Hari Kunzru - Gods Without Men

“Well, I—”

“And you need to tell me what in hell’s name your conscience has to do with the price of rice.”

This was not a conversation Jaz wanted to have, not today. Preferably not ever, but particularly not today. He thought of asking Willis whether he could call him back, but that wasn’t really an option. If right now was when Fenton wanted to talk about Cy Bachman and the Walter model and all the rest of the shit Jaz had hoped to keep in a holding pattern over the fan for another few days, then right now it would have to be. It was obvious what Bachman had been saying. Their relationship had never been straightforward, and now — after their argument — he wanted Jaz off the team. Fenton was doing him a courtesy, allowing him to defend himself, but it was probably a fait accompli . He assumed his security pass had been deactivated. They were probably boxing up his personal effects for the courier.

This had been coming for a while.

He’d first set eyes on Cy Bachman two years previously, over lunch at a steakhouse in the Financial District, the kind of place Willis favored for meetings, where you could eat eighty-five-dollar Wagyu burgers and wash them down with bottles of Opus One. Bachman turned out to be vegetarian, a fact Willis evidently knew and had ignored when making the booking. While the CEO told a boring story about a horse he was thinking of buying from a stable in Saratoga, Jaz had watched an elegant, fiftyish, shaven-headed man shoot his French cuffs and tackle an enormous bowl of arugula, whose size appeared to be the kitchen’s consolation for the meal’s total absence of protein. It occurred to him the salad was a joke — the place was known for the “no rabbit food” motto emblazoned on its creamy letterpress menu. Bachman affected neither to notice nor to care.

When Willis finished the horse story, Bachman smiled at Jaz and complimented him on a paper he’d coauthored at MIT, outlining a simplified statistical technique for describing the behavior of certain assemblies of particles. Jaz was disarmed, but at the same time wary. Bachman had a reputation as one of the most talented financial engineers on Wall Street; it was an open secret that Willis had poached him from one of the big banks to head a new research team. He assumed the lunch was because Willis wanted him to work under Bachman. The comment was his new boss’s way of letting him know he had prepared. Later he’d discover that this care and meticulousness was carried through to every aspect of Bachman’s life, from his fastidiously stylish dress to his almost neurotic concern for the visual presentation of data. A trailing zero could drive him into a rage. He insisted his team was “properly attired” even if all they were doing was writing code.

Willis seemed untouched by Bachman’s aura, his WASP sense of entitlement and large personal fortune providing an effective shield against intellect. “Enjoying your meal, Cy?” he chortled.

Bachman made a face. “This is revenge,” he explained. “I took him to a raw-food place in the Village.”

“Bastards made me a coffee out of pistachio nuts.”

Jaz laughed heartily. He knew better than to be fooled by Fenton’s bluff manner. Behind the genial clubman’s mask, the oak-paneled three-martini smokescreen he put up to fool the credulous, a ruthless tactician lurked. When it came to the acquisition of money, he was entirely pragmatic, prepared to act without prejudice or sentiment. In this respect, he was quite brilliant. Jaz couldn’t help but connect this ability to suspend judgment, to take each new situation entirely on its own merits, with the image of a man crawling down a tunnel with a gun in his hand, feeling his way in the dark.

“So Jaswinder. Cy’s taken a look at your work and he thinks he could use you on Walter.”

“Walter?”

“It’s a new global quant model.”

“Goddamn theory of everything, isn’t that right, Cy?”

“If you say so, Fenton. Everything would be kind of a large dataset.”

Jaz was intrigued. “What stage are you at?”

“Personally,” interrupted Willis, “I think it’s just great already. If it was up to me I’d go live right now, start counting my winnings. But Cy says the bastard’s got a half-life of about twenty seconds, and if we go off all premature we’ll blow the chance of a bigger payday down the line.”

“But it is down to you, Fenton. Just say the word.”

“Cy. If you tell me I can have a dollar today or three tomorrow, I’ll take the three bucks. Deferred gratification — it’s what separates civilized man from chimps and children. We’re getting OK returns on the established models, so I’m happy to wait. Just as long as Renaissance or those bastards at Goldman don’t get there before us.”

“Fenton, I’d be very surprised if they had any interest in this strategy.”

“Well, I wouldn’t. Probably bugging the damn table decorations in this joint, paying off the sommelier. Speaking of which, let’s get another bottle.”

When Jaz presented himself at Bachman’s office the following day, he expected to be shown some kind of formula. The Walter setup was very cloak-and-dagger, a separate address, pin codes and biometrics to get through the door. Bachman had a view of the Hudson and a display case full of curios behind his desk that Jaz avoided scrutinizing too closely, in case it led to a conversation about basketry or ceramics or netsuke, topics that would quickly lead him out of his comfort zone. Luckily Bachman got straight to business. He told him the best way to understand the model was to work with it, which seemed sensible enough. When Jaz asked about its basic principles, he waved the question away.

Bachman’s model was conventional in that it relied on discovering certain predictable behaviors in the market — regularities, trackable cycles — and using that knowledge to trade. But as far as Jaz could grasp from the initial presentation, which took almost three hours and left him feeling like he’d been sparring with some kind of higher-dimensional gorilla, the type of regularities Walter sought were particularly fleeting and unstable. The model was being trained not simply to exploit some temporary price disparity but to identify and track entirely ad hoc constellations of five, six, seven variables, brief but dazzling phenomena, lightning flashes of correlation. The math, Jaz thought, was some of the most beautiful he’d ever encountered. The problem that would come to tug at him like an importunate child was something else. Something about Walter’s responsiveness, its voracious thirst for data. It was more like an organism than a computer program. It felt alive .

For the first few months he had little to do with Walter’s guts, the software that identified patterns and executed trades. His job was to take certain datasets and hunt for statistical relationships, what Bachman called “rhymes.” The material (prepared according to some arcane process Bachman refused to discuss) came in discrete clusters, little clots of seemingly unrelated numbers. Some of it was familiar: commodity and share prices, government bond yields, interest rates, currency fluctuations. But there was other data: on shopping-mall construction, retail-sales figures, drug-patent applications, car ownership; on the incidence of birth defects, industrial injuries, suicides, controlled-substance seizures, cell phone tower construction. Walter consumed the most esoteric numbers: small-arms sales in the Horn of Africa; the population of Gary, Indiana, between 1940 and 2008; the population of Magnitogorsk, Siberia, for the same years; prostitution arrests in major American cities; data traffic over the TPE trans-Pacific cable; the height of the water table in various subregions of the Maghreb.