108: Round 3: Redpeak Brotherhood

The match loaded at six AM and Redpeak Brotherhood was already in their formation.

Sixty-seven avatars arranged in three clean layers across the south end of the Riverbed terrain map. Three tanks in the first layer. Two each in the second and third. Twelve archers in a spread formation behind the DPS cluster that sat between the second and third layers. Their guild leader — a Lv 33 Warrior in Redpeak's crimson-and-black colors — stood at the center of the second layer with the kind of stillness you built by running the same formation for three hundred bracket matches.

The Riverbed terrain map was the committee's standard choice for mismatched bracket rounds — not a flat arena, but a river-gravel plain with low banks running east to west, two meters of elevation change at each bank edge that could break line-of-sight if you used it. Most guilds ran straight formations on this map because the bank-crossing cost movement time that smaller formations couldn't afford to spend. I'd mapped the banks during the preview window and decided the same thing: straight formation, use the flat central strip, trust the plan.

Twelve of us at the north end.

I held position and counted heads. In the old timeline, the kill-cam from a match like this would show the smaller guild either charging immediately — desperation — or backing toward the terrain wall — failure — or doing something so unexpected that the larger guild spent its first thirty seconds recalibrating and never quite recovered.

The arithmetic of five-to-one wasn't just about numbers. It was about the psychological weight of that ratio on the smaller side. Sixty-seven avatars in formation had a presence that twelve avatars didn't — the visual density of a deep-layered formation, the sound of sixty-seven players' movement in the terrain, the simple fact that when the first layer moved, the second and third layers moved behind it like a wall with internal momentum. The standard smaller-guild response to that visual pressure was to react to the wall rather than to what was behind it. That was why most smaller guilds lost: not because they were outplayed, but because they played against the thing they could see rather than the thing they knew.

We were going to need option three.

*Ding!* [Continental Committee: SEVERING LIGHT vs REDPEAK BROTHERHOOD. Match begins. Duration: 20 minutes maximum. Elimination condition: all opponents eliminated or flag captured. Begin.]

"Phase 2," I said.

"Phase 2," Old Wolf answered.

The Scattered Fan spread to its opening position. East point, Wanqing. West point, me. Center, Old Wolf and Iron Plum. The remaining eight members held center-rear, in the clustered reserve position we'd used as cover — they looked, from a distance, like a standard back-line that hadn't deployed yet.

Redpeak's first layer took three steps forward. Even pacing, even interval, the formation staying locked. Their guild leader said something on their voice channel that I couldn't hear but which caused the archers behind the second layer to shift their spread spacing by two meters — widening it, which was what you did when you expected the opposing formation to converge on a center point and you wanted more coverage for the expected convergence.

They were expecting Phase 1.

I was at Lv 34 as of Wednesday. The gap-crossing survival window was 5.7 seconds.

"Go."

***

The entry was 0.6 seconds and I'd run it fourteen times.

The difference between practice timing and match timing was the thing Zhu Yuhan had flagged in her pre-match preparation notes, and she was right about it. In the drill instance the gap had been clean geometry and known opposition behavior. In this match it was sixty-seven people who had been drilling their own formation for longer than I'd been building mine, and their first layer reacted a fraction faster than the drill opposition set — the two flanking tanks' response coming at 0.3 seconds instead of the projected 0.4. I had 0.3 seconds of additional resistance on entry that I hadn't modeled.

I crossed it anyway.

I crossed the four-meter gap in the half-second between the two flanking tanks' response time — they registered the movement at 0.3 seconds, began their closure rotation at 0.5 seconds, and I was through at 0.6 seconds before the closure completed. The gap zone had the particular quality of a compressed space: the DPS line in front of me, the first-layer tanks re-forming behind, the cage of the Scattered Fan east and west points holding the layer's flanks from outside. It felt like being inside a fist that hadn't closed yet. Ironbody Stance activated on entry. Rending Fury was pre-charged.

[Skill: Rending Fury — Lv 2] I hit the DPS line.

The line had fourteen players between the second and third layers. Two of them were Lv 33 Warriors acting as secondary DPS; the rest were Mages and Archers in light armor with no tank coverage. In 5.7 seconds I could reach approximately eight meters of line.

I reached eight meters of line.

[System: Rending Fury struck REDPEAK DPS-1 (Mage Lv 31): -3,240 HP (Rend applied).] [System: Crescent Moon Slash struck REDPEAK DPS-2, DPS-3 (3m arc): -2,880, -2,960 HP.] [System: Rending Fury struck REDPEAK DPS-4 (Archer Lv 30): -3,120 HP (Rend applied, critical).]

Three confirmed kills. Two at sub-30 percent HP.

Then something happened that hadn't been in Wenqing's analysis.

The Redpeak guild leader — from his position in the second layer, which was ten meters away — activated a formation-leader skill I hadn't seen before. It rendered as a gold ring expanding from his position at floor level.

*Ding!* [System: REDPEAK BROTHERHOOD — Formation Skill: Unyielding Shield Wall activated. All Redpeak Brotherhood members within 15 meters: DEF +35%, Knockback immunity for 8 seconds.]

I was inside the fifteen-meter radius.

The two Warriors I was about to finish closed their DEF deficit to zero. The archers I'd half-burned through regenerated enough effective HP via the DEF increase that my next hit would no longer kill them in a single strike. The first-layer closure rotation behind me — which I'd planned to outrun on exit — was now immune to knockback, which had been the secondary exit mechanism Wenqing had built into the survival window.

5.7-second gap window. Two seconds used. 3.7 seconds remaining. Exit compromised.

I had a choice. Exit immediately and eat the unplanned damage, which was significant. Or use the remaining 3.7 seconds productively and trust Zhu Yuhan.

I used the remaining 3.7 seconds.

[System: Rending Fury struck REDPEAK DPS-5 (Mage Lv 32): -3,040 HP (Rend applied).] [System: Crescent Moon Slash struck REDPEAK DPS-6, DPS-7: -2,920, -3,010 HP. Both opponents at critical HP.]

Then I exited.

The first-layer closure hit me as I came through. Full contact, both flank tanks landing simultaneous blows while the DEF+35% from Unyielding Shield Wall was still active on them and I was running the Ironbody Stance at reduced percentage because I'd extended my time inside.

[System: REDPEAK TANK-1 struck BLADELESS: -1,840 HP.] [System: REDPEAK TANK-2 struck BLADELESS: -1,760 HP.]

I had 4,120 HP before entry. I had 520 HP after exit.

Zhu Yuhan's correction hit at 0.3 seconds.

[System: YUHAN (Priest) — Greater Restoration: +1,640 HP. Sustained Recovery applied: +160 HP/2s for 8 seconds.]

I was at 2,160 HP and standing.

Old Wolf was at my left. "Still alive," he said.

"Yes."

"That wasn't in the script."

"No."

The match hadn't stopped. The Redpeak guild leader's Unyielding Shield Wall had an eight-second duration. They were at four seconds elapsed. Wanqing and the east point were holding the first layer's left flank. The west point was scrambling because I'd exited at the wrong angle and disrupted the cage timing.

"Phase 2 reset," I said. "West point re-anchor."

"West point re-anchoring," the point commander confirmed.

I had to trust them to re-anchor in 3.5 seconds and I had to do something about the Redpeak DPS line before their two wounded survivors recovered enough to contribute to their formation again.

I didn't go back in.

The decision happened in the fraction of a second between exiting and registering what the match situation actually was. The west point was scrambled. Going back in during an unanchored cage would leave the gap behind me uncontrolled — the cage was the mechanism that kept the entry viable, and without the cage the first layer would simply close it and trap me on the wrong side permanently. The plan had accounted for this contingency even if I'd hoped not to need it. Hold position, buy the re-anchor time, trust the west point to recover.

In the 3.5 seconds that the west point was re-anchoring I pushed the edge of my position to the flank of the first layer — not crossing the gap, but pressuring the layer's left-side tank into a two-direction defense problem. Old Wolf mirrored the pressure on the right. The first layer compressed inward by two meters, which was enough.

When the Shield Wall expired at eight seconds and the west point had re-anchored, the Scattered Fan cage reformed around a formation that was now two meters narrower and four members lighter.

"Phase 2, second pass," I said.

I went back in.

*Ding!* [System: Rending Fury struck REDPEAK DPS-2 (Mage Lv 31, critical HP): elimination.] [System: Crescent Moon Slash struck REDPEAK DPS-6, DPS-7: both eliminated.]

The rest of the match took seven minutes.

It was a different kind of fighting than the gap zone — methodical, patient, the formation work of twelve people who'd drilled the same sequences enough times that the adjustments happened in the moment without announcement. Old Wolf held the re-formed west anchor. Wanqing worked the east pressure with a precision that, from my position, looked almost relaxed. The Redpeak guild leader, Chen Mang, tried two more formation variations in the seven minutes remaining — a compressed-center push and then a split-flank maneuver — and both of them died on contact with a formation that had stopped reacting to what they expected and started doing something they hadn't prepared for.

***

*Ding!* [Continental Committee: SEVERING LIGHT defeats REDPEAK BROTHERHOOD — 12 remaining vs 41 remaining at time limit. Score: 26-0. SEVERING LIGHT advances to Round 4 (Sat Mar 21). Next opponent TBD.]

Not a clean elimination. Twenty-six of their sixty-seven were down. The rest had survived the time limit. On points, we'd won by enough that the match wouldn't be re-disputed, but anyone watching the kill-cam had seen a fight where the smaller guild had nearly had its key player eliminated in the gap zone, improvised a reset under fire, and won by the skin of its teeth.

There was a specific calculation in letting that be visible. A guild that won immaculately showed capability. A guild that nearly lost and recovered showed something different — adaptability under pressure, healer performance in an unplanned scenario, a guild leader willing to extend into a compromised exit because he trusted his team. Those were harder things to counter-plan for than a clean formation win. You could develop a counter to a formation you'd analyzed. You couldn't fully counter a team that improvised well, because the improvisation was the variable.

Wang Jian would see this on the kill-cam. He'd note the near-miss. That was the calculation: the near-miss was information, and the information I wanted him to have was that we recovered, not that we'd been perfect.

I was at 2,160 HP at match-close. I'd gone in with 4,680. The gap-crossing had cost me more than half my health pool, a prepared counter had stripped the exit mechanism, and Zhu Yuhan's 0.4-second margin had been the difference between that being survivable and not. The kill-cam was going to show exactly how close it had been. There was nothing to be done about that. Wang Jian would see it and draw his conclusions. The conclusions would be accurate: we'd won because a healer he didn't know existed had built a safety margin into her pre-positioning that even my own plan hadn't specified.

Wenqing's post-match analysis arrived in thirty seconds: *Formation Skill: Unyielding Shield Wall not in the bracket-history data. First recorded use in competitive play. Source analysis: Redpeak Brotherhood guild leader Chen Mang is a former professional player from the disbanded server-3 top guild Ironspire. Ironspire used a similar formation-leader skill called Iron Bulwark in their 2013 match records — this appears to be a server-1 adaptation. Implication: Chen Mang came into this match with a prepared counter to our Phase 2 gap-crossing. He knew we would run it. He received advance intelligence.*

I read it once. Then again.

"Wang Jian's communication chain," Wanqing said, on the bonded thread.

"Yes."

"He had more than eighteen hours."

"He had the week. Wenqing's analysis was right about the timing — we had eighteen hours of preparation margin after Round 2. But Wang Jian didn't need to prepare a counter in eighteen hours. He prepared it in the week, based on Xu Ming's Round 2 formation-testing."

A pause.

"Then why didn't the counter work."

"Because Chen Mang used it too early. The Shield Wall at second four was designed to eliminate me in the gap. It would have — if Zhu Yuhan's pre-positioning hadn't already accounted for the extended exit window."

Zhu Yuhan, on the guild channel: "The exit window was flagged in my pre-match data. I added 0.4 seconds to the standard correction pre-position. I didn't know why — your entry and exit practice timing was clean. I added the margin because practice timing is never the same as match timing."

Old Wolf: "You expected the extra damage."

"I expected something I hadn't modeled correctly. I didn't know what. I just built in the margin."

A long pause on the guild channel.

Old Wolf: "Good."

There was a specific quality to Old Wolf's *good* in that context. It wasn't approval in the praise sense. It was the word he used when something had demonstrated its own value and the demonstration was complete and didn't require additional comment. He'd said it about Iron Plum's tank rotation after the fourth formation drill. He'd said it about Wenqing's flat terrain analysis. He used it when the evidence had accumulated to a threshold and he'd concluded. Zhu Yuhan had now crossed that threshold in a high-stakes match situation, with a variable she hadn't been given and a margin of her own construction. That was a complete demonstration. Old Wolf's *good* closed the file.

Outside the pod block the Hangzhou morning was cold and clear. The same overcast that had been running since the start of March had finally broken, and the March 14 morning had actual direct light coming from the east — pale, still winter-angled, but present. It hit the pod block's concrete exterior at a low angle and made sharp shadows that the overcast days had been missing. I stood in it for a moment.

I was thinking about what Chen Mang's counter meant for the next arc. He'd come in with a preparation that was almost good enough — almost, because the last 0.4 seconds hadn't been in his model. That wasn't a failure on his part. That was a healer he'd had no data on, doing something that wasn't in any public dataset. His intelligence was good. It just hadn't reached far enough.

The asymmetry that had saved us in Round 3 was the same asymmetry I'd been building since October: the team had members that Wang Jian's intelligence hadn't fully mapped. Zhu Yuhan's solo protocols, Wenqing's private analysis logs, the formation drills in a private instance with access-flagged containment — all of it produced a guild that was more capable at specific moments than the public record suggested. That gap between the public record and the reality was the asset. Wang Jian would work to close it. The work I had to do was to keep generating new gaps faster than he could close the existing ones.

Round 4 was not yet confirmed. The Round 4 opponent and the Round 4 preparation were problems for the week following March 14. Today was March 14, seven minutes past match-close, and the problem was still forming. The formation analysis Wenqing would have ready by noon. The opponent the committee would confirm by afternoon. The preparation would start as soon as those two things arrived and not before. In the meantime I was standing in cold March light with 2,160 HP and a bun from the overnight cart, and that was enough for right now.

Wang Jian would note that. He'd note the gap in his intelligence and he'd work to close it. The next time he sent a counter, it would be more comprehensive. The question was how far the "next time" would be from now.

Somewhere in Wang Jian's operations room a kill-cam was playing of a twelve-member guild that had survived a prepared counter-strategy by a margin thin enough to see through.

He would file it under "no one" and sharpen his pencil.

I was counting on that too.