← THE GREY FILE  ·  Mission Briefs
Mission Brief

Three Days in Marseille

A client thought he had a team on him. My job was to find out whether he was right, or just tired.

The call came through a lawyer, which is how the good ones usually come. He had a client — call him the principal — who was convinced he was being followed. Convinced enough to stop sleeping, not convinced enough to do anything sensible about it. The lawyer wanted a sober answer to a single question before anyone spent money on bodyguards and armored glass: is this real, or is he just frightened?

Names changed. Details moved. The lessons are real.

I took the train down. You learn a city better from the ground than from the air, and besides, I wanted three days in REDACTED before I gave anyone a verdict. Frightened men are wrong most of the time. They are also, occasionally, exactly right, and the work is refusing to decide which until the city tells you.

Day one — the baseline

You cannot tell what is abnormal until you know what is normal. So the first day I did not look for watchers at all. I looked for the principal's life.

Pattern-of-life is dull and it is everything. Where he lived, where he worked, the café where he took his morning ████, the route between them, the one bridge he had to cross because the map gave him no other. Static coverage owns the places a man cannot avoid — his door, his office, the only crossing — so I learned those first, the way a hostile observer would. I stood on his street and asked the question they ask: where would I sit to see the door and the cars and not be seen? The good vantage points in any street are finite. An upper window across the way. A café table with the right sightline. A van that never moves. I marked them and moved on.

I did this from a distance, and I did it as a man with somewhere else to be. The cardinal sin on day one is looking like you are surveilling the person you are protecting. I wanted his rhythm, nothing more. By evening I had it: when he left, which way he turned, where he was predictable. He was very predictable. They usually are.

Day two — making the city show its hand

You do not stumble onto a surveillance team. You will never feel it the way the films promise. You detect it on purpose, on a route built for the job, and you do it without ever letting a watcher believe you are looking. That route has a name — the surveillance detection route — and it is not a chase. You are not trying to lose anyone. You are building a sequence of small situations where a follower has to choose between exposing himself and losing you, and then you watch which he chooses.

I ran the principal through one. Gave him an errand with a real reason behind it — the pharmacy, the bank, a parcel to collect — so every leg of the route had an innocent face. Cover for action carries the whole thing. If anyone asks, he was buying aspirin.

The route changed texture deliberately. Walk, then a tram, then walk again. A route that is all one speed in one medium gives a follower an easy ride; one that keeps changing forces him to keep adapting, and adaptation is where amateurs slip. I built it out of corners and crossings, because a turn lets you sweep your eyes back the way you came without ever craning your neck against the flow. I used the rest, too — the side mirror of a parked car, a shop window, the dark glass of the tram. Reflections are the backbone of the whole craft. They cost you nothing in posture. A man looking at a watch in a window is not, to anyone behind him, a man looking behind him.

Then the channels. Marseille gives you these for free if you know where to walk — a long stair, a narrow lane with no parallel route, a stretch where everyone funnels through one opening. In a channel a watcher cannot hang back a block over and cannot melt into a side street. He is on your path or he is not behind you at all. I added a timing stop: the principal sat for a coffee he did not want, in the open, for eleven minutes. A stop forces a decision on anyone following — stop too, conspicuously, or walk past and surrender the eye. Both answers tell you something.

By the end of the day I had two faces I did not like.

What a team looks like, and what a lone man looks like

Here is the part the frightened never get right, and it is the only part that matters.

A real foot team is almost never one person. One follower behind you is a liability — if you turn, he has nowhere to go but past you or into a doorway, and both read wrong. So a competent team spreads the load. One takes the eye, directly behind or across the street. Others walk parallel a block over, sit ahead of the likely path, drift behind as relief. They rotate. The man who had you for four hundred meters peels off and a fresh face picks you up at the next corner. The entire point of the rotation is to deny you the one thing that confirms everything — the same individual, seen three times, in three places that only make sense if he is following you.

A vehicle team runs the same logic faster. A single car cannot hold a tail through traffic without crowding your bumper or losing you at every light, so they leapfrog — one car holds the eye and drops back, another that was running parallel takes over, and they accept losing visual for short stretches because a teammate has you. A lone car that makes your every turn, holds the same two-car gap through six changes of direction, and reappears after you thought it gone is not a team. It is one committed man who has already burned himself.

So the diagnostic is not do I see a suspicious person. Anyone can feel watched. The diagnostic is interest plus repetition plus correlation. Ordinary people do not bend their path around yours. They move toward their own business and never adjust to keep you in view. A watcher's pace changes when yours does. A watcher loiters when you stop and finds a reason to move when you move.

And the standard, the one I carry as a sentence and gave to the lawyer verbatim: once is nothing, twice is coincidence, three times is enemy action.

Day three — the answer

My two faces from day two did not survive the standard.

The first was a man I saw at the tram and again near the bank. Twice. In the same quarter, an hour apart. Coincidence of a small city — he worked there, or shopped there, and his path crossed ours by the geometry of a place where everyone uses the same three streets. The third sighting never came. I watched for it across a change of mode and direction that only a follower would mirror, and he simply was not there. I logged him and discarded him.

The second was a parked car with someone in it on a street where nobody sits in parked cars. That one I took seriously for half a morning. Then it left, on its own errand, and a man got out and bought bread and did not look at the principal's door once. Static coverage that does not watch the thing it is supposedly covering is not coverage. It is a man eating breakfast in his car.

Three days, two SDRs on foot, one short loop by vehicle, and not a single face that appeared three times across the changes I had engineered. No rotation. No leapfrog. No interest that bent around the principal's movement. The electronic layer I could not clear from the street — a tracker, a phone reporting home, none of that shows on a foot run, and I said so plainly in the report. But for human coverage, the city had been given every chance to show a team and had shown me nothing.

So I told the lawyer the truth, soberly, because a sober answer was the whole job: the man was not under hostile surveillance. He was under the weight of his own fear, which is a real condition and deserves a real answer, just not the one he wanted. We swept his car and his phone to close the technical question. Both clean.

He was disappointed. Frightened men often are. A threat you can name is easier to carry than a fear you cannot, and I had taken away the name without taking away the feeling.

Anyone can feel watched. The discipline is telling the coincidence of the city apart from the cost of being followed — and the city pays that cost in repeated faces, three times over, not two.

— M.