Durrik Tells a Story

I was three hours into the replacement briefing package when Durrik came back.

He didn't announce himself. He walked into the equipment bay the way he walked everywhere, which was like a man who had decided where he was going before he left his previous location and didn't need to adjust once he arrived. He pulled a crate from the stack near the secondary entrance, set it upright, and sat on it. He didn't look at what I was doing.

I kept working. The next notation section covered the rail line's elevated approach from the service tunnel junction, and I needed to get the firing arc coordinates right before I lost the spatial picture I'd been building in my head for the last twenty minutes.

Durrik was quiet for long enough that I finished the section and submitted it. Then he said, "You know a Chethari named Ossver."

It wasn't a question, and it wasn't his usual opening. His usual opening was a direct assessment of something specific. I saved the current file and turned to look at him.

"No," I said.

"Right. You wouldn't." He had his field recorder on his knee but he wasn't looking at it. He was looking at the middle distance, the same place he looked when he was running through something in his head before he said it out loud. "Chethari soldier, about eight years back. I was running the advanced tactical assessment program at Kethavar Station."

I waited.

"Best scores I'd seen in the advanced program in four years," he said. "Spatial reasoning, decision speed under fire, stress response calibration. He had a read on terrain that most soldiers take six years of combat assignments to develop and he had it in the assessment environment on simulated data." He turned the field recorder over once in his hands. "He also had an opinion about senior officer judgment that he didn't keep very well contained."

The equipment array ran its four-minute self-diagnostic. The cycling sound came and went.

"His unit ran into a collapsing position in the Verath Perimeter sector," Durrik said. "Standard Dominion push, larger than the tactical assessment predicted, and the senior officer called a fallback. Ordered it. Clean order, properly communicated." He set the recorder on his knee again. "Ossver assessed the fallback route as less viable than a secondary hold position about sixty meters to the northeast. He was probably right. The terrain data supported his read."

"He didn't fall back," I said.

"He didn't fall back. He took two soldiers with him to the secondary hold position, which was his call and theirs, and they held it for about forty minutes while the rest of the unit executed the fallback. The senior officer's assessment about the fallback route turned out to be partially wrong, so the unit's extraction took longer than planned. Ossver's position held for thirty-eight of the forty minutes. Then a Dominion gifted soldier found a line on the hold that Ossver's terrain read hadn't accounted for, because the gifted soldier's movement capabilities put them in a position the terrain data said shouldn't have been viable."

I said nothing.

"Ossver was twenty-nine," Durrik said. "The two soldiers with him were twenty-six and thirty-one. The senior officer's fallback route, even with the extraction delay, got the rest of the unit out." He looked at the recorder. "Ossver's assessment was good. His read on the terrain was accurate. His judgment about when his read was better than the senior officer's was the thing that killed him and the two people who trusted his judgment over the order."

The bay was quiet. Outside, the perimeter walker made his turn at the eastern corner, boots steady on the approach path.

"I'm not saying you had a better option tonight," Durrik said. "The contract structure is the contract structure and Vasht negotiated it knowing what it was. I'm saying that being right about a situation isn't the same thing as being right about when your read of the situation supersedes the collective judgment of people who have context you don't." He stood up and picked up the crate and set it back on the stack with the same deliberate economy he'd used to take it down. "Ossver was technically brilliant and tactically stubborn and he died at thirty because he knew the right answer and couldn't tell the difference between knowing the right answer and knowing better than the order."

He walked to the secondary entrance and stopped there without turning around. "The eastern choke point notation. The rail line section."

"Four submissions left," I said.

"Good." He walked out.

I sat with the display off for about a minute.

The thing about Durrik was that he didn't tell stories. In the six weeks I'd been working with him across the assessment program and the training curriculum and the station's tactical review sessions, he had communicated everything he needed to communicate through direct assessment and calibration adjustments and one-sentence observations delivered without ceremony. He didn't illustrate. He didn't use examples from his own experience to make a point.

Ossver cost him something to say. I knew that without being able to explain exactly how I knew it, the same way you know when a man is carrying something in a specific register versus just talking. The story was the most personal thing he'd said to me since we met on the training deck, and he'd delivered it the way he delivered everything, which was with the professional flatness of a man who didn't put decoration on the things that were already heavy enough on their own.

I understood what he was telling me. I understood it clearly enough that I didn't need to turn it over and examine it from multiple angles, which was how I handled most things. He wasn't telling me to obey orders I hadn't agreed to. He was telling me that there was a specific failure mode in people who were technically right, and that the failure mode wasn't about being wrong on the object level, it was about miscalibrating the gap between their own situational read and what they didn't know they didn't know.

I knew Durrik had been working with Chethari soldiers for long enough that a specific one dying eight years ago was still specific enough to tell as a story. That was its own information.

I pulled up the next notation file, the one covering the rail line approach from the service tunnel junction, and I looked at it for a moment without reading it. Then I read it and got back to work.

The remaining sections took another two hours. By the time I submitted the final notation and the package was complete in the replacement briefing queue, the station's interior lighting had shifted to its low-cycle setting, which was the facility's version of evening. The perimeter rotation had changed twice.

I stretched, got a canteen from the supply shelf, and drank about half of it. My hands felt fine. The integration's background hum sat in the bone behind my right ear at its resting frequency, which I'd stopped consciously registering weeks ago except at moments like this one, when the room was quiet enough that there wasn't much else to hear.

The Aeluun interface modification that Sonn had been suggesting for two weeks was sitting in my review queue, unopened.

I opened it.

Sonn had documented the proposal carefully, which was how Sonn documented everything: in precise, layered notation that communicated the important information first and the qualifications second, so you could read the first three lines and know the substance and then read the rest if you wanted the mechanism. The substance was: a calibration adjustment to the integration's pattern recognition signal could expand its effective range by approximately forty percent, based on the current development state's baseline. The mechanism was a targeted resonance modification to the interface junction points, adjustable in the field with the right tools, which Sonn had already sourced and placed in my equipment kit two weeks ago on the grounds that I might decide I wanted them eventually.

That was very Sonn. The tools had been in my kit for two weeks, catalogued in the secondary compartment under "calibration support," and I hadn't asked about them because I hadn't been ready to ask about them.

The qualification was the one Sonn had flagged in the proposal's fourth line, which I'd known was coming: the calibration adjustment would improve pattern recognition range, but it would also increase the integration's active output signature by a measurable margin. At range, a sensor package looking for Aeluun integration signals would detect me approximately thirty percent more easily than the current configuration allowed.

More capability, more visibility. Better reach, louder signal.

I thought about the fourteen-day review period and what it was there for. The review period was specifically designed around Sonn's cascade risk concern, the one from Chapter ten's 0200 station conversation: high-intensity repeated stress over a compressed period was the triggering condition for deep resonance cascade. The review period was a brake. It was there to make sure the development wasn't accelerating faster than the system could stabilize.

The calibration adjustment wasn't an integration advancement. It was a tuning modification, not a progression step. Sonn had been careful to document that distinction in the proposal's second section, with the kind of precision that suggested they'd known I'd need the distinction to be clear before I'd be willing to look at the proposal seriously.

The range expansion would be real. The signature increase would be real. The question was whether the tactical calculus, given where we were heading, supported accepting the signature increase in exchange for the range improvement.

I thought about the probing attacks. The way the contacts had been reading as information-gathering rather than seizure attempts. The way someone on the Dominion side was treating each engagement as a data collection exercise.

Better pattern recognition range would let me detect that kind of calibrated, sensor-aware approach further out. It would also make my own signal more visible to whoever was running those approaches.

I went to my equipment kit and opened the secondary compartment. The calibration tools were exactly where Sonn had put them: a resonance adjustment stylus, two frequency calibration chips the size of my thumbnail, and a small coupling wrench with markings in Thuvari that I'd had Sonn translate into a working order of operations three weeks ago. The order of operations was written on a card in my handwriting, tucked under the coupling wrench.

I read through the order of operations twice. It wasn't complicated. It required precise placement of the calibration chips at two of the interface junction points on the device's exterior surface, a frequency adjustment using the stylus that had to be completed within about ninety seconds of chip placement to prevent the calibration from locking in at the wrong resonance, and then a settling period of about twenty minutes during which the interface should not be subjected to active-use load.

Twenty minutes of no active use. That was the constraint. Twenty minutes sitting still in an equipment bay while the system recalibrated, which was fine for tonight and would be a problem during any engagement where I needed the modification to already be in place.

I set the tools out on the equipment table in order of use. Then I sat for a moment and made sure I'd actually decided this rather than just arrived here through a series of steps that felt like deciding.

The range improvement would matter in the engagements that were coming. The signature increase was a real cost. Both things were true. Sonn had assessed it as a net positive given current tactical conditions, and Sonn's assessments of the integration's state had been accurate enough that I trusted the methodology even when I pushed back on the conclusions.

Durrik's story wasn't about this specifically. But the part of it that sat with me was the part about context you don't have, and the fact was that Sonn had more context on integration calibration than I did by a significant margin, and the proposal had been sitting in my queue for two weeks because I'd been delaying a decision that I'd been treating as complex when the actual block was that accepting the modification meant accepting that someone else's read on the right tradeoff was better than my instinct to leave the configuration alone.

That was worth noticing.

I picked up the first calibration chip and checked its contact surface, which was clean and aligned correctly. I placed it at the junction point Sonn's documentation indicated as primary. Then I placed the second chip at the secondary junction. I picked up the stylus.

The order of operations card said: set resonance to 4.7 adjusted units from current baseline. Rotate clockwise. Hold for five seconds. Remove chips in reverse order.

I set the resonance. The interface's background hum shifted fractionally in the bone behind my right ear, a sliding adjustment that settled into a frequency I could feel was different from the previous baseline before I'd consciously processed the difference. It didn't hurt. It felt like tuning a radio with your hand already on the dial.

Five seconds.

I removed the chips in the correct order, set the stylus down, and looked at the tools on the table.

The settling period was twenty minutes. The integration was running its recalibration, which I couldn't feel directly but which Sonn had said would show up in the readings as a slight fluctuation before it resolved. I had nothing to do except wait and not put active-use load on the system.

I put the tools back in the secondary compartment in the order I'd taken them out, which wasn't required but made sense. The coupling wrench, the calibration chips in the small protective case, the stylus. The order of operations card went back under the coupling wrench. I closed the compartment.

Outside, the perimeter walker was at the northern corner. His footsteps were the steady, measured cadence of someone who'd learned the route by repetition rather than by counting steps, which meant he'd been on enough rotations to internalize the timing. That was the kind of thing you noticed more readily than you used to, with the pattern recognition reading gait and interval as a matter of course.

The integration's hum had a slightly different texture now, richer at the upper end of the frequency range I caught through bone conduction. Not louder. Different. I didn't know yet whether the range improvement was something I'd be able to detect subjectively or whether I'd only notice it operationally, the way you don't notice your vision has improved until you try to read something at distance.

Seventeen minutes left in the settling period.

I thought about Ossver, who'd been twenty-nine and technically brilliant and dead because he'd been right about the terrain and wrong about the weight that gave him.

I thought about Vasht, who'd spent political capital she couldn't recover to keep me in the field and who was now going to have to make a case to people with legal staff in the room while her primary argument was an asset who'd just refused a direct Command request on record.

I thought about what I'd said to her. I'm not going to agree to conditions I can't walk away from. Which was true. And I thought about what she'd said back. You may not be right about your judgment being better than the collective analysis of people with more experience in integration development.

Also true. Both of them, at the same time.

The calibration modification was a data point in that calculation. Sonn's assessment was that it was the right tradeoff for current conditions. I'd delayed the decision for two weeks on instinct and then made it tonight, sitting in an equipment bay at low-cycle lighting after Durrik told me a story about a Chethari soldier who'd died at thirty. I wasn't sure the timing made the decision better or worse. I was sure the decision was right on the object level, which was the thing I could actually evaluate.

What I didn't know, and couldn't currently know, was whether I'd delayed it for two weeks because I was correctly calibrating the risk or because I didn't like accepting that someone else's read on the tradeoff was better than mine. Sonn had been careful and patient with the proposal, which was Sonn's way of giving me space to arrive at it without pressure, and I'd used that space the way I used most space, which was to hold the decision at arm's length while I finished other things.

Probably both. That was usually the answer when the choice was between two explanations that both fit the evidence.

Eleven minutes.

I got up and refilled the canteen and drank the second half of it, which I'd been meaning to do for the last hour. My back had the specific stiffness that came from sitting in the equipment bay's folding chair for four hours, which wasn't built for someone my height and didn't pretend to be. I walked to the observation window at the bay's south end and looked out at the sector's exterior, which at low-cycle lighting showed the grey-brown of the industrial terrain and the lit lines of the approach paths and the dark bulk of the rail infrastructure cutting across the mid-distance.

The rail line was the choke problem. The elevated approach it provided was the coverage gap in the standard defensive geometry, the one I'd spent the last four hours documenting for the replacement soldiers. Looking at the actual structure from this angle, I could see the shadow it cast on the ground approach, which was exactly the shadow I'd been trying to describe in coordinate notation for the past hour. It would have been easier to just bring the replacement soldiers to this window and point at it.

That wasn't an option, which was why the notation existed. But it was one of those gaps between what documentation could do and what direct experience could do, and it was always going to be there.

Seven minutes.

I was thinking about the gap between documentation and experience when it occurred to me that Durrik's story was partly about that specific gap. Ossver had the documentation, had the terrain data, had the assessment scores. What he didn't have was eleven years of watching exactly what happens to technically brilliant soldiers who mistake their read of the terrain for the totality of the relevant variables.

Durrik had those eleven years. The story was him handing over the thing the eleven years had cost him, in the most efficient form available, which was a specific instance with specific outcomes rather than a general principle.

I'd received it. I wasn't going to pretend the receiving meant I'd immediately recalibrated every future decision accordingly, because that wasn't how it worked. But I'd received it, and that was something.

Five minutes.

The observation window showed the approach path's northern section, and the perimeter walker was visible at the far end of his route, the one who'd internalized the timing. He moved with the calm of someone who'd done this enough times that the alertness had gone from conscious to structural, which was different from relaxed. The alertness was still there. It had just moved to a different level.

The integration's hum in my right ear had settled. The recalibration texture I'd noticed earlier was gone, replaced by the baseline frequency running at its normal register, except that the normal register was now the modified register, which would be indistinguishable from normal in another twenty-four hours.

Two minutes.

I went back to the equipment table and ran a brief passive diagnostic through the interface the way I'd run a function check on a rifle after a maintenance session, checking that the system responded normally to passive query and that nothing in the recalibration had created an unexpected artifact in the junction response. The readings came back clean. The junction points were calibrated correctly. The resonance was sitting where the order of operations card said it should sit.

Good.

I'd need to contact Sonn tomorrow morning and let them know the modification was done, so they could run a proper reading and confirm the calibration from their end. Sonn would probably have questions about why I'd done it tonight instead of flagging it for review, and the honest answer was that Durrik had told me a story and I'd been sitting in a quiet bay with nothing left to do on the replacement briefing package and the tools had been in my kit for two weeks.

That was the honest answer. Sonn would understand it, I thought, even if they logged it differently in the integration record.

The settling period was done. The modification was in.

I sat down in the folding chair with the display off and the equipment bay quiet, and the integration ran its modified frequency in the bone behind my right ear, and outside the perimeter walker made his turn and started back on the southern route. The replacement briefing package was complete and sitting in the queue. The defensive geometry was documented. The dead angle at the north corridor position was corrected and marked for priority reading.

I'd done what I could do tonight. Most of it, anyway.

There was one more thing I'd been putting off. I pulled up the comm interface and drafted a message to Vasht, short and specific: I was aware that the Command review board situation was going to require something from me beyond the current posture. I was willing to have a conversation about what that looked like before Palleth called back with legal staff on the line. Not a concession on core conditions, but a conversation about what the Compact could have from this arrangement that they weren't currently getting.

I read it back. It was accurate. It didn't promise anything I couldn't deliver and it didn't give away anything that mattered.

I sent it and shut the display down.

The equipment bay had the particular silence of a space that had been occupied for hours and was now preparing to be empty again. The integration ran steady. The perimeter walker's footsteps marked the rotation's eastern corner.

Tomorrow Sonn would need to see the calibration readings. Tomorrow the replacement soldiers would start working through the briefing package, and the ones with sense would ask questions about the northern choke point's dead angle. Tomorrow the replacement briefing process would begin moving from documentation to direct terrain experience, and there would be gaps between what the documentation said and what the terrain did, and those gaps were what the next two weeks were for.

Tonight I'd done what I came to the equipment bay to do, which was finish the notation work, and I'd done something else besides, which was make a decision I'd been delaying for reasons I was still sorting out.

I left the bay to the self-diagnostic's four-minute cycling and walked back toward quarters, and the integration ran its new baseline frequency all the way there.