by Liam Vaughan
CHAPTER 8
A BRIEF HISTORY OF SPOOFING
Nav’s wager on the financial crisis left him with more firepower than ever, but the reality was that scalping the market was becoming harder and harder for human traders. More than a year had passed since That’s a Fugazi had revealed his intention to build a program to compete in an increasingly automated world, and he was now ready to put it into action. Nav’s grand vision was to induce other participants to trade how he wanted them to. It was only a more sophisticated version of what traders had been doing, one way or another, for centuries.
When Daniel Defoe wasn’t writing novels like Robinson Crusoe and Moll Flanders or spying for the monarchy, he spent his days in eighteenth-century London working as a merchant, buying and selling wine, hosiery and perfume. A prolific social commentator, Defoe turned his gaze in 1719 to Exchange Alley in the City of London, a narrow, grey-stone street near the Bank of England where, in rowdy coffeehouses, some of the earliest recorded trading in stocks and commodities took place. ‘’Tis a trade founded in fraud, born of deceit and nourished by trick, cheat, wheedle, forgeries, falsehoods and all sorts of delusions,’ Defoe wrote in his essay ‘Anatomy of Exchange Alley’. One passage, recounting the methods of one of the most successful ‘stockjobbers’, a future member of Parliament and governor of the East India Company named Sir Josiah Child, offers an early description of the art of spoofing.
If Sir Josiah had a mind to buy, the first thing he did was to commission his brokers to look sower, shake their heads, suggest bad news from India; and at the bottom it followed, ‘I have commission from Sir Josiah to sell out whatever I can,’ and perhaps they would actually sell ten, perhaps twenty thousand pound. Immediately the Exchange was full of sellers; nobody would buy a shilling, ‘till perhaps the stock would fall six, seven, eight, ten per cent, sometimes more; then the cunning jobber had another set of men employed on purpose to buy, but with privacy and caution, all the stock they could lay their hands on, ‘till by selling ten thousand pound, at four or five per cent lost, he would buy a hundred thousand pound stock at ten or twelve per cent under price; and in a few weeks by just the contrary method, set them all a buying, and then sell them their own stock again at ten or twelve per cent profit.
Fast-forward 275 years to the 1990s, and similar antics could be observed over a much shorter time frame on the Liffe floor at the Royal Exchange, which was situated less than a hundred meters from Exchange Alley. ‘I would get an order to sell two thousand lots and the client, say Goldman Sachs, would tell me to make as much noise as possible,’ recounts a former broker. ‘A thousand offered at ninety-nine! A thousand lots at ninety-eight! A thousand at ninety-seven! I’d sell a few, the price falls, and guess who was waiting at ninety-five bid somewhere else in the pit with a different broker? Goldman, who then starts buying everything, pushing the price back up. When the other traders came back to me and said they’d take the contracts, I’d tell them it was all gone.’
Spoof was the name of a card game invented in the 1880s by British comedian and music hall performer Arthur Roberts. The game revolved around trickery, and the verb ‘to spoof’ came to mean to hoax or deceive. Students play a derivation that involves guessing how many coins are concealed in their fists to decide who’s buying the next round. In tech circles, it refers to an attempt to imitate someone’s identity to gain access to data or money. One of the earliest references to spoofing in relation to financial markets was a 1999 New York Times article titled ‘Chasing Ghosts at Nasdaq’, which documented a rise in cancelled orders in the stock market.
In the pits, spoofing was to some extent curbed by the fact that traders could see whom they were competing against. Serial offenders were liable to be taken outside and made to understand the error of their ways. But when anonymous, screen-based trading came along, that safeguard disappeared. Another accelerant was the introduction of the ladder. For the first time, market participants could see, not just the current best bid and offer, but where orders were waiting up and down the order book. This provided a valuable insight into supply and demand, but it also created new opportunities for deception. One man who famously capitalised was Paul Rotter, aka ‘The Flipper’. Rotter, a skinny, unassuming German with a rebellious streak, started his career in 1994 in the drab back office of a bank in Munich, where he spent his days punching customer trades into a DTB machine, one of the earliest incarnations of an electronic trading screen. The job was tedious but Rotter found he had a gift for spotting patterns in the way prices moved, and within a year he moved to Frankfurt to take up a position as a junior trader. His arrival coincided with the explosion of electronic trading and by the time the pits closed, he was a lethal online scalper. Making money from Luddite open-outcry traders desperately trying to apply what they’d learned in the pits on a computer was embarrassingly easy. ‘It was paradise,’ recalls Rotter. ‘All these locals were used to seeing JPMorgan and Goldman orders and front-running, but they couldn’t do that anymore because it was anonymous and they had no idea what they were doing. There was no algorithmic trading yet and the regulators hadn’t brought in all these rules on what you could and couldn’t do.’
At twenty-four, Rotter moved to Ireland to set up his own fund with some associates. Capitalising on the herdlike behavior of the locals, he would, according to reports, load up the ladder with buy orders and wait for others to line up alongside him. Then, once the market had gone up by a few ticks, he’d cancel his bids and quickly sell into the unsuspecting traders who’d followed him for a higher price than he paid, closing out his position for a profit. By trading bigger than anyone else, he found he could control the market. However, he encountered a problem: with so much buying and selling, Rotter found he sometimes ended up transacting with himself, a prohibited practice known as wash trading. When he started receiving warning letters from the exchange, he contacted his software provider, Trading Technologies (TT), and said he needed a way to stop hitting his own orders. TT came up with a new function called ‘avoid orders that cross’ that solved the problem but inadvertently helped Rotter take his exploits to another level. Where Rotter had previously had to cancel his orders before swapping direction, now he could do both simultaneously with a single click of his mouse. Traders who had tried to piggyback his orders had no chance of getting out of the way when he switched back on them, instantaneously flipping from buyer to seller and vice versa. At one stage, Rotter was trading two hundred thousand contracts a day and, in a good month, raking in $7 million. Along the way, he made staunch enemies among a community of former locals who, not knowing his identity, christened him ‘The Flipper’ and put pressure on the exchanges to ban him. Rotter’s cover was blown on a forum in 2004, and he was confronted in the audience of an industry event in London.
‘It was a little bit frightening at that time,’ he says. ‘These guys were like, “You should look behind your shoulder.” I got messages saying, “We’ll come after you.”’ Rotter pulled back when HFT arrived and new regulations came in. Today he takes much longer-term positions from his home in the Bahamas, but he’s unrepentant. ‘I’ve met many people in the market, stockbrokers from the eighties who had all this inside information. I never did anything like that. I was trading the order book and seeing what other guys were doing and reacting. They thought they had a right to wait for orders and front-run them, and I took advantage. But my money was always at risk. It was a fair market.’
Inheriting Rotter’s mantle as the bogeyman of futures was Igor Oystacher, a Russian day trader who, within a few years of learning to trade, became one of the biggest e-mini traders in the world. Oystacher grew up in a well-off but abstemious family in Moscow where his father, an engineer, introduced him to chess when he was six years old. By age ten, Oystacher could beat his dad, and not long after he was crowned city champion for his age group. He finished high school, an institution for gifted children focused on physics and astronomy, early and moved to Detroit to live with relatives before
enrolling in the maths programme at Northwestern University in Illinois. In his third year, he secured an internship at a prop shop in Chicago called Gelber Group, where he was so successful he quit his degree to trade full-time.
Liberated from the strictures of Russian life and away from the influence of his father, Oystacher devoted himself to making money. By now, he’d abandoned chess for speed chess, a variant of the game that requires players to make decisions in seconds, and with his rapid-fire reactions, pattern recognition skills and instant recall, he took to the ladder immediately. Dour and inscrutable, Oystacher was known as ‘Snuggles’ among the traders on Gelber’s two-hundred-strong trading floor. He approached trading like chess, as a battle with moves and countermoves, and in 2004 he became the focus of unwanted attention when complaints about a mysterious entity with the tag 990 started appearing on the forums. Back then, each firm was given a numeric ID that showed up on post-trade confirmations. Gelber’s was 990, and it didn’t take long for people to deduce that the perpetrator was the firm’s aggressive new Russian whiz kid.
Like Rotter, Oystacher used TT’s ‘avoid orders that cross’ function to switch directions rapidly. Buying and selling in huge volumes, he blew through everything in his path. ‘I know it sounds hard to believe that one person can control a world market but trust me that is what is occurring,’ wrote a disgruntled e-mini trader on one of the forums. ‘He started doing this with 300 lots … now he has made so much money doing it that he is up to 2000 lots,’ worth around $120 million. Oystacher ignored the complaints, which he believed to be inaccurate and unfounded, and passed his methods on to a new group of traders, including an unassuming Chinese mathematician named James ‘Jimmy’ Chui whose homework he had copied at Northwestern. In 2007, Chiu was headhunted by Jump Trading, the Chicago high-frequency trading giant, where he would become an even bigger player than Oystacher and pass on his knowledge to a new generation.
Nav, Oystacher and Chiu were part of an elite cadre of big-ticket human scalpers, and the three sometimes clashed in a market increasingly inhabited by preprogrammed machines. Years later, investigators would talk about the ‘Spoof Wars’ of the late 2000s, an ultra-high-stakes game of brinkmanship in which players would try to wipe each other out by hitting their opponents’ spoof orders and getting them to ‘puke’ – close out positions they never intended to consummate for a devastating loss. Nav had read about Oystacher on the boards and throughout his career would complain bitterly to the CME and his peers about ‘The Russian’, who he was convinced was trying to spoof the market higher when he was shorting it. In truth, although the two men did exchange blows, Nav was more often than not duking it out with some other trader who happened to have an MO similar to Oystacher’s.
Amid such ferocious competition, Nav looked to other avenues to make money. One tactic was to focus on the ‘pre-open’. During the week, the e-mini traded twenty-four hours a day apart from between 3.15 p.m. and 3.30 p.m. CET, when the market shut down. In that fifteen minutes, traders could place nonbinding orders that the CME used to calculate an ‘indicative opening price’, or IOP, based on the level where the greatest number of buyers and sellers matched. As traders placed, modified and cancelled their orders, they could see the IOP rising and falling until, with thirty seconds to go, it locked up and the opening price was set. Prices fluctuated wildly around the open, and any trader who positioned themselves correctly for the number the CME spat out could make a fast profit. There were rumours that some traders had found a way to affect the outcome by repeatedly placing large orders above or below the prevailing price, inducing other participants to follow suit, then cancelling their orders just before the cutoff. It was certainly a lucrative period for Sarao. ‘I saw him come in for the open, make a hundred thousand dollars, then go home again,’ recalls one former colleague. But in March 2009, the CME contacted Nav, reminding him that all orders must be ‘bona fide’.
From a hired desk in a small trading arcade in London, it wasn’t always clear what bona fide meant. The US Commodity Exchange Act of 1936 made it a felony to ‘manipulate or attempt to manipulate the price of any commodity in interstate commerce, or for future delivery’, but pre-2011, when new legislation was introduced, the policing of futures markets was patchy and ineffectual. Specific rules on spoofing hadn’t been brought in yet, and proving manipulation in court was so difficult that the industry’s main watchdog, the Commodity Futures Trading Commission, rarely brought cases. Day-to-day oversight of the market was left to exchanges like the CME, which preferred to view market participants as valued customers rather than regulated entities. Anyone suspected of attempting to manipulate prices was given a series of warnings before eventually being hit with a fine that often didn’t cover their alleged gains.
As futures markets had evolved, supervision had failed to keep up. Trading was now so fast, and the data it created so voluminous, that the authorities were simply unable to monitor what was happening on their watch. Every time Nav loaded up the ladder he saw spoofing, wash trading, momentum ignition and other forms of chicanery, with seemingly no consequences. When he complained to the CME about it, which he did frequently, they ignored him or told him he was wrong. If he retaliated and used the same tactics himself, nobody seemed to care. The impression he was left with was that, in the era of the algos, anything went. As he’d commented on the boards back in 2007, ‘With the volume I do I know the exchange will turn a blind eye.’
If Nav had worked at a bank or a hedge fund, there would have been a compliance officer to step in when he strayed into dodgy territory, but as an independent day trader he was on his own. The brokers at GNI were officially supposed to ‘diligently supervise’ his activities, but they were philosophically and financially inclined to stay out of it. When a warning did arrive from the CME, like the one about Nav’s trading in the pre-market, they simply passed it along.
Nav believed he was engaged in an existential battle against opponents with superior powers and advantages. If he were to stand any chance, he reasoned, he would have to build a machine of his own. Unlike the HFT firms, he didn’t have the resources to create something powerful and rapid from scratch. His approach, instead, would be to take the off-the-shelf software package he already owned and pimp it like a gearhead souping up an old Ford. Despite the gulf in technology, Nav felt confident he could succeed because the robots had an inherent weakness he knew he could exploit. They were followers, blindly reacting to signals contained in trading data in preordained, predictable ways. Nav knew he couldn’t analyse the order book as efficiently as a machine, and he was always going to be second best in a speed race. But if he could fuck with the signals themselves, he could get the robots to respond to his commands and take back some control. Oh sheep, come join thy shepherd!
CHAPTER 9
BUILDING THE MACHINE
Ever since Nav started trading at Futex, he’d used a suite of programs provided by the Chicago-based software vendor Trading Technologies. Founded in 1994, TT was run by a futures legend named Harris Brumfield who started using the software after leaving the pits and liked it so much that he bought the company. The software, which was ubiquitous among day traders, gave users the tools they needed to observe the market, place assorted order types, and connect to the exchanges with minimal delays. TT also offered a product called Autotrader that allowed customers with no background in programming to create their own algos in Excel, offering a low-cost way to introduce an element of automation into their trading. On Friday 12 June 2009, Nav emailed his broker at GNI Touch asking to be put in touch with ‘a TT technician that will be able to programme for me extra features on TT … Obviously I would be prepared to pay for their time / excellence.’ A few days later, he followed up with a message to TT’s London-based sales rep with the subject line ‘Matrix’. It read:
Hello mate,
What I need are the following functions
i)The cancel if close function we discussed. I would also like to be able to
alternate the closeness ie one price away or three prices away etc etc
ii)Join bid/offer function which will work like the stop market function in that you can hit into orders but you don’t get filled. Your order simply joins the bid/offer if it goes bid/offered
iii)A facility to be able to enter multiple orders at different prices using one click
iv)The ability for my orders to rest on a particular size, ie my order will be pulled if there are not x amount of orders beneath it. Of course to make this work we will have to stay at the back of the book,
v) this can be done by increasing/decreasing my order by a 1 lot every time a new order is detected where I am resting
vi)To be able to enter an integer and my order will stay working until a clip has entered the order that is of equal or greater value than the integer.
vii)The ability for my orders to only allow 1 clip to go into them. Hence, if I am working 500 lot and a 2 lot trades, the 498 balance is removed immediately.
All of the above shouldn’t be difficult to do, since there are people using these matrices in every market I have traded so it is fairly common.
best regards
Nav
Nav’s blueprint may seem complicated, but his goal was as straightforward as the Flipper’s or Lord Josiah Child’s in Exchange Alley: to distort the picture of supply and demand enough to mislead other market participants, allowing him to buy cheap and sell high. The spoofing machine he envisaged would contain a raft of features that could be switched on and off with a mouse as Nav traded in real time, augmenting his natural abilities. Principal among these was what he called the ‘cancel if close’ function, and the US authorities would later describe as a ‘layering algorithm’. When activated, it would allow Nav to place a number of large sell orders a designated number of ticks above the best offer. As the market moved higher and lower, Nav’s orders would move in lockstep, always maintaining the specified distance from the current price to minimise the chances of being hit. In the complex world of HFT, it was a surprisingly rudimentary mechanism based around the principle that when algos noticed a jump in the number of sellers relative to buyers, they would also start selling and the market would fall. As the layering algo blew prices lower like an industrial fan, Nav would profit by simultaneously selling some e-minis manually, waiting for the market to fall by a couple of ticks, then exiting his position by buying the same number back and cancelling the spoofs. A few minutes later, when conditions had stabilised, he’d start the process again.
