188: The Draw

The CW III group stage ran over two weeks.

Fourteen days, six scheduled matches, four wins required to advance as a top-two finisher from the group. The fixed member cap for group stage matches was 80 — the tournament organizers used a reduced cap for the early rounds to limit the advantage of high-population guilds and to test formation quality at a scale that was harder to compensate for with roster depth alone.

Wenqing had factored the 80-cap into his models in July. The core six formation members — Old Wolf, Ironmark, TwilightTide, Zhu Yuhan, Cloudrift, and the two primary damage dealers in the inner formation — were all within the Sovereign's Reach radius regardless of total formation size. At 80 members rather than 133, those six represented a higher proportion of the total formation. The math of the reduced cap was simply favorable: less formation to cover with the same resonance zone.

Match 1: Unnamed Guild, unseeded and unranked. October 14.

We ran the full cap — 80 members, the formation configured for the reduced size. I'd been running 133-member sessions for ten months; the 80-member feel was the early formation configurations, before the November-December expansion. The smaller formation had a different texture from the inside — less noise in the peripheral channels, fewer decisions in the outer ring that needed to be tracked and left to others, the core coordination tighter simply because there was less of it. The smaller formation felt cleaner at the center. The resonance radius still covered the core formation correctly. The formation density was lower at the perimeter but the council members were within 12 meters. The core was the same. The edges were different.

Match result: 60–0. Forty-two minutes. No complications.

Wenqing's post-match note: *Output was 94% of the Floor 20 benchmark. Expected — the reduced formation size limits resonance coverage at the outer ring. The core six are within range; the outer formation members are not. Recommend adjusting the outer-formation composition for subsequent matches to prioritize resonance-range members over raw output.*

The outer formation. The players at positions thirteen through eighty who were currently outside the resonance zone. If we filled those positions with guild members whose roles benefited most from resonance augmentation — the secondary damage dealers whose output multiplied best under the timing optimization — the effective coverage would improve even without changing the total formation size.

I forwarded Wenqing's note to the formation captain roster. Old Wolf, TwilightTide, and Zhu Yuhan reviewed the outer-ring composition and submitted a revised lineup for Match 2 within eighteen hours. The turnaround was characteristic of how the council worked at this stage — they'd been operating together for long enough that the gap between a direction and its implementation had compressed into a day.

Match 2: River Stone Guild, mid-server, unseeded. October 18.

The revised outer-ring composition ran the resonance augmentation to nineteen of the eighty formation members rather than fourteen — a 36% increase in effective coverage within the fixed cap. The outer formation closed the perimeter faster in Phase transitions. The timing on the match felt cleaner than Match 1 despite the better opponent — the formation had found its 80-member rhythm. Match result: 60–8. Forty-seven minutes.

Match 3: Amber Ascent, server rank 12, unusual formation composition. October 21.

Ningxia's pre-match analysis had flagged Amber Ascent in the July document. Her post-draw note was more specific: *Amber Ascent uses a mobile defense configuration — rather than a fixed formation, they rotate their damage dealers in a shifting-radius pattern around a central healer cluster. The pattern is designed to prevent resonance-class or AoE-class abilities from locking onto a consistent formation position. This is unusual. It is likely a preparation specifically for our class mechanic.*

She'd been watching the server's competitive community absorb the April 13 class activation. Someone in Amber Ascent's analytical layer had read Wenqing's posted combat log, understood the resonance's dependency on proximity, and had built a formation specifically designed to counter it. Not an improvisation — a deliberate preparation over four months. They'd seen the public data and had done good analytical work with it. The kind of work that deserved acknowledgment even from the person it was designed against.

They had adapted their formation specifically for Sovereign's Reach.

Wenqing: *If the rotation pattern is consistent, the resonance will still apply at the moments when rotation-cycle members pass within 12 meters. The augmentation windows will be intermittent rather than continuous. Recommend tracking the rotation cycle and timing attacks to the window when the largest cluster is within range.*

*Timing attacks to the rotation window,* I sent.

*Yes. The resonance compounds when multiple adjacent members are in range simultaneously. If their rotation cycle brings eight or more members within 12 meters every 40 seconds, those 40-second windows are where the pressure should concentrate. Outside those windows, maintain standard pressure — don't telegraph the timing by going passive between cluster windows.*

He'd designed an approach to the counter-design before I'd asked him to. He'd been thinking about Amber Ascent since Ningxia's July flag. The specific quality of his preparation — present at the analysis stage, not improvised at the deployment stage — was what made the match manageable.

Match 3. The rotation cycle was approximately 38 seconds — close to Wenqing's 40-second estimate. In the first two cycles I tracked the pattern without varying my attack timing, running standard pressure to avoid signaling the shift. I was timing the rotations in the peripheral corner of my attention: the specific 38-second count, the moment when the largest cluster converged inside the 12-meter zone, the three or four seconds of maximum resonance density before the rotation carried members back out. The counting became automatic after the first two cycles. The rhythm of it was its own information, the way a cadence tells you where you are before you look at the clock.

In the third cycle I concentrated the heaviest chains in the window when the cluster reached maximum density.

The resonance intervals in those windows: 28% augmentation average on the clustered members, significantly above the 8–15% standard for dispersed formation positions.

TwilightTide on the healer channel: *The pressure spikes are irregular but high when they come. I'm redistributing the heal cycles — holding load during the rotation's extended range and releasing in the cluster window.*

*Yes.*

*It's the inverse of the standard approach,* she said. *Usually I'm managing peak load continuously. This is managing load in concentrated pulses. The skill requirement is different.*

*How is it landing.*

*Better than I expected. The cluster windows are predictable — the rotation pattern is consistent enough that I can anticipate the pulse timing. The rotation team is running the same cycle. They don't vary it.*

She'd adapted to the inverse pattern in three cycles. I noted that and kept the timing. The cluster window at second 38 of cycle 4 came with the formation ready for it.

Match result: 60–31. Fifty-four minutes. The longest match of the group stage.

Amber Ascent's approach was genuine. It wasn't incompetent preparation — it was a reasonable counter-design that had required our formation to adapt in real time, and the adaptation had worked by a margin that wasn't comfortable. Sixty to thirty-one meant they'd scored thirty-one points against our 60-point threshold. They'd gotten through our formation. Wenqing's post-match note: *They identified the resonance mechanic and designed against it. Their analysis of our combat log data was accurate. Their execution was also clean — the rotation team coordinated well. They're a better opponent than their server ranking suggests.*

They'd see us again in the knockouts.

Match 4: Ironcloud Alliance. October 25.

The seeded match. Wenqing's note from July had described the split-anchor configuration. I'd reviewed the configuration analysis twice before the match: two damage dealers at the central aggro position, the formation built to force a prioritization choice on the opposing guild — which of the two anchors to pressure, knowing that pressuring one allowed the other to operate freely.

With Sovereign's Reach, the answer was: neither, exclusively.

The resonance augmented the formation's output across all adjacent members simultaneously. When Ironcloud split their anchor, the augmentation was applying to six of our formation members — enough to maintain simultaneous pressure on both anchors without a prioritization choice. Our formation didn't face the dilemma their configuration was designed to create. The split-anchor formation's design logic assumed the opposing formation would face a resource allocation problem. We didn't have the problem.

TwilightTide, at minute fourteen: *They're recalibrating. Their healer layer is repositioning — they hadn't anticipated the bilateral pressure. They're not set up for it.*

*Hold positions.*

Phase 4 of the Ironcloud match — the competition format ran tournament Phases rather than dungeon Phases — hit at minute twenty-two. Their split-anchor formation had been holding but losing ground steadily, two-anchor pressure creating attrition they hadn't modeled. At Phase 4, they consolidated to a single anchor. It was the correct adjustment — the bilateral pressure had been costing them more than the split was gaining. But the adjustment had come four Phases into a match that had been running against them since the second Phase.

Too late. The momentum was already ours and the Phase 4 consolidation hadn't recovered the deficit.

Match result: 60–17. Forty-nine minutes.

Group stage results: four matches, four wins, top finish in Group A. Amber Ascent finished second, which meant we'd face them again in the knockout rounds. They knew our timing by now and we knew their rotation. The next match would be different in the specific way that second encounters were different — not the initial read of an unknown formation, but the adaptation of an adaptation.

***

Wanqing at the bench on October 28. The mid-fall maple — the specific mixed-color phase, orange and green in the canopy at the same time, the stage that arrived after the early-turn and lasted for two weeks before the leaves fell. The bench had its October configuration: a degree cooler than September, the thermos more necessary than in the summer. The mixed-color canopy cast a different quality of light than the spring green — warmer in tone, the orange filtering the afternoon sun into something that felt like the season itself was visible in the light.

"Four and zero," she said.

"Yes. The Group A final was Amber Ascent versus Ironcloud — Amber Ascent advanced second. We'll face them in the knockouts."

She looked at the mixed-color maple. "They adapted for your class mechanic."

"Ningxia flagged the formation type in July. The analysis held."

"And you adapted for their adaptation."

"Wenqing's timing approach worked. The rotation cycle was consistent enough that concentrating attacks in the cluster window was reliable. TwilightTide adapted the heal cycle to match."

"TwilightTide inverted her approach."

"She said it was better than she expected. The rotation pattern was consistent." I looked at the maple. The October color had come in the way it did every year — not gradually but in a week, the tree shifting from late-summer green to mixed-color in a single week in mid-October. "Their formation design was genuine work. Four months of preparation based on the public combat log data."

"Which you posted."

"Wenqing posted the relevant parts. The soul-bind data wasn't in the posted log. The Void Severance skill wasn't in the posted log. The resonance radius and timing data were."

Wanqing turned a page of the fall seminar materials. "Strange," she said.

"What."

"That the lower member count makes the class more effective relative to its opponents." She looked at the maple. "You've been optimizing for the 133-member formation since December. The tournament's reduced cap makes the ratio more favorable."

"The resonance covers the core six regardless of total formation size. At 80 members, those six represent a higher proportion of the formation than at 133. The formation you built for 133 is proportionally more resonance-dense at 80."

"So the guild that optimized for precision over scale benefits from the precision-testing format."

She'd derived the competitive logic from the mechanic description. The same conclusion Wenqing had noted in his July model — the 80-cap format created conditions where precision-based formations outperformed scale-based ones relative to the full-roster context — but arrived at through reasoning from the mechanism rather than from the model. Both routes led to the same place. She didn't need the model to see the logic. She needed the mechanism and she built the logic herself.

"Yes," I said.

"Good," she said. And went back to the problem set. The campus around us had the specific mixed-color quality of the October canopy — the orange and green that wouldn't last more than two weeks before the leaves fell and the grey arrived. Two weeks of this particular light.

The knockout rounds were next. Amber Ascent in the quarterfinals, at a different member cap, with a formation that had already adapted once and would adapt again. The second encounter was a different problem from the first — both guilds had data the other one had seen produce, which changed what mattered. The relevant unknown was not the formation mechanics, which were both visible from the group stage record. The relevant unknown was who had found the better counter to the thing they'd seen.

November was one week away.