186: Void Severance

The first test deployment of Void Severance happened August 4.

Wenqing had recommended waiting until August — after the CW III opponent guild analysis was complete and Ningxia's preliminary bracket simulations had identified the competitive contexts where the skill would have the highest yield. He'd been aware of the skill since April 13, when the full class tree had populated on my interface. I'd described the tooltip to the council at the April 14 session briefing. Since then, by council agreement, the skill had been private: not deployed in the regular Floor 20 sessions, not mentioned in any channel visible outside the council, not documented in any posted combat log.

Four months of running the Floor 20 sessions without using it, which had required tracking the cooldown opportunity windows each run — noticing the moments in Phase transitions where the skill would have been most effective — and setting those moments aside. The discipline of not deploying a capability you understood in a context where you could use it effectively was a particular kind of patience. Not the patience of building toward something unknown. The patience of holding something known in reserve, which was harder in some ways because the cost was visible and the benefit was distant.

Every Phase 1-to-Phase 2 transition I'd watched the cooldown timer sit at zero. The 12-second window. The formation repositioning. The moments where I knew what the skill would do and had chosen not to use it. Four months of those choices, across twenty-odd Floor 20 sessions. The council had agreed to it without discussion because they understood the logic.

The discipline of holding a capability in reserve was a different kind of patience than the patience of building toward the capability. Both were necessary. Four months had confirmed they were compatible.

The test was not in the regular session. We ran a controlled version on an alternate instance — a Floor 20 reset that Wenqing had calculated would produce comparable boss mechanics at the same difficulty calibration. The goal was to test the skill's effect on the Phase transition windows and document the results before deploying it in competition. Testing first, then using what the test confirmed.

The Phase 1-to-Phase 2 transition for the Abyssal Sovereign was the critical window: Phase 2 initiated the void field expansion and the formation's clustering requirement. The twelve seconds of transition — boss animation, field expansion, formation repositioning — was the highest-risk interval of the encounter. If Phase 2 opened with two or three formation members out of position, the void damage in the first cluster-formation cycle could cascade into a wipe.

In thirty-eight Floor 20 clears, we'd had two Phase 2 opening wipes. Both from the same cause: the outer formation's repositioning taking thirteen seconds when the window was twelve. A second and change was the margin between a clean clear and a reset. I'd known this. I'd been watching the twelve-second window every clear since December, noting the outer formation's timing, knowing what was possible in the alternate instance.

Void Severance's tooltip: *Applies a disruption to one target's next three attack sequences, causing their damage windows to misalign with standard timing. Effect duration: 12 seconds.*

The Phase transition window was 12 seconds.

The alignment was exact.

***

At the Phase 1 kill threshold, I activated Void Severance.

The visual effect was minimal — a brief darkening of the target aura on the boss, gone in under a second. Nothing that announced itself. Nothing visible from outside the skill's targeting system as anything other than a momentary graphical artifact. The subtlety was intentional design: a capability that revealed itself in the combat log's timing data rather than in a visible effect was harder to identify and counter.

The boss entered Phase 2.

The first three attacks of Phase 2 arrived with irregular timing. Not delayed. Not absent. But desynchronized — the attack intervals arriving approximately 0.4–0.7 seconds off the standard Phase 2 timing that the formation had memorized over thirty-eight Floor 20 clears. The formation's muscle memory expected the attacks on a specific cadence. The attacks arrived on a different cadence.

TwilightTide on the healer channel: *The Phase 2 opening pattern is wrong. The timing is off. Not wrong-wrong — off. The attacks are coming but not when they should.*

*Off is correct,* I sent. *Void Severance disrupts the Phase transition attack timing. Adapt or ignore — your call.*

*Adapt,* she said immediately. She didn't have to think about it. Then, four seconds later: *The disruption is fading. Attack 3 just came back to the standard window.*

The disruption lasted exactly 12 seconds. Three attacks with shifted timing, then back to the Phase 2 standard pattern. The tooltip's description had been exact — the skill did what it said it did. No more, no less. Exactly calibrated.

Wenqing's log, posted in the council channel within two minutes of the Phase transition: *Phase 2 formation repositioning completed 6.4 seconds faster than the August average. The disruption gave the formation an extended window to reach the clustering positions before the first AoE escalation cycle. The additional 6.4 seconds is sufficient for the outer formation's complete repositioning under current movement speed parameters.*

6.4 seconds faster repositioning.

Over a five-hour clear, that was one instance of 6.4 seconds. But Phase transitions were the high-risk windows. The skill's value wasn't in DPS output — it didn't add to anyone's damage numbers directly. It was in transition management: buying the formation time in the moments where time was most scarce. The two Phase 2 opening wipes in thirty-eight clears had been the outer formation's timing by one second. The skill erased that margin and gave six seconds of buffer.

One second had been the margin between a clean clear and a reset. The skill gave six. That was five seconds of pure structural protection, visible only to the person who knew what had been at risk.

*Second test,* I sent. *The Phase 2-to-3 transition.*

The Phase 2-to-3 transition was less dangerous than Phase 1-to-2 — the formation was already clustered in the void field formation, so the repositioning requirement was minimal. But the soul-bind selection in Phase 3 happened within the first eight seconds of Phase 3 initiation. If Void Severance disrupted the opening attack cycle of Phase 3, the soul-bind timing would shift.

I ran the cooldown to zero. 180 seconds. The test instance continued running, the formation holding Phase 2 at controlled DPS to manage the kill timing. At the Phase 2 kill threshold, I activated Void Severance again.

Phase 3 opened.

TwilightTide: *The soul-bind targeting is delayed. The first attack cycle is off — same pattern as Phase 2.*

*Yes.*

*The soul-bind is going to target — it's targeting Ironmark. Standard soul-bind target would be—* she paused. *In thirty-eight clears, the soul-bind has never targeted Ironmark.*

*No.*

*You shifted the target selection.*

*The Phase 3 selection algorithm uses the first attack cycle timing as part of its target selection logic. Shifting the timing shifts the target.*

Wenqing: *Logging. This is not in any public analysis of the Phase 3 soul-bind mechanic. The selection algorithm's dependence on Phase-initiation timing has not been documented on any server. If this is correct, it means the soul-bind is not random — it's calculated, and the calculation can be influenced.*

*Not publicly,* I sent. *I know the mechanic.*

Previous timeline. Floor 20 in the previous timeline's competitive context — the version of this dungeon that had run on a server I would never return to, in a guild that no longer existed, alongside people who were currently children or hadn't met me yet. The Phase 3 soul-bind had targeted TwilightTide in every run because of her healing output profile in that guild. I'd noticed the pattern. I'd tested the algorithm's logic using a different approach in a different guild structure. I'd confirmed the calculation. The knowledge had been sitting in the ground layer of everything since then, waiting for the version of the game where it could be used correctly.

*I'll add it to the private analysis document,* Wenqing said. *This does not leave the council. The soul-bind targeting influence is too significant to share in the posted combat log data.*

The council agreed. No one argued for disclosure. The soul-bind mechanic, once known, could be used by any guild that had the information. The correct time to share it was after CW III, if at all.

***

Wanqing was at the bench on the evening of August 4. Summer Thursday, the full canopy, the thermos. The bench was in its late August configuration — the canopy still at full density but with the first suggestion that the tree was aware of what was coming in October. Not visible as color change yet. Something in the quality of the light through the leaves, the angle the late-afternoon sun came through. August light through a full maple canopy had a particular quality — not the bright filtered green of June, not yet the heavy late-summer quality of September, but something between them that lasted only a few weeks. The bench caught it at the angle where the canopy was densest.

"The Void Severance test," she said.

"Yes. Three functional applications confirmed: Phase transition repositioning window, Phase initiation timing disruption, soul-bind target selection influence."

She turned a page of the seminar materials — the summer seminar's final unit, the dense applied sections that she'd been working through for two weeks. "When will you use it in competition."

"Ningxia's bracket model has it deployed in the quarterfinal round or later. The rounds with the most complex formation opponents. Using it in early rounds reveals the capability before the high-stakes context where the reveal would have the highest yield. The pattern is the same as holding the soul-bind mechanic back from the public combat log — keep the capability private until the context justifies it."

"The same logic as not posting it in the public log."

"Yes. The skill is private until the context justifies the reveal."

She looked at the summer maple. The late August quality of it — the full canopy holding, but the sense of something preparing to turn. "The Phase 3 soul-bind selection," she said. "The previous timeline."

Not a question. The same pattern as May — she tracked when information implied the previous timeline's knowledge and made the inference.

"Yes."

She was quiet for a moment. The summer bench, the evening light filtering through the full canopy. A student on a distant path, moving at the summer pace. The campus was quieter in August than in any other month — even quieter than the exam-period quiet, because at least the exam period still had people in it. August had the quality of a campus holding its breath between semesters.

"You carry a lot of that," she said.

"Yes."

"And you deploy it when it's useful and set it aside when it's not."

"That's the approach."

She turned another page. "It must be strange," she said. Not quite a question. The specific register she used for observations that were also, if I chose to answer, invitations. She'd used it twice before — the first time in March 2016, the second time in January 2017. Each time was a different depth of the same question, which she only asked again when she thought I might have more to say about it than I'd had before.

I considered it. The question deserved a real answer.

"It was strange in the first weeks," I said. "In March 2015, sitting in the hospital waiting room, knowing the fund could be built and the transplant would succeed. Knowing that if I did the work correctly, the outcome that had happened the first time wouldn't happen the same way." I looked at the summer maple — the full summer density, the same tree it had always been at this bench in this August. The tree was the same tree every August. It didn't carry the weight of its previous Augusts. I did. "The strangeness was a kind of grief. For the previous timeline. For the version of things where I'd learned all of this too late."

She was very still. Not the problem-set stillness — the other kind. The kind where all of her attention was in one place.

"It's less strange now," I said. "It's just the layer the decisions are built on. The previous timeline is — not behind me, exactly. More like the ground the current one is standing on. The weight of it doesn't go away. But it's not strange anymore. Strange was the early feeling. Now it's just what's true."

She looked at me. The specific quality of attention she gave things that required it — full and quiet at the same time. Not waiting for more, but not closing the door on more. "The layer the decisions are built on," she repeated. Not questioning — tasting the phrase. The way she repeated phrases she was going to keep.

"Yes."

She went back to the seminar materials. The August bench. The full-summer maple with its first whisper of what was coming. The campus between semesters, holding the specific quality of August quiet.

The test had confirmed three applications of Void Severance. Wenqing had four months to incorporate them into the competitive model. Ningxia would update the bracket simulations with the new capability data. The skill's deployment context was set: quarterfinals or later, maximum-stakes context.

November was three months away.