Marn had rejected six reconnaissance exercises before choosing the seventh. There had been nothing wrong with the first six. Covert observation, route reconnaissance, target acquisition, counter-surveillance, hostile tracking and sensor-denied navigation were all perfectly respectable ways for a reconnaissance specialist to spend a morning. The seventh, however, was labeled **Adversarial Prediction: Irregular Doctrine**, with a description explaining that opposing-force behavior could deviate from established tactical logic.

Marn read it twice before selecting it. There were easier ways to spend a morning, but easy wasn't why he'd chosen this particular exercise.

The simulator changed around him while he checked the reconnaissance equipment secured across his harness. His tricorder rode close to his left side, configured primarily for passive collection rather than the broader active sweeps that could announce his presence to anyone equipped to notice them. Two monitoring beacons occupied narrow slots along his belt, a compact optical sensor sat in another pouch, and the last piece was the recon probe.

Marn pulled the probe free and turned it once in his hand. It was scarcely larger than his palm, built around a compact antigravity system with recessed sensor apertures and a deliberately unremarkable exterior. Optical, thermal, electromagnetic and life-sign detection were tied directly into his tricorder, allowing it to follow terrain, hold an observation point or sit somewhere Marn had no intention of putting his own head.

Useful equipment, but still only equipment. A probe could be detected, jammed or deceived, and perfectly accurate sensor data could still produce a completely wrong conclusion if the person looking at it misunderstood what it meant. Marn trusted the little machine exactly as much as it deserved.

By the time he returned it to its pouch, the smooth walls of Deck Eighteen had disappeared beneath rock, scrub vegetation and a dull gray sky. A shallow valley stretched below him, divided by low ridgelines and the remains of an abandoned colonial settlement. Most of the buildings were little more than shells with missing roofs and partially collapsed walls.

Marn crouched behind a rise overlooking the valley and did nothing for several minutes. That was the first part of reconnaissance people tended to misunderstand. It did not begin when you started moving. It began when you stopped long enough for the terrain to tell you something.

The tricorder remained secured at his side and the probe stayed in its pouch while he watched the valley. A neglected road ran through the settlement. The eastern side rose sharply enough to offer concealment but poor movement. West was lower, broken by ruined structures and the remains of an irrigation trench. The northern ridge offered the cleanest approach to the objective and the best elevation over the route Bulwark would supposedly have to use.

Marn already disliked the north, not because anything was wrong with it but because too much was right.

"Computer, mission brief" he said quietly.

The computer informed him that Bulwark would cross the valley from south to north and recover intelligence from structure seven. Hostile strength was estimated at six to ten personnel with light energy weapons, portable sensors and limited communications capability. When Marn asked about known doctrine, the computer explained that the opposing force historically favored elevated firing positions, overlapping fields of fire and protected withdrawal routes.

That made the northern ridge even more obvious.

There had been no recent contact, hostile sensor coverage was unknown, and no civilians were expected in the area. Marn lingered on that last word for a moment before telling the computer to begin.

From his current position, the northern ridge would have been his first choice if someone handed him the opposing force and told him to build the ambush himself. It dominated the approach to structure seven, provided good cover and offered several withdrawal routes.

It was the correct position, which was precisely what bothered him.

Only after several more minutes of visual observation did Marn bring out the tricorder. The first returns were inconclusive. Residual power signatures remained scattered through the settlement, mineral interference degraded several areas, and a weak thermal anomaly near the northern slope could have been equipment, an animal or the simulator deliberately giving him something obvious to chase.

Two hundred meters farther west, the remains of one building showed recent disturbance along the outer wall. Loose material had been displaced and one section of masonry was noticeably cleaner than the surfaces around it.

Marn marked both locations before pulling the recon probe from his harness.

"Passive profile. Terrain follow. Maximum altitude twenty meters" he instructed while linking it to the tricorder.

The probe lifted almost silently from his palm and dropped below the ridge rather than climbing over it. Marn sent it toward the northern slope first, where it found disturbed soil near one shelf, residual heat beneath another outcrop and a weak power signature consistent with a portable sensor. Farther along the probable withdrawal route, it found more signs of recent movement.