August in Paris was the city at its most honest.
The tourists arrived in force, which the city accepted with the weary indifference of someone who had been hosting uninvited guests for centuries and had made a functional peace with the arrangement. The Parisians who could leave did. The ones who stayed inhabited the city differently — slower, more present, as though the departure of half the population had restored some original quality of the place that normal months obscured.
I stayed.
I stayed because the work did not take August off. Because the Ashworth window was running regardless of the calendar. Because Catherine Pryce was in London building a presentation that needed input every three days. Because the Amsterdam position was moving — slowly, correctly, the supply chain thesis developing exactly as the combined analysis had predicted — and the panel was showing 81% now, up three points from July.
I stayed also because Margaux stayed.
Her Sciences Po programme had a summer placement requirement — she was doing hers at a financial consultancy in the 9th arrondissement, three days a week, which left the other four for JE Capital's research function and the apartment on Rue du Bac and the Tuesday evenings that had become as fixed in the weekly structure as the boulangerie at eight and the call with Hugo at eight-thirty.
She had stopped saying she was coming over.
She just came.
I had stopped noting the second coffee cup.
It was simply there.
The packaging took three weeks and four days. Not four weeks — Catherine Pryce did not use the full allocation when the work could be completed in less.
What she built was forty-one pages. Not the same forty-one pages as Margaux's Q1 analysis — different register entirely. Margaux wrote for internal use, for the decision-making function, in the precise language of someone who understood both the analytical process and the person who ran it. Catherine wrote for external use, for British institutional due diligence committees, in the language of someone who had sat in those committees for twenty years and knew exactly what the people on the other side of the table needed to see before they could say yes.
The document had three sections.
Section one: performance record. Three years of audited returns, presented against the relevant benchmarks — CAC 40, Euronext pan-European, and a custom peer group of European equity funds of comparable vintage. The comparison was not flattering to the peers. It was not designed to be.
Section two: investment process. Not the real investment process — the presentable version. The documented thesis methodology. The research framework that Margaux had mapped in the Q1 analysis. The pre-notification protocol with Marchand Capital described as a structured information exchange with a complementary European investment vehicle, which was accurate without being complete.
Section three: team and structure. Smith Capital Group SA, Luxembourg. JE Capital SAS, French entity. The Berenger LP as an institutional reference. Hargreave Partners as a British institutional relationship. Maître Dubois as legal counsel. Gérard Fontaine as certified accountant.
On paper, constructed this way, it looked like something a British pension committee could recommend without embarrassment.
I read the final version on a Wednesday morning and called Catherine.
"The peer group in section one," I said. "You've included three funds that underperformed more severely than average. The comparison makes us look better than a median peer comparison would."
A pause.
"Yes," she said.
"Is the peer group selection defensible?" I said.
"All three are funds of comparable vintage and strategy focus," she said. "The selection is technically accurate. It is also favourable." She paused. "I wanted your view on whether favourable is acceptable."
I thought about Chassagne's document on page sixty-seven. About the distinction between what was true and what was complete. About the founding condition of every relationship I had built — transparency about what I knew and deliberate silence about what I could not say.
"Change it to the median peer group," I said. "Our returns are strong enough without the selection bias. And the first time a due diligence committee does their own peer analysis and finds a different comparison, we've lost them."
A silence.
"You're right," she said. "I knew you'd say that. I wanted to see if you would."
I looked at the phone.
"You tested me," I said.
"I test everyone," she said, without apology. "It's how I know who I'm working with." She paused. "You passed. Obviously." Another pause. "The revised version will be with you by Friday."
She ended the call.
I sat with that for a moment. Then I wrote in the notebook: Pryce tests. Good. Median peer group — correct call.
Margaux found the priority change on a Thursday.
She was at the desk — her desk now, in the mornings — with the Ashworth client list that Catherine had compiled and the performance records Catherine had accessed through her network of former colleagues who remained, in various capacities, adjacent to the information she needed.
I was on the sofa reading the Amsterdam company's latest filing — they had released an interim update that confirmed the supply chain thesis in three specific line items.
She said: "James."
Not James, look at this — just the name, in the register she used when she had found something that required immediate attention.
I looked up.
She turned the laptop to face me.
"The Ashworth client list," she said. "Catherine ranked them by estimated unhappiness and mobility — how likely each is to move when the FCA findings drop." She paused. "I've been cross-referencing against their current allocations and their publicly stated investment mandates." She paused. "The third client on Catherine's priority list — the Durham University endowment — I've moved it to first."
"Why?" I said.
"Because their investment mandate was revised fourteen months ago," she said. "The revision expanded their alternative allocations significantly and explicitly included European equity funds of certain vintage and return characteristics." She paused. "JE Capital meets every criterion in the revised mandate. Which Catherine's client ranking did not account for because the mandate revision was disclosed in a supplementary committee minute that was publicly filed but not indexed by any of the standard databases."
"Where did you find it?" I said.
"Durham University's governance website," she said. "Committee minutes archive. Page forty-three of the November meeting from fourteen months ago."
I looked at her.
"You read committee minutes on the Durham University governance website," I said.
"I read everything that's publicly available," she said, without inflection.
I held her gaze for a moment.
She looked back at me with the expression I now had a complete name for. Steadfast attention. The expression of someone looking at something they have decided they are not going to stop looking at.
"Send it to Catherine," I said. "She should know Durham moves to first."
"Already sent," she said. "Three minutes ago."
She turned back to the screen.
I returned to the Amsterdam filing.
The apartment was quiet in the specific way of two people working in the same room who have been working in the same room long enough that the quiet had its own texture — not absence of noise but presence of concentration.
Outside, August Paris went about its honest self.
My phone buzzed.
Catherine: Durham first. Margaux is extraordinary. How did she find that?
I replied: She reads everything that's publicly available.
Catherine's reply: So do I. I missed it.
Then, after a moment: Tell her from me.
I looked up.
"Catherine says you're extraordinary," I said.
Margaux did not look up from the screen.
"Tell her I read committee minutes for fun," she said.
I typed this to Catherine.
Catherine's reply: I believe her. That's the most alarming thing anyone has said to me this week.
The panel's 47.2 million figure arrived on a Tuesday in the third week of August.
Not from the positions I had been watching. Not from Rheinmetall, which was up nineteen percent and closing toward the thesis target. Not from Amsterdam, which was building correctly. Not from any trade I had executed or anticipated.
From an email.
The email arrived at eleven forty-three on a Tuesday morning, from an address I did not recognise, with a subject line that read: Re: Smith Capital Group SA — Allocation Enquiry.
The sender was a man named Robert Ashby. His email signature identified him as the Chief Investment Officer of the Pemberton Family Office, Edinburgh. Assets under management: not disclosed, but the domain of the email resolved to a family office registered in Scotland with a public filing history that I pulled in four minutes — and the filing history showed a total asset base of approximately two hundred and thirty million euros.
He had found us.
Not through Catherine. Not through Hargreave. Through the Luxembourg registration of Smith Capital Group SA, which was a public filing, and through the performance data attached to the Berenger LP reference, which Petra Novák's team had apparently discussed at a Geneva conference in July.
He had done his own research.
His email was six lines.
Dear Mr. Smith,
I have been aware of JE Capital's performance record for approximately eight months. I was directed to Smith Capital Group's Luxembourg registration last month and have since reviewed your audited returns and your Berenger Capital LP relationship.
I would like to discuss a potential allocation. Our family office has a mandate that your returns profile appears to match.
I am in London next week. Are you available Wednesday?
Robert AshbyCIO, Pemberton Family Office
I read it three times.
Then I sat back and thought about the panel's 47.2 million.
JE Capital's current AUM was forty million.
A Pemberton allocation — even a conservative one, sized at five percent of their total assets — would be eleven point five million.
Forty million plus eleven point five million was fifty-one point five million.
Not forty-seven point two.
But a conservative allocation was not the only possible allocation. A ten percent mandate — which their filing history suggested was within their typical single-manager exposure — would be twenty-three million.
Forty million plus seven point two million was forty-seven point two.
Seven point two million was thirty-one percent of twenty-three million.
The panel had not calculated the full allocation. It had calculated the initial deployment — the first tranche, while they assessed the relationship before committing the remainder.
I looked at the panel.
Blank, patient, waiting.
I had circled 47,200,000 in the notebook three weeks ago and put a question mark beside it.
The question mark had an answer.
I picked up my phone. Called Catherine.
"Robert Ashby," I said. "Pemberton Family Office, Edinburgh."
A pause.
"I know him," she said immediately. "Not well — I've met him twice. He ran the investment function at a Scottish pension fund for twelve years before moving to the family office two years ago." She paused. "He is meticulous. Slow to commit. When he commits, he commits significantly." She paused. "James. How did he find you?"
"Geneva conference in July," I said. "Berenger's team apparently discussed us."
"Petra Novák," Catherine said. "She mentioned you at the conference."
"Yes," I said.
"That was not accidental," Catherine said. "Petra does not mention things accidentally."
I thought about this.
About the ten-million allocation increase Petra had approved in March. About the good hunting message. About a forty-two-billion-euro fund's head of European equities who had decided, apparently, that making JE Capital visible in the right rooms was part of how she managed her own portfolio.
She was protecting her allocation by growing the manager.
That was institutional thinking at its most elegant.
"Wednesday in London," I said. "Can you be there?"
"I can be there," Catherine said. "But, James — let me say something."
"Tell me," I said.
"Robert Ashby will not be impressed by you arriving with support," she said. "He will be more impressed by you arriving alone." She paused. "He is the kind of man who allocates to managers, not to organisations. He wants to sit across from the person who makes the decisions and understand them." She paused. "I should be in the building. Not in the room."
I thought about this.
She was right.
"In the building," I said.
"In the building," she confirmed. "Bookable meeting room. Two floors up." She paused. "If you need me, text. I'll be there in four minutes."
"Thank you," I said.
"James," she said.
"Yes."
"The packaging," she said. "The forty-one pages. He won't have read it — he found you through Berenger, not through our materials. He will have done his own research." She paused. "Go in without the document. Just you and what you know. That's what he's actually there to see."
"I know," I said.
"I know you know," she said. "I'm saying it anyway."
She ended the call.
I put the phone down.
Looked at the notebook. At the circled 47,200,000 with the question mark that now had its answer.
Opened to a fresh line. Wrote: Pemberton Family Office — Ashby. Wednesday London. Do not bring the document. Be what you are.
I told Margaux that evening.
She was at the desk — it was seven, the working day technically over, but she had found something in the Durham due diligence background that she wanted to finish reading before closing the laptop.
"Pemberton," I said.
She looked up.
"Robert Ashby emailed today," I said. "He found us through Geneva. He wants a meeting Wednesday."
She was quiet for a moment.
"The panel's number," she said. "47.2 million."
"Yes," I said.
"It was Pemberton," she said. "The initial tranche."
"Yes," I said.
She looked at the notebook on the desk between us. At the circled number with the question mark.
"Three weeks ago," she said.
"Yes," I said.
She thought about this.
Not about the money — I knew from the quality of her silence when she was thinking about money and when she was thinking about something else. She was thinking about something else.
"James," she said.
"Yes."
"The panel produced that number from its synthesis of the full situation," she said slowly. "Three weeks before Ashby's email. Before there was any public signal that he was going to contact you." She paused. "Which means the synthesis included — what? His research process? His conversations at the Geneva conference? The connection between Petra's mention and his due diligence timeline?"
"I don't know," I said. "I don't know what inputs produced it."
"But it knew," she said.
"It synthesised," I said. "Which is different."
"Is it?" she said.
I thought about this.
About phase-three integration at seventy-three percent and advancing. About four new synthesis domains active. About the panel producing a number without context or label — just a number, sitting at the edge of my vision like something that had been waiting to become real.
"I'm not sure anymore," I said.
She held my gaze.
"That answer," she said, "is more honest than the distinction."
"Yes," I said. "It is."
A silence.
"James," she said.
"Yes."
"Whatever it is — whether it synthesises or whether it knows — it is part of you." She paused. "That means it is working toward the same things you're working toward." She paused. "Which means you can trust it."
I looked at her.
"I've been trusting it for two years," I said.
"I know," she said. "I'm saying you can trust it more." She paused. "You've been holding it at arm's length even while integrating with it. Treating its outputs as inputs to verify rather than as part of your own thinking." She paused. "Elise said you were thinking with it from day one. I'm saying — let yourself think with it. Fully."
I sat with this.
The arm's length she described was real. I had maintained it deliberately — the discipline of treating the panel's output as signal to be verified rather than conclusion to be accepted. The documentation. The independent thesis for every position. The refusal to act on signal alone.
All of that was correct practice.
All of that was also, I understood now, a form of distance. A way of maintaining the distinction between James Smith and the system in his eye.
A distinction that Elise had told me did not exist.
That had never existed.
"Wednesday," I said.
"Wednesday," she said.
"Go with what I know," I said. "Not the document."
"Yes," she said.
"Which is—" I paused. "Everything the panel has synthesised, and everything I have understood from it, and everything I have built on top of both." I paused. "Together."
"Yes," she said. "That's what you are. All of it. Together."
She turned back to the Durham background.
I turned to the notebook.
At the bottom of the page, below the circled 47,200,000, I wrote one more line:
She's right. Trust it more. The arm's length was always the wrong distance.
Then I closed the notebook.
Outside, August Paris made its honest way toward evening — the light going amber over the rooftops, the city finding its slower register, the specific quality of a summer that was almost done but was not done yet.
The east windows held it for a moment.
I let them.
End of Chapter 33
The Probability of Wealth — By J. Valmont
[Next Chapter → Chapter 34: "Ashby"]Robert Ashby arrives in London on a Wednesday. He sits across from James in a Bishopsgate meeting room with no document between them. He asks seven questions. The seventh one is the one nobody has ever asked — and James's answer, which he gives without thinking, is the most honest thing he has said about himself in two years. Catherine Pryce, two floors up, receives one text message. It reads: We're done. She books dinner.