different from the operational layer. The operational layer is the match, the session, the bench. The planning layer is what makes the operational layer possible."

She turned a page. "What else."

"FrostDragon's documentation issue. The narrative record is in the database now, but the early period is still unprotected. The Lu Yifan network's probing of the certification infrastructure — Wenqing's been tracking 12 access requests since February 2019. It's been quiet since September. They're not probing actively. They're waiting."

"Waiting for what."

"Either for a vulnerability to appear," I said, "or for us to make a move that creates one."

She looked at the January bench.

"You haven't created one," she said.

"No. The documentation layer is what it is. The blockchain record holds. The certification is what it was in April 2018. Nothing has changed that creates a vulnerability."

She turned a page. "What about the approach that doesn't attack the certification."

"Yes," I said. "That's the approach I'm not seeing yet. If they've been studying the certification infrastructure since February 2019, they know the blockchain anchor is unbreakable. They'll try something different."

"Something different," she said.

She was quiet for a moment.

"The platform function they haven't used yet," she said.

I thought about what Wenqing had documented: four platform functions tested between 2017 and 2018 — seeding algorithm, performance verification, ladder ranking, session validity checker. All four had been tested and then backed off when the response was effective.

"They've tested the platform functions," I said. "All four. None produced a sustained advantage."

"Are there other platform functions," she said.

I thought about it.

"Yes," I said. "Player account verification. The function that confirms account ownership and playtime authenticity." I looked at the January maple. "They've never tested that one. If they could question the playtime authenticity of the formation's core members — suggest that the class activation was performed by non-genuine accounts, boosted rather than developed through play —"

"The certification would still hold," she said. "It's timestamped, blockchain-anchored."

"Yes. But the argument wouldn't be against the certification. It would be against the guild's competitive eligibility status. Different regulatory mechanism."

She turned a page. "Not Wenqing's documentation layer. A different regulatory layer."

"Yes. The tournament committee has a separate eligibility verification process. It's less rigorous than the certification database. Older infrastructure. Single-point verification."

She looked at the January bench.

"Wenqing needs to certify the account authenticity before that approach arrives," she said.

"Yes."

"When will he see it."

"I'm going to send him the question today," I said.

I did.

Wenqing's reply came two hours later: *The account authenticity function. I saw three access requests to the eligibility verification API in November. I had them categorized as standard pre-CW V tournament administration. I'm reviewing now.* He paused. *The November requests weren't administration. They were structural queries — the same pattern as the certification infrastructure probing.*

He'd seen the data. He'd miscategorized it.

*Can the eligibility verification be pre-certified,* I sent.

*The eligibility verification system has a self-attestation function that generates a timestamp record,* he said. *If we submit the account authenticity data to the self-attestation function now — before the eligibility challenge arrives — the timestamp protects the attestation against retroactive challenge.*

*Submit it,* I sent.

*January 18,* he said. *I need a week to compile the complete account data. All 133 members.*

Seven days.

*Submit on January 18,* I said.

He submitted on January 18. The timestamp record was created at 10:47 AM: 133 member accounts, playtime authenticity attestations, match history, session logs. Everything the eligibility verification system could ask for, already in the system, already timestamped. The approach that hadn't arrived yet now had nothing to find.

This was the same principle as the blockchain certification in April 2018, applied to a different regulatory layer. The value of the timestamp was in its precedence — it existed before the challenge, which meant the challenge could only argue against the timestamp itself, not against the substance it protected. The approach that worked against the certification database worked here too. The same logic, a different surface. Wenqing's pattern was consistent: find the next surface before it becomes the point of attack.

***

The January-February formation sessions.

Floor 20 in January: 2h 48m. 81% efficiency. The asymptotic curve moving by one percentage point — the slowest single-point improvement in the guild's history.

Wenqing: *One percentage point in two months. The rate is consistent with the asymptotic model at this efficiency level. The theoretical floor is 2h 15m. We're at 81%. The remaining 19 percentage points will require either a technique advance or a formation structure change.*

*What kind of technique advance.*

*I don't know,* he said. *I've been watching for it since the plateau began. TwilightTide's healing layer is the most likely source — the accelerating in-match calibration from CW V could translate to session efficiency gains that I haven't modeled.*

TwilightTide's note for January: *The Tuesday sessions have shifted. The rhythm is arriving at minute 36 now — 8 minutes earlier than the October baseline. I don't know why. The composition was the August change; January feels like something else.*

Something else in January. A second shift, four months after the first.

*What's different in January,* I sent.

*The CW V semifinals,* she said. *The match at 69 minutes — the 7-second cycle and finding each new rhythm faster than the previous. The match did something to the session practice that I'm still working out.*

The match had changed the sessions. The under-pressure calibration in December had changed what the 3 AM window felt like in January.

*The practice is learning from the match,* Wenqing sent when I shared this. *Not the reverse. The standard model is that practice prepares for the match. The match is prepared for by practice. In this case, the match is teaching the practice.*

The match teaching the practice. A new direction in the relationship between the two.

***

Mu Qingyao's February message.

*The six-month drill data is complete. The timing signal member's read time is at 0.7 seconds.*

0.7 seconds. The formation was reading at near-Black Dragon formation speed, without the resonance or the Pioneer's Path class data.

*Full-formation floor-clearing at 0.7 seconds read time,* she continued. *Our Floor 15 record cleared yesterday at 4h 12m. The previous record was 5h 01m. One hour faster.*

One hour faster on the most complex floor they could currently reach. The constrained-repetition drill, the coordinating foundation rebuild, the calibration acceleration that had appeared in training — all of it arriving in the floor time.

*Wenqing is going to want the data,* I sent.

*I know,* she said. *I've already formatted it for him.*

She'd formatted the data for Wenqing's archive before I'd asked. She knew what the documentation layer needed.

*Good,* I sent. *Send it directly to him.*

*Already sent,* she said. *He's been tracking the Tianhe formation's development since October 2018. He should have the full data for the record.*

Wenqing's reply to her data, forwarded to me: *The Tianhe formation's development path from October 2018 to February 2020 is the cleanest external validation of the resonance integration principles I've seen. They're implementing the same design through different mechanisms and arriving at similar efficiency curves. This belongs in Volume 3.*

The Tianhe formation in Volume 3 of the Black Dragon archive. A formation on a different server, with different class mechanics, developing in the same direction through the same principles.

The documentation was expanding to include what it had propagated. The archive had started as a record of one formation's development. By Volume 3, it was becoming a record of the development of a development method — the principles spreading and producing independent results across formations that had never shared a session.

That was not something that could have been planned at the start. In October 2015 the archive was a record of 12 members on Floor 3. There was no model then for what happened when the documentation reached outside the formation and produced independent development elsewhere. The model had grown from the data, the way every useful model in the archive had grown — observed first, named after.

***

Bai Yueran's February message.

*The third layer is at page 7 of the design document. I'm not telling you anything about it until October registration.*

*Good,* I sent.

*FrostDragon,* she sent. *He messaged me.*

That was unexpected.

*What did he say.*

*He wants to talk to both of us. Not in-game. He's been watching both guilds — ours and yours — and he says he has something to share that's relevant to what we're both building.* A pause. *I told him I'd ask you.*

FrostDragon initiating contact with both guilds simultaneously.

*When,* I sent.

*He suggested March,* she said. *In person, if possible. He's in Beijing.*

In person. The first time he'd proposed IRL contact.

*Tell him March 15,* I sent. *We'll find the format.*

She sent back: *March 15.*

***

Wanqing at the February bench.

The late winter — the bare maple, the month before spring, the bench in its coldest-of-year quality.

"FrostDragon wants to meet," I said. "In person. March 15."

She looked at the bench.

"All three," she said.

"Yes. He has something to share that's relevant to what both guilds are building."

She turned a page.

"He's been watching for fourteen months," she said. "Whatever he's built in the watching — he's ready to share it."

"Yes."

She looked at the late-winter maple.

"Good," she said. "That's the right timing."

The late winter bench had its own quality distinct from January's. In January the cold was the cold of the year just starting, which had a kind of blankness — the months ahead not yet inhabited, the structure not yet loaded with the details that would fill it. By February the cold was the same temperature but the structure was already filling in. The semester had found its rhythm. The projects had their early shapes. The bench in late February held the cold of a month that knew what March would be, even if it hadn't arrived yet.

She turned to the power grid data.