Maître Dubois's office on Rue Childebert felt different at four in the afternoon than it had at ten in the morning.
The light came from the west at this hour, flatter and more honest, the kind of light that made things look exactly as they were rather than as they preferred to appear. The old paper smell was the same. The certainty was the same. But something in the quality of her attention when I sat down — something in the way she had cleared her desk of everything except a single legal pad and a pen — told me that this meeting was different from every meeting we had held before it.
She had prepared herself to receive something important.
She looked at me.
"Start at the beginning," she said. "Not the fund. Before the fund. Before Étienne's account." She paused. "The accident."
I had known this was coming from the moment she said bring everything. I had spent the two hours between the phone call and this office deciding exactly how much of the truth I could give her — not the panel, never the panel, that was a door I would not open for any reason — but everything around it. Everything adjacent. Everything she needed to construct a defence without knowing the weapon at the centre of it.
"The explosion," I said. "October. Corridor B, Sasung Lyon facility."
"The power system failure," she said.
"Yes."
"Which Sasung's legal team attributed to a routing malfunction in the facility's primary electrical infrastructure." She paused. "I've read the incident report. The one they filed with the French workplace safety authority — DREAL. I requested a copy three weeks ago."
I went still.
She had been doing her own research. For three weeks, without telling me.
"Why?" I said.
"Because Henri Beaumont is thorough," she said. "And because when a thorough man sends a letter to your client's director, the correct response is to understand everything about the thread he's about to pull before he pulls it." She folded her hands. "The DREAL report is twelve pages. It identifies the cause as a fault in the power routing relay — a component that had been flagged for replacement in the facility's maintenance log fourteen months before the incident."
"Park Seo-yeon's articles documented that," I said.
"Park Seo-yeon's articles documented the executive decision to defer the replacement," she said. "The DREAL report goes one step further." She paused. "The relay that failed was not simply old. It had been modified."
The word landed quietly. The way significant words sometimes did — without drama, simply occupying the space where the previous understanding had been.
"Modified," I said.
"The DREAL inspector noted a non-standard adaptation to the relay's thermal cutoff mechanism. The adaptation was not in the original installation records. It was not in any subsequent maintenance log." She looked at me steadily. "The inspector categorised it as potentially consistent with either an unauthorised upgrade attempt by facility staff — the charitable interpretation — or a deliberate alteration designed to cause the relay to fail under specific load conditions."
The afternoon light came through the window in its flat, honest way.
"Someone tampered with it," I said.
"The report is careful not to say that," she said. "DREAL investigators are engineers, not detectives. They note the anomaly and pass it to the relevant authority for further assessment." A pause. "The relevant authority in this case would be the judiciary. Specifically, a juge d'instruction — an examining magistrate — if someone were to file a formal complaint."
"And has someone?"
"Not yet," she said. "The DREAL report is eleven months old. It has been sitting in a file." She paused. "I found it because I was looking for it. Beaumont will find it because he is thorough." She looked at me. "When he does, his argument changes shape. He is no longer alleging that JE Capital had advance knowledge of a media cycle. He is alleging something considerably more serious."
I worked through the logic. I had been doing this — the patient, sequential architecture of following a problem to its end — since I was fourteen years old. I did it now.
Beaumont represents a family office that lost thirteen million euros on a Sasung long position. He is looking for evidence that JE Capital's short was not based on public information. The DREAL report shows that the Sasung explosion — the event that triggered my Sasung short, indirectly, by enabling Park Seo-yeon's investigation — may not have been an accident.
And the person who had been in Corridor B when it happened was me.
"He'll argue that I knew," I said. "Not that I had a media source. That I had direct knowledge of the sabotage. That I knew the explosion wasn't an accident. That I knew it would lead to an investigation. That I positioned accordingly."
"Yes," she said.
"That's not what happened."
"I know it isn't," she said. "You were eighteen years old doing an unpaid errand for a researcher you had introduced yourself to on the strength of a career event. You had no conceivable motive to be involved in industrial sabotage." She paused. "But James — motive is what a criminal court requires. A civil proceeding is simpler. Beaumont only needs to demonstrate that you had a connection to the event, and that the event was not what it appeared to be, and that you benefited from its consequences. Three facts that are all verifiably true."
"Connection, anomaly, benefit," I said.
"Yes."
I sat with this.
The panel had shown 89% on the Sasung short before I had read Park Seo-yeon's articles. Before the French media had picked up the story. Before any public signal had been available that would explain the trade. I had told myself, in the weeks following, that the panel had simply processed the same public information faster than the market — the Korean articles, the regulatory filings, the pattern in the maintenance budgets.
But the DREAL report had not been public. It had been sitting in a government file for eleven months.
The panel had shown 89%.
I thought about this very carefully.
The panel was not supposed to have access to non-public information. It was a probability engine — I had assumed it operated on market signals, on the same data that professional analysts used, synthesised faster and more completely than any human analyst could manage. That was the framework I had built around it. That was the story I told myself to maintain the distinction between what I was doing and what the AMF called insider trading.
But the DREAL report.
If the panel had known what the DREAL report contained — if it had incorporated information from a non-public government file into its 89% assessment — then the framework I had built was wrong.
Then I did not fully understand what the panel was.
Then the question Chassagne had been asking — how do you know — had an answer I had not anticipated and could not give.
I sat in Maître Dubois's office in the flat afternoon light, and for the first time in nineteen months, I felt genuinely uncertain about the nature of the thing in my eye.
"James," she said.
I looked at her.
"You've gone somewhere," she said. Not accusatory. Observational.
"I'm thinking," I said.
"Think out loud," she said. "Whatever it is, I need to hear it. Not all of it — I understand there are things you will not tell me. But whatever bears on our legal position."
I made a decision.
"The Sasung short," I said. "My conviction on it was higher than my documentation supports. Not because I had a source — I didn't. Not because I knew about the DREAL report — I didn't know it existed until thirty seconds ago." I paused. "But my confidence in the trade came from somewhere I cannot fully explain, and that somewhere is not entirely accounted for by the Korean articles and the maintenance budget analysis."
She looked at me for a long time.
"You had a signal," she said. Carefully. Precisely.
"Yes."
"A signal you cannot disclose."
"Yes."
"And you don't know its source."
I paused.
"I thought I did," I said. "I'm less certain now."
She wrote something on the legal pad. I could not see what it was.
"Is the signal still active?" she said.
"Yes."
"Has it been wrong?"
"Occasionally. Infrequently."
She wrote something else.
"Here is what I need you to understand," she said, setting down her pen. "If Beaumont files a civil claim, JE Capital's trading methodology will be subject to disclosure. Not the specific signal — your signal is proprietary and we can protect it under trade secret provisions up to a point. But the existence of a signal, and its nature — whether it draws on public or non-public information — will become a central question." She paused. "If there is any possibility that your signal incorporates non-public information, even inadvertently, even without your knowledge, then I cannot build a defence that relies on the claim that all your trades were based on public data."
I said nothing.
"Which means," she continued, "that I need you to do something I have never asked any client to do before." She looked at me steadily. "I need you to genuinely investigate your own edge. Not to explain it to me. Not to explain it to the AMF. But to understand it yourself. Because if you are using information you don't know you have, the solution is different from the solution if you are simply very, very good."
The room was very quiet.
This was the problem in its purest form. Not Beaumont. Not the civil claim. Not Étienne's exposure. The root problem, the one that had been there since October of the previous year: I did not understand the panel. I had accepted it, used it, built a structure around it, created documentation to support its outputs — but I had never genuinely interrogated its source.
I had been afraid of what the answer might be.
"Alright," I said.
She nodded. Picked up her pen again.
"For the immediate matter — Beaumont's letter," she said, returning to her precise professional register. "I will respond within seventy-two hours. Formal, through this office, all communications to be directed to counsel. I will request disclosure of the full client identity and the specific nature of the claim before any meeting is considered." She paused. "In the meantime, I need you to do two things."
"Yes."
"First — suspend all trading in Sasung-related securities. Any Sasung entity, any company connected to the Sasung supply chain. Until this is resolved."
"Done," I said.
"Second — I need a written account of your movements and communications in the seventy-two hours before you entered the Sasung Europe short position. Every person you spoke to. Every document you accessed. Every source you read." She looked at me. "Not for me to read, necessarily. For you to have, precisely documented, in case we need it. Memory degrades. Documents don't."
"I'll have it to you by tomorrow morning."
She nodded. Made a final note.
"One more thing," she said.
"Yes."
She set the pen down entirely. The gesture of someone stepping outside the professional conversation for a moment.
"James. The Sasung explosion." She paused. "If someone modified that relay — if the incident was deliberately caused — then you were in that corridor by circumstance. You were there because you had introduced yourself to Dr. Arnaud and offered to carry coffee. A chain of events that began with a career event you attended because you were looking for any connection to the financial world you could find."
"Yes," I said.
"Which means you were not a target," she said. "You were incidental."
"I know."
"But." She paused again. "If Beaumont's client — the family office, the one that lost thirteen million — if they begin investigating the explosion themselves, in the course of building their civil claim, and they find the DREAL report, and they find the relay modification—"
She stopped.
I followed the logic.
If Beaumont's client investigated the explosion, they would find what Maître Dubois had found. The modified relay. The non-standard adaptation. The DREAL inspector's careful, engineer's language around deliberate alteration.
And they would ask: who modified it, and why?
And if the answer to that question, when fully traced, led somewhere unexpected — somewhere that intersected with the family office's own history with Sasung, with their investment thesis, with the reasons they had held a significant long position — then the civil claim against JE Capital would become the least interesting thing in the room.
"They might find something they didn't expect to find," I said.
"Yes," she said. "And that cuts in two directions. It might destroy their claim entirely, if what they find suggests the explosion was engineered by a party connected to their own client." A pause. "Or it might make their claim irrelevant, because the thing they've uncovered is so much larger that a securities dispute becomes noise."
I thought about Park Seo-yeon in Seoul, and her six articles, and the Sasung executives placed on administrative leave, and the Korean parent suspended on the KRX.
I thought about the relay in Corridor B, modified in a way that was not in any maintenance log.
I thought about the eight thousand euros that had arrived in my account within two weeks of the accident, with a full NDA clause buried in the settlement paperwork.
Silence money.
I had thought that in the hospital, on the night the DREAL report did not yet exist in my knowledge. And I had filed it and moved on, because moving on was what I did.
"Isabelle," I said.
She looked at me.
"What if the explosion wasn't meant to cover up negligence?" I said slowly. "What if it was meant to destroy something? Something in the lab. Something in Corridor B or adjacent to it." I paused. "And the cover-up — the NDAs, the compensation payments, the legal preparedness fund — wasn't about hiding maintenance failures. It was about keeping people who were in the building from looking too closely at what they'd seen."
The office was very still.
Maître Dubois looked at me with the expression of a person who has just watched someone arrive, by their own logic, at a conclusion she had reached three days ago and had been deciding whether to share.
"I raised that possibility with a colleague in criminal law last week," she said. "Off the record. Hypothetically."
"And?"
"And she told me that if I had a client who had been in that corridor, I should make absolutely certain that client understood one thing." She paused. "The moment the explosion becomes a crime rather than an accident, everyone who was present becomes either a witness or a suspect." She held my gaze. "And the most dangerous person in that situation is not the one they suspect of the crime. It is the one who subsequently profited from its consequences."
The flat afternoon light lay across the floor between us.
I thought about Corridor B. The forty-one steps I never finished. The ceiling on fire. The warm wall. The scuff mark on the baseboard that I had focused on when my legs stopped working.
I had been incidental. A seventeen-year-old carrying coffee.
I had also made money from what followed.
"What do I do?" I said.
"Right now?" She picked up her pen. "You go home. You write the account of your movements before the Sasung short. You do not discuss any of this with Étienne or with anyone else until I have spoken to my criminal law colleague properly, with names, not hypothetically." She paused. "And James — the investigation you said you needed to conduct. Into your own edge. Into the nature of your signal."
"Yes."
"Do it quickly," she said. "Because if I'm going to defend you against Beaumont, and potentially against something larger, I need to know whether the thing at the centre of this is what you believe it is."
I nodded. Stood. Picked up my jacket.
At the door, I paused.
"The DREAL report," I said. "The relay modification. Has anyone else seen it?"
"It's a public document," she said. "Filed with a government agency. Anyone who requested it could access it."
"But nobody did. For eleven months."
"No," she said. "Nobody did."
I thought about that.
"Beaumont's client," I said. "The family office that held forty million in Sasung Europe. Who are they?"
She paused.
"The disclosure in the commercial court filing was redacted," she said. "I have a name from a secondary source — a contact in the Paris legal community who knows Beaumont's practice." She paused. "The family office is called Groupe Aldric. They are Luxembourg-registered. Their principal is a man named Édouard Aldric." Another pause. "He is seventy-one years old. He is one of the largest private shareholders in the history of the Euronext Paris. And he has not lost a securities dispute in twenty-two years."
I absorbed this.
"Until now," I said.
"He doesn't know it's until now," she said. "He thinks he's found a student fund that got lucky and he intends to recover his money." She looked at me. "He has not yet understood what he's actually found."
I put on my jacket.
"Isabelle," I said. "Thank you."
"Go home, James," she said. "Write the account. And figure out what your signal is."
I walked back through Lyon as the evening came in, the city shifting from its afternoon register to its early evening one — people moving faster, the light going amber, the cafés beginning to fill.
I walked and I thought.
About the relay. About the DREAL report. About the panel showing 89% before I had any public reason to believe in that number.
About what Maître Dubois had not quite said: that if the panel was drawing on non-public information, then I had a problem that went beyond Beaumont and beyond the AMF and beyond anything a legal structure could contain.
And about Édouard Aldric. Seventy-one. Twenty-two years without a loss. Luxembourg-registered. Forty million in Sasung Europe.
A man who had bet that much on one company was not simply an investor. He was a believer. And believers didn't take thirty-three percent losses philosophically. They took them personally.
I stopped at the corner of Rue de la République.
Took out my phone.
Opened the notes application and began to type. The account of my movements in the seventy-two hours before the Sasung short, precisely, as Maître Dubois had asked. Every source. Every document. Every contact.
I typed for eleven minutes, standing on the corner while the city moved around me.
Then I stopped.
Because I had just typed, in the sequence of my research on the evening before I entered the short position, a detail I had forgotten until this moment.
A detail so small I had not registered it at the time as significant.
The evening before I entered the Sasung short, I had been searching for background on the Sasung Lyon facility — its history, its function, what it actually did. I had found a corporate profile page, now deleted from Sasung's website but cached in an archive service, that described the Lyon facility as Sasung's European computational research division, specialising in advanced predictive systems development.
Predictive systems.
I had read that and moved on to the maintenance budget data and not thought about it again.
But the panel. The thing in my eye. The probability engine.
Advanced predictive systems development.
I stood on the corner of Rue de la République with the city going amber around me and felt the ground shift slightly under everything I had believed about my own story for nineteen months.
The panel had appeared after an explosion in a facility that specialised in building predictive systems.
The panel was a predictive system.
The explosion had not been an accident.
I thought about the three things Maître Dubois said Beaumont would try to establish: connection, anomaly, benefit.
I had all three.
What I had not understood, until this moment, was that I was the anomaly.
Not the trade. Not the timing. Not the win rate.
Me.
Whatever had been in that facility — whatever had been destroyed or damaged or released in the explosion — some part of it had not been destroyed.
Some part of it was in my eye.
End of Chapter 9
The Probability of Wealth — By J. Valmont
[Next Chapter → Chapter 10: "Predictive"]James makes one search. One night of research into Sasung's Lyon facility that changes the shape of everything. And Étienne, alone in his apartment, opens the door to a visitor he did not invite.