Chương 84: The snare

The spring-snare lesson started because Marco had been watching him build them for three days and had finally run out of patience.

"Show me," Marco said. Not a question. The tone of someone who had been gathering information through observation and had reached the limit of what observation alone could produce. He showed him. The mechanism was a spring-snare with a balanced trigger: bent sapling under tension, trigger stick set at the pressure point, trip line at the correct height for the target species.

He'd been building them since Day 19, reflexively, efficiently, with the speed of something that had moved from skill to reflex over a hundred and fifty days of repetition. He built them the way he'd learned to build them, which was by doing them wrong until they worked, and then doing the working version until it was automatic. He showed Marco the working version. Marco watched with the focused attention he applied to things he'd decided to learn. Then: "Why that angle." He started to answer.

And then, halfway through the sentence, he stopped. The angle. He'd set the trigger stick at forty-five degrees relative to the spring tension line, which was the correct angle, which he'd been setting it at for a hundred and fifty days, which, when he reached for the reason, produced a notable absence where the reason should have been. He knew the angle was correct.

He'd established it was correct empirically, on Day 22, when he'd built seven snares at varying angles and the forty-five-degree one had been the only one that triggered cleanly. He'd run the test because the first angle he'd tried had failed, and he'd varied from there until something worked, and the forty-five-degree angle had worked, so he'd kept using it. He had never, at any point in this process, examined why the forty-five-degree angle worked. He stood there with this gap for a moment. Marco waited.

"The trigger geometry," he said, and then stopped again, because he was building the explanation in real time now rather than retrieving it, and building it in real time meant actually working out the underlying mechanics rather than packaging something he already knew. The trigger stick at forty-five degrees distributed the trip-line tension across both axes simultaneously, lateral and downward, which meant the threshold force was lower, which meant smaller animals could trigger it reliably, which meant...

he kept going, building it out, and by the end of the explanation he had a model he hadn't had forty-five seconds ago. He explained the model to Marco. Marco listened in the way he listened to things that were mechanically useful, completely and without interjecting. Then he said: "So the angle's not about the spring tension."

"It's about both." Finn paused.

"I thought it was about the spring tension. I was wrong about what I thought. The angle works because of both axes." Marco nodded, twice, slowly, in the way that meant he understood and was integrating.

"Show me the wrong angle." He built a comparison snare at thirty degrees. Then sixty. They tested the trigger threshold on both. The results illustrated the model with a cleanness that made him slightly annoyed at himself for not having run this test a hundred and thirty days ago, or at least having built the explicit model rather than operating on an implicit one he'd never examined. Teaching XP, he supposed, was for discovering this kind of thing. He was going to look at that optimistically.

The notification arrived while he was watching Marco build his first attempt: 【Teaching Session — Methodology】 Subject: spring-snare trigger geometry | Recipients: Marco, Yuki INT: +0. 2 (reasoning behind established practice, first formal articulation) Teaching XP: +45 | Methodology model: externalized (dual-axis trigger geometry) Trap geometry, explained out loud for the first time. Marco is learning the system. Yuki is evaluating the system.

These are different activities, and only one of them is going to produce feedback he finds uncomfortable. The traps, for the record, are catching approximately sixty percent of what they should be catching. The Watchers have been noting this for some time. They elected not to mention it. They are mentioning it now. The timing is not accidental. He read the last line and looked at Yuki.

She was three meters to his left, working on the herbalism kit with her usual quiet attention, which was the quality of attention that suggested she was focused on her own work while also being peripherally aware of everything in the immediate area. He'd come to understand that Yuki's peripheral awareness was actually her primary mode. She just led with the focused part to keep people from feeling surveilled. She was looking at him now.

"The tension anchor," she said.

"You're setting it at the base. Move it to the mid-point of the sapling. You'll get a faster release and a cleaner snap angle." Not a suggestion. The flat delivery of someone who had already run the calculation and found it complete. He looked at the trap. He looked at where the anchor was sitting. He moved it. The next test-pull snapped two full beats faster than his standard configuration. He looked at her.

"You didn't know you knew that," she said. Meaning the trigger geometry, the articulated model, the whole lesson.

"No," he said. She nodded once and went back to the kit. That was the exchange. Six words of trap correction, four words of observation, one on his end, and somehow he was going to be thinking about it for the next several hours. The sixty-percent line from the Watchers was sitting in his head now, doing the math he should have been doing. Sixty percent.

His snares had been catching sixty percent of what they should have been catching, for an indeterminate number of days, and Yuki had apparently known why, and she'd been sitting three meters away working on her herbalism kit, and she'd just corrected it in one sentence, which raised a question he was not sure he wanted the full answer to: what else had she been quietly evaluating? He ran a mental inventory. His route planning. She'd corrected that once, sideways, on Day 141. His biological proximity indicators.

Corrected once, directly. His camp placement. She'd repositioned herself twice in the first week in a way that suggested his original layout had a coverage gap he hadn't identified. His perimeter sweep pattern. She'd started running the watch rotation without being asked, which could mean she'd evaluated the rotation and found it acceptable, or could mean she'd evaluated the rotation and decided the most efficient correction was to just do it herself. Love what you're reading?

Discover and support the author on the platform they originally published on. The list was getting longer than was comfortable. Yuki had been running what amounted to a parallel quality-assurance audit of his entire operational methodology, and the audit had been generating findings, and the findings had been deployed one at a time with the patient precision of someone who understood that delivering all of them at once would have caused a system crash in the recipient. He respected this.

He also found it deeply, specifically unsettling in the way of discovering your work had been peer-reviewed without your knowledge and the reviewer had been kind about it. He'd been managing these exchanges with Yuki for three weeks now, the ones that were short and landed with more weight than their length suggested, and he still hadn't fully calibrated to them. He kept expecting the short ones to be incomplete and finding instead that they were precise. She didn't say more than she meant.

He'd been operating the same way for a hundred and fifty days solo, had considered it a personal efficiency, and was now encountering it in someone else and finding it slightly unnerving in a way he couldn't fully account for. Yuki just corrected my trap geometry and then told me I didn't know I knew something and she was right on both counts, he thought, in the internal register. The group has been teaching me things about what I know by asking me to explain it. I have been treating this as one-directional.

The one-directional assumption was the part worth examining. He'd been running the calculation as: he knew things, the group didn't know things, he transferred the things. Clean, linear, his model exiting his head and entering theirs. This was accurate as far as it went.

It did not account for what happened when the transfer required verbalization, and verbalization required him to build explicit what had previously been implicit, and building explicit produced new information that hadn't been in the implicit version.

It also did not account for Yuki, who was apparently running her own parallel assessment the entire time and had just demonstrated that his traps had been underperforming for a hundred and fifty days, and that she'd been aware of this, and that she'd chosen today to mention it, and that the Watchers had also been aware of this and had also chosen today to mention it, and the coordination between those two timing decisions was either coincidental or evidence that the universe was collaborating with Yuki to manage his ego in stages.

He was going to need a new category. Marco's first snare attempt had the trigger at forty degrees. Close. He corrected it without making the correction a correction, just showed the adjustment and the reason, the way he'd been learning to show things rather than tell them, which was a distinction the teaching had clarified over the last ten days. The forty-five-degree angle. The dual-axis distribution.

Marco rebuilt it at the correct geometry, tested the threshold, and made the sound of someone satisfied with a thing they'd figured out correctly.

"I can set these now," Marco said.

"Yes."

"Add two per camp?" Finn considered the camp footprint and the likely game movement for the terrain they'd be crossing. Small ground game, the thorn-rats and the smaller foragers that moved through the undergrowth at dusk and dawn, approached from upwind as a basic predator-avoidance behavior. They'd smell the fire's smoke and the camp's human scent and avoid the downwind corridor entirely, which meant the viable snare positions were on the approaches where the air was cleanest.

In the current terrain, prevailing wind ran south to north in the evenings, which put the northern approach upwind and the eastern approach in a crosswind corridor that still carried enough clean air to draw traffic.

"Two on the northern approach," Finn said.

"That's where the game comes in. Upwind from the fire, clean air, and the root density up there funnels the ground-foragers through natural channels between the larger trees. One on the eastern side for the crosswind traffic. Set them at dusk, check at first light. Anything past sunrise, the day predators are moving and you're competing for the same catch."

"Got it." Marco picked up the materials and started on the camp placement with the efficiency of a man who had received instructions in a format that matched the way his brain worked and was now simply executing them. Someone at DataVault, a developer named Gina who sat two rows over and had opinions about everything including his lunch, had hiked the Pacific Crest Trail. Six months. Mexico to Canada.

She'd come back with the kind of weathered confidence that only came from voluntary extended suffering and had shown the office approximately four hundred photographs of sunsets that all looked like the same sunset. Finn had listened politely and thought, in the privacy of his own head, that walking through a forest for six months sounded like a thing people did because they hadn't found a sufficiently engaging spreadsheet.

He was now, by any reasonable metric, on a version of the PCT that was worse in every conceivable dimension. There was no trail. There was no resupply schedule. The wildlife was not a scenic feature but an active participant in the question of whether he'd be alive tomorrow.

And he was setting snares for rabbits on a northern approach because the wind direction meant the rabbits couldn't smell the fire, which was a sentence Gina would have recognized as trail life and Finn would have recognized, six months ago, as a sign that something had gone fundamentally wrong with his career trajectory.

The difference between the PCT and this was the difference between choosing to test yourself against nature and being entered into a competition you hadn't applied for, with entry requirements that included being alive and exit requirements that were still under negotiation. He was not going to tell Gina about this if he ever saw her again. She would be insufferable. This was the thing about Marco he kept noticing in new ways.

The man's execution, once he had the correct model, was reliable in a way that produced outcomes. Not elegant, not optimized, but reliable. He set things up and they held. He built things and they functioned. He held positions and they stayed held. There was a coherent capability in there that Finn had been valuing incrementally and probably hadn't yet reached the correct total for. He added this to the running assessment.

The party skills register had been getting more accurate as his valuation methodology improved. He was going to need more pages. At camp that evening he updated the bark-sheet. Two new lines.

The first was the dual-axis trigger model, the actual reasoning behind a practice he'd been running correctly without understanding correctly, which meant his prior execution had been right by the empirical test and not by the underlying logic, which meant there were probably other practices in his methodology in the same condition. He'd been treating worked empirically as understood correctly and these were not the same criterion and he'd just found an example of the gap.

The second line was the correction to his teaching model. Externalization produces new information. Direction of transfer: bidirectional. He looked at this line for a moment. It was technically two lines. He counted it as one entry and did not examine the counting. He read both lines back. They were correct. He was reasonably sure they were correct, anyway.

He'd need to test the second one more carefully, because it was a claim about his own cognitive process, which was the kind of claim that was easy to over-generalize. But the dual-axis model was definitely correct, because he'd built two comparative snares and tested them and the data was unambiguous, and the bidirectional transfer claim was supported by at least three instances he could name. What else is there, he thought, looking at the bark-sheet without reading it.

What else am I doing correctly without knowing why. The list didn't announce itself, but it had the quality of something longer than he wanted it to be. Which was probably the correct length, and probably going to require more conversations with people who asked him to explain things out loud, and probably something he was going to have to be a different kind of okay with than he currently was. He closed the bark-sheet. The fire was Marco's placement, warm, distributed, minimal smoke.

Dez had fallen asleep over his own terrain diagram again, which was a habit Finn had stopped correcting because Dez always woke up fine and sometimes the diagram had new notations in the morning that suggested the dreaming brain was more productive than the waking one, which was a hypothesis he was not going to test on himself. Yuki was on the first watch rotation, moving the perimeter with the quiet efficiency of someone who'd picked up the watch pattern by observation and was now running it without being asked.

He hadn't asked her. He noted that he hadn't asked her, and she'd done it anyway. He noted that the perimeter she was running was slightly different from his standard pattern, tighter on the eastern arc where the crosswind carried sound less reliably. He noted that the adjustment was correct. That was probably another entry.

And probably another conversation he'd need to have with himself about what it meant that Yuki's quality-assurance audit was, entry by entry, making his systems better, and that the appropriate response to this was not defensiveness but something closer to the relief of a person who'd been carrying a weight alone and had just discovered that someone had been quietly redistributing it. He was not going to say any of this out loud.

He was going to note it, internally, and let the bark-sheet catch up when the bark-sheet was ready.