The Hour Between Dog and Wolf: Risk-taking, Gut Feelings and the Biology of Boom and Bust

The trouble stems from the fact that our visual system is surprisingly slow. When light hits our retina, the photons must be translated into a chemical signal, and then into an electrical signal that can be carried along nerve fibres. The electrical signal must then travel to the very back of the brain, to an area called the visual cortex, and then project forward again, along two separate pathways, one processing the identity of the objects we see, the ‘what’ stream, as some researchers call it, and the other processing the location and motion of the objects, the ‘where’ stream. These streams must then combine to form a unified image, and only then does this image emerge into conscious awareness. The whole process is a surprisingly slow one, taking, as mentioned, up to one tenth of a second. Such a delay, though brief, leaves us constantly one step behind events.

Neuroscientists have discovered another problem with the idea that we are watching the world live. An important part of this idea is the notion that our eyes objectively and continuously record the scene before us, much like a movie camera. But eyes do not operate like this. If we continuously recorded the visual information presented to us, we would waste a great deal of time (and probably suffer constant headaches) looking at blurred images as our eyes pan from one scene to another. More importantly, we would be swamped by the sheer amount of data, most of which is irrelevant to our needs. Live streaming takes up an enormous amount of bandwidth on the internet, and it does so as well in our brains. To avoid a needless drain on our attentional resources, our brain has hit upon the tactic of sampling from a visual scene, rather than filming it. Our eyes fix on a small section of our visual field, take a snapshot, then jump to another spot, take a snapshot, and quickly jump again, much like a hummingbird nervously flitting from flower to flower. We are largely unaware of this process, and do not see a blur when our eyes shift location because, remarkably, the visual system stops sending images up to consciousness while it jumps from scene to scene. Furthermore, we are unaware of these jumps and intervening blackouts because our brain weaves these images seamlessly into something that does appear much like a movie. We can perform up to five of these visual jumps per second, the minimum amount of time required for a shift in view being therefore one fifth of a second.

If we return to sports, we can see that some numbers do not add up. How can a cricketer at silly point catch (or duck) a ball in under a tenth of a second if he is not even aware of it yet? How can he direct his attention to the ball if it takes him twice as long just to move his eyes? And when dealing with these numbers we have not even begun to consider the additional 300–400 milliseconds required for an elementary cognitive decision or inference, and the 50 milliseconds or so it takes for a motor command to be communicated by nerves to our muscles. The picture conjured by these numbers is one of an infielder frozen in the readiness stance, eyes fixed like a waxwork model, while a projectile shudders past his immobile and fragile head.

The same questions we ask about athletes can be asked, and with more urgency, beyond the cricket pitch. How can we humans survive in a brutal and fast-moving world if our consciousness arrives on the scene just after an event is over? This is a baffling question. But asking it allows us to see what is wrong with the notion of the brain as a central processor, taking in objective information from the senses in the manner of a camera, processing this information rationally, consciously and discursively, deciding on the appropriate and desired action, and then issuing motor commands to our muscles, be they larynx or quadriceps. Each of these steps takes time, and if we were indeed programmed to behave this way, then life as we know it would be very different. If we had to think consciously about every action we took, sporting events would become odd, slow-motion spectacles that few people would have the patience to watch. Worse, in nature and in war we would have long ago fallen prey to some quicker beast.

I, CAMERA?

It turns out that there is something wrong with each step in this supposed chain of mental events. The eye takes snapshots rather than movies; but even these snapshots are not a photographic and objective record of the outside world. All sensory information comes to us tampered with. Like the news on TV, it is filtered, warped and pre-interpreted in a way designed to catch our attention, ease comprehension and speed our reactions.

Take for instance the ways in which the brain deals with the problem of the one-tenth-of-a-second delay between viewing a moving object and becoming consciously aware of it. Such a delay puts us in constant danger, so the brain’s visual circuits have devised an ingenious way of helping us. The brain anticipates the actual location of the object, and moves the visual image we end up seeing to this hypothetical new location. In other words, your visual system fast-forwards what you see.

An extraordinary idea, but how on earth could we ever prove it to be true? Neuroscientists are devilishly clever at tricking the brain into revealing its secrets, and in this case they have recorded the visual fast-forwarding by means of an experiment investigating what is called the ‘flash-lag effect’. In this experiment a person is shown an object, say a blue circle, with another circle inside it, a yellow one. The small yellow circle flashes on and off, so what you see is a blue circle with a yellow circle blinking inside it. Then the blue circle with the yellow one inside starts moving around your computer screen. What you should see is a moving blue circle with a blinking yellow one inside it. But you do not. Instead you see a blue circle moving around the screen with a blinking yellow circle trailing about a quarter of an inch behind it. What is going on is this: while the blue circle is moving, your brain advances the image to its anticipated actual location, given the one-tenth-of-a-second time lag between viewing it and being aware of it. But the yellow circle, blinking on and off, cannot be anticipated, so it is not advanced. It thus appears to be left behind by the fast-forwarded blue circle.

The eye and brain perform countless other such tricks in order to speed up our understanding of the world. Our retina tends to focus on the front edge of a moving object, to help us track it. We process more information in the lower half of our visual field, because there is normally more to see on the ground than in the sky. We group objects into units of three or four in order to perceive numbers rather than count them, a process, known as subitising, that comes in handy when assessing the number of opponents in battle. We rapidly and unconsciously assume an object is alive if it moves in certain ways, regularly changing direction say, or avoiding other objects, and then pay it closer attention than we would if it was inanimate.

Our reaction times can also be speeded up by relying more on hearing than vision. That may seem counter-intuitive. Light travels faster than sound, much faster, so visual images reach our senses before sounds. However, once the sensations reach our eyes and ears, the relative speeds of the processing circuits reverse. Hearing is faster and more acute than seeing, about 25 per cent so, and responding to an auditory cue rather than a visual one can save us up to 50 milliseconds. The reason is that sound receptors in the ear are much faster and more sensitive than anything in the eye. Many athletes, such as tennis and table-tennis players, rely on the sound a ball makes on a racket or bat as much as on the sight of its trajectory. A ball hit for speed broadcasts a different sound from one sliced or spun, and this information can save a player the precious few milliseconds that separate winners from losers.

If we now add up all the time delays between an event occurring in the outside world and our perceiving it, we discover the following lovely fact. For events occurring at a distance, we see them first and hear them with a delay, as we do, for example, when seeing lightning and hearing the thunder afterwards. But for events taking place close to us, we hear them, because of our rapid auditory system and relatively slow visual one, slightly in advance of seeing them. There is, though, a point at which sights and sounds are perceived as occurring simultaneously, and that point is located about ten to fifteen metres from us, a point known as the ‘horizon of simultaneity’.

Could our more rapid hearing provide traders with an edge over competitors? Right now, all price feeds onto a trading floor are visual images on a computer screen. But the technology does exist for supplying audio price feeds. These have already been supplied to blind people, and apparently they sound much like an audiocassette on fast forward. Such a feed could give traders a 40-millisecond edge. That is not much time. But who knows, it could prove decisive when hitting a bid or lifting an offer during a fast market.

Bringing a trader’s hearing into play may have a further advantage. Research in experimental psychology has found that perceptual acuity and general levels of attention increase as more senses are involved. In other words, vision becomes more acute when coupled with hearing, and both become more acute when coupled with touch. The explanation ventured for these findings is that information arriving from two or more senses instead of just one increases the probability that it is reporting a real event, so our brain takes it more seriously. Many older trading floors may have inadvertently capitalised on this phenomenon, because they came equipped with an intercom to the futures exchanges, with an announcer reporting bond futures prices: ‘One, two … one, two … three, four … fours gone, fives lifted, size coming in at six …’ and so on. With the advent of computerised pricing services, many companies felt this voice feed was antiquated and discontinued the service. Yet by bringing in a second sense it may have been an effective way of sharpening attention and reactions among the traders.

KNOWING BEFORE KNOWING

All these ad hoc adjustments to the information being transmitted to your conscious brain keep you from falling hopelessly behind the world. But the brain has an even more effective way of saving you from your fatally slow consciousness. When fast reactions are demanded it cuts out consciousness altogether and relies instead on reflexes, automatic behaviour and what is called ‘pre-attentive processing’. Pre-attentive processing is a type of perception, decision-making and movement initiation that occurs without any consultation with your conscious brain, and before it is even aware of what is going on.

This processing, and its importance to survival, has nowhere been better described than in the extraordinary book All Quiet on the Western Front, written by Erich Maria Remarque, a soldier who served in the trenches during the First World War. Remarque explains that to survive on the front soldiers had to learn very quickly to pick out from the general din the ‘malicious, hardly audible buzz’ of the small shells called daisy cutters, for these were the ones that killed infantry. Experienced soldiers could do this, and developed reactions that kept them alive even amid an artillery bombardment: ‘At the sound of the first droning of the shells,’ Remarque tells us, ‘we rush back, in one part of our being, a thousand years. By the animal instinct that is awakened in us we are led and protected. It is not conscious; it is far quicker, much more sure, less fallible, than consciousness. One cannot explain it. A man is walking along without thought or heed; – suddenly he throws himself down on the ground and a storm of fragments flies harmlessly over him; – yet he cannot remember either to have heard the shell coming or to have thought of flinging himself down. But had he not abandoned himself to the impulse he would now be a heap of mangled flesh. It is this other, this second sight in us, that has thrown us to the ground and saved us, without our knowing how.’

Neuroscientists have long known that most of what goes on in the brain is pre-conscious. Compelling evidence of this fact can be found in the work of scientists who have calculated the bandwidth of human consciousness. Researchers at the University of Pennsylvania, for example, have found that the human retina transmits to the brain approximately 10 million bits of information per second, roughly the capacity of an ethernet connection; and Manfred Zimmermann, a German physiologist, has calculated that our other senses record an additional one million bits of information per second. That gives our senses a total bandwidth of 11 million bits per second. Yet of this massive flow of information no more than about 40 bits per second actually reaches consciousness. We are, in other words, conscious of only a trivial slice of all the information coming into the brain for processing.

A fascinating example of this pre-conscious processing can be found in a phenomenon known as blindsight. It became a topic first of curiosity and then of medical concern during the First World War, when medics noticed that certain soldiers who had been blinded by a bullet or shell wound to the visual cortex (but whose eyes remained intact) were nonetheless ducking their heads when an object, such as a ball, was tossed over their heads. How could these blind soldiers ‘see’? They were seeing, it was later discovered, with a more primitive part of the brain. When light enters your eye its signal follows the pathways, described above, back to your visual cortex, a relatively new part of the brain. However, part of the signal also passes down through an area called the superior colliculus, which lies underneath the cortex, in the midbrain (fig. 5). The superior colliculus is an ancient nucleus (collection of cells) that was formerly used for tracking objects, like insects or fast-moving prey, so that our reptilian ancestors could, say, zap it with their tongues. Now largely layered over by evolutionarily more advanced systems, it nonetheless still works. It is not sophisticated: it cannot distinguish colour, discern shape or recognise objects, the world appearing to the superior colliculus much like an image seen through frosted glass. But it does track motion, capture attention and orient the head towards a moving object. And it is fast. Fast enough, according to some scientists, to account for a batsman or a close fielder’s rapid tracking of a cricket ball. Lastly, blindsight operates without us ever being aware of it.

Fig. 5. The visual system. Visual images travel by electrical impulses projected from the retina to the visual cortex at the back of the brain. They are then sent forward along the ‘what’ stream, which identifies the object, and the ‘where’ stream, which identifies its location and movement. An older, faster route for visual signals travels down to the superior colliculus where fast-moving objects can be tracked.

To what features of the world do we pre-attend? When a close fielder is crouched at the ready, frozen like a statue, his eyes fixed and unable to scan, what in his visual field captures the interest of his pre-conscious processor? We do not yet know a complete answer to this question, but we do know a few things. We attend pre-consciously, as in blindsight, to moving objects, especially animate ones. We attend to images of certain primitive threats, such as snakes and spiders. And we are strongly biased to aurally attend to human voices, and visually to faces, especially ones expressing negative emotions such as fear or anger. All these objects can be registered so rapidly, in as little as 15 milliseconds (this does not include a motor response, of course), that they can affect our thinking and moods without our even being aware of them. In fact we often know whether we like or dislike something or someone well before we even know what or who it is. The speed and power of pre-conscious images, especially sexual ones, were once used in subliminal advertising as a way of biasing our subsequent spending decisions. More usefully, this pre-conscious processing can affect motor commands for reflex actions and automatic behaviours.

One of these reflexes is our startle response, a quick and involuntary contraction of muscles designed to withdraw us, like an escaping octopus, from a sudden threat. It can be initiated by both sights and sounds. A loud bang will trigger the startle, as will a rapidly approaching object in our visual field. The way we visually detect an object on a collision course with us is ingenious: our startle is initiated by a symmetrical expansion of a shadow in our visual field. The expanding shadow indicates an incoming object, and its symmetry indicates that it is heading straight for us. Apparently this pre-conscious object tracking is so well calibrated that if the shadow is expanding asymmetrically our brain can tell within five degrees that the object will miss us, and as a result the startle response is not triggered. The startle, from sensory stimulus to muscle contraction, is exceptionally fast, your head reacting in as little as 70 milliseconds and your torso, since it is farther from your brain, in about 100 milliseconds. Coincidentally, that is roughly the time required for a fielder at silly point to catch a ball coming off a bat. It is entirely possible that close fielders rely on the startle response to achieve the almost inhuman response times they display. If so, then, conveniently, perhaps the fielder can catch or avoid a ball in the little time allowed him only if it is coming straight for his head.

Besides the startle response, how can we react fast enough to meet the challenges sports, and daily life, throw at us? As we saw in the previous chapter, humans have adopted a wide range of movements, like those found in sports and dance and modern warfare and even trading, for which evolution has not prepared us. How can these learned movements become so habitual that they approach the speeds needed for sporting success or survival in the wild? To answer this question we should recognise a basic principle at work in our reflexes and automatic behaviours: the higher we rise in the nervous system, moving from the spine to the brain stem to the cortex (where voluntary movement is processed), the more neurons are involved, the longer the distances covered by nervous signals, and the slower the response. To speed our reactions the brain tends therefore to pass control of the movement, once it has been learned, back to lower regions of the brain where programmes for unthinking, automatic and habitual actions are stored. Many of these learned and now-automatic behaviours can be activated in as little as 120 milliseconds.

A glimpse into this process has been provided by a brain-scanning study of people learning the computer game Tetris. At the beginning of the study, large swathes of the trainees’ brains lit up, showing a complex process of learning and voluntary movement; but once they had mastered the game their movements became habitual, and brain activity in the cortex died down. Their brains now drew much less glucose and oxygen, and their speed of reactions increased markedly. Once the players had the knack, they no longer thought about playing the game. This study, and others like it, supports the old saying that when learning begins we are unconscious of our incompetence, and proceed to a stage where we are conscious of our incompetence; then when training begins we move to conscious competence; and as we master our new skill we arrive at the end point of our training – unconscious competence. Thinking, one could say, is something we do only when we are no good at an activity.

One last point. As fast as these automatic reactions may be, they still do not seem quite fast enough for many of the high-speed challenges we face, and may therefore leave us slightly behind the ball, so to speak. The trouble with these reaction times is just that – they are reactions. But good athletes are not in the habit of waiting around for a ball or a fist to appear, or opponents to make their move. Good athletes anticipate. A baseball batter will study a pitcher and narrow down the likely range of his pitches; a cricket close fielder will have registered a hundred tiny details of a batsman’s stance and glance and grip even before the ball has left the bowler’s hand; and a boxer, while dancing and parrying jabs, will pre-consciously scan his opponent’s footwork and head movements, and look for the telltale setting of his stabiliser muscles as he plants himself for a knockout blow. Such information allows the receiving athlete to bring online well-rehearsed motor programmes and to prepare large muscle groups so that there is little to do while the ball or fist is in the air but make subtle adjustments based on its flightpath. Skilled anticipation is crucial to lowering reaction times throughout our physiology.

Let us finish by listening to Ken Dryden, a legendary goalie in ice hockey and one of the most articulate athletes ever, on the importance of anticipation and automatic behaviour: ‘When a game gets close to me, or threatens to get close, my conscious mind goes blank. I feel nothing. I hear nothing, my eyes watch the puck, my body moves – like a goalie moves, like I move; I don’t tell it to move or how to move or where, I don’t know it’s moving, I don’t feel it move – yet it moves. And when my eyes watch the puck, I see things I don’t know I’m seeing … I see something in the way a shooter holds his stick, in the way his body angles and turns, in the way he’s being checked, in what he’s done before that tells me what he’ll do – and my body moves. I let it move. I trust it and the unconscious mind that moves it.’

To sum up, we humans have been equipped over our long evolutionary training period with a large bag of tricks designed to increase our speed of reactions. In the foregoing discussion I have rummaged in this bag and pulled out only a few of our amazing gadgets. But demonstrating how they work should be enough, I hope, to show just how reliant we are on these quick responses for survival in the wild and in war, for success in sports, and for buying back a large block of bonds sold to DuPont.

WHAT LIES BENEATH

In fact, so fast are our reactions that consciousness is frequently left out of the loop. Given that sobering fact, we have to ask: what role does consciousness play in our lives? We experience our consciousness as something residing in our heads, peering out through our eyes much as a driver peers through a windscreen, so we tend to believe that our brain interacts with our body just as a person interacts with a car, choosing the direction and speed and issuing commands to a passive and mechanical device. But this belief does not stand up to scientific scrutiny. As George Loewenstein, an economist at Yale, points out, ‘There is little evidence beyond fallible introspection supporting the standard assumption of complete volitional control of behavior.’ And he is right, for the stats on reaction times tell us otherwise: we are for the most part on autopilot.

The news gets even worse for the Platonists among us. In the 1970s, Benjamin Libet, a physiologist at the University of California, conducted a famous series of experiments that has tormented many a scientist and philosopher. These experiments were simplicity itself. Libet wired up a group of participants with what are called EEG leads, small monitors attached to the scalp which record the electrical activity in the brain, and then asked them to make a decision to do something, like lift a finger. What he found was that the participants’ brains were preparing the action 300 milliseconds before they actually made the decision to lift their finger. In other words, their conscious decision to move came almost one third of a second after their brain had initiated the movement.

Consciousness, these experiments suggested, is merely a bystander observing a decision already taken, almost like watching ourselves on video. Scientists and philosophers have proposed many interpretations of these findings, one of which is that the role of consciousness may not be so much to choose and initiate actions, but rather to observe decisions made and veto them, if need be, before they are put into effect, much as we do when we practise self-control by stifling inappropriate emotional or instinctive urges. (We may be on autopilot for much of the day, but that does not mean we cannot take responsibility for our actions.) Libet’s experiments, suggesting as they do that consciousness is largely an override mechanism, led one particularly witty commentator, the Indian neuroscientist V.S. Ramachandran, to conclude that we do not in fact have free will; what we have is free won’t.

It seems that consciousness is a small tip of a large iceberg. But what exactly lies below it? What lurks beneath our rational, conscious selves? The eighteenth-century German philosopher Immanuel Kant proposed a particularly intriguing answer to this question: we do not know what is down there. Kant believed that our consciousness – that is, our experience of a unified and understandable world, and of a continuing person experiencing this world – is possible only because our mind constructs this unified experience. If our mind did not organise our sensations the world would be a whirling, blooming confusion. But the mind does: it provides organising constructs, such as space and time, so that we experience a continuing world, just as it does another construct, that of cause and effect, which ties succeeding events together into a coherent story. Kant thought all these unifying constructs applied only to the veil of sensations, and not to the entities creating or lying behind the sensations. These objects we can never know. Inaccessible to rational analysis, forever mysterious to science, these hidden beings can be groped at and suggestively discerned only through art and religion. And it is in this dark world that the soul belongs, putting it too beyond the ken of rationality and beyond the domain of cause and effect. It was upon this argument that Kant rested his belief in free will.

Kant’s philosophy left a deep imprint on German thought. Freud, inspired by Kant’s vision, argued that below the façade of our rational selves, deep in our subconscious, there boils a devil’s cauldron of envy and sexual perversion and patricidal tendencies which warps our judgement. Nietzsche too found beneath our delusions of rationality and morality a dark urge for power and dominance. Modern neuroscience, however, has lifted the lid off this hitherto mystifying brain and found something far more valuable than the entities proposed by nineteenth-century German philosophy – a meticulously engineered control mechanism. More valuable because it has been precisely calibrated over millennia to keep us alive in a brutal and fast-moving world. And we can thank our lucky stars for it, otherwise we would long ago have been battered to extinction. Lifting the lid of our brain does not reveal the nether world of Kant’s unsayable, nor the volcanic will of Nietzsche’s superman, nor yet the hellish subterranean den of Freud’s subconscious. It reveals something that is a lot closer to the inner workings of a BMW.

FAST TIMES ON THE TRADING FLOOR

Let us now return to the financial world, and consider the importance of fast reactions to the success and survival of risk-takers. Traders like Martin frequently face high-speed challenges which demand an equally fast response. The challenges may not demand quite the same speed of reactions as fielding at silly point, but traders nonetheless regularly face time constraints, and when they do their decision-making and trade execution must bypass conscious rationality and draw instead on automatic reactions. This is especially true when markets begin to move fast, as they might in a frantic bull market. Then Martin is obliged to sell bonds to clients or risk alienating the sales force, and must scramble to buy them on the broker screens or from other clients before losing money. At times like this trading is much like a game of snap, and the fastest person wins.

This simple point carries unexpected implications for economics. It is not often appreciated that financial decision-making is a lot more than a purely cognitive activity. It is also a physical activity, and demands certain physical traits. Traders with a high IQ and insight into the value of stocks and bonds may be worth listening to, but if they do not have an appetite for risk then they will not act on their views and will suffer the fate of Cassandra, who could predict the future but could not affect its course. And even if they have a good call on the market and a healthy appetite for risk, yet are shackled with slow reactions, they will remain one step behind the market, and will not survive on the trading desk – or anywhere else in the financial world, for that matter.

Treasury traders, like flow traders more generally (a flow trader is one who trades with clients, handles the flows coming off the sales desks), therefore require a battery of traits: they need a high enough IQ and sufficient education to understand basic economics; a hearty appetite for risk; and a driving ambition. But they also need the physical build. They must be able to engage in extended periods, hours at a time, of what is called visuo-motor scanning, i.e. scrutinising the screens for price anomalies between say the ten-year and the seven-year Treasury bond, or between the bond and currency markets. Such scanning requires concentration and stamina, and not everyone can do it, just as not everyone can run a four-minute mile. And once a price discrepancy has been identified, or a high bid spotted during a sell-off, a trader must move quickly to trade on these prices before anyone else. Not surprisingly, most flow-trading desks, be they ones trading Treasury or corporate or mortgage-backed bonds, usually employ one or two former athletes, a World Cup skier, say, or a college tennis star.

The physical nature of trading is even more apparent on other types of floors. On the floor of a stock exchange or the bond and commodity pits at the Chicago Board of Trade, a trader’s job can resemble a day spent in a wrestling ring. Hundreds of traders stand together, jostling each other and vying for attention when trying to trade with each other, something they do with an arcane system of hand signals. When markets are moving fast and a trader needs the attention of someone on the other side of the pit, then height, strength and speed are of paramount importance in executing a trade, as is the willingness to elbow a competitor in the face. Needless to say, there are not a lot of women in the financial mosh pits.

Another style of trading that makes punishing physical demands is what is called high-frequency trading. This activity involves buying or selling securities, say a bond or stock or futures contract, sometimes in sizes amounting to billions, but holding the positions for only a few minutes, sometimes mere seconds. High-frequency traders do not try to predict where the market is going in the next day or two, let alone the next year, as do asset managers who invest for the long term; they try to predict the small moves in the market, a few cents up or down. As a general rule, the shorter the holding period for a style of trading, the greater the need for its traders to have fast reactions.

Having said all this, there are good reasons for expecting the physical aspect of trading to decline in importance in the financial world. More and more activities are now carried out electronically. The first and most dramatic sign of such a change was the closing down of physical stock exchanges, such as the London Stock Exchange. In their place mainframe computers took over the task of matching buyers and sellers of securities. Today only a few physical exchanges, with tumultuous floors and face-to-face execution of trades, remain, the New York Stock Exchange and the Chicago Board of Trade being the most famous.

The same evolution has begun in bond and currency trading at banks. Many banks began to post the prices of the most liquid securities, beginning with Treasuries and mortgage-backed bonds, on computer screens, and then allowed their clients access to the screens. That way they could execute trades themselves, without the need of going through a salesperson like Esmee. Normally traders like Martin post prices on these screens for a limited size, say $25 to $50 million, and these will be executed electronically by clients; but for bigger trades, like DuPont’s, clients still prefer to call their salesperson. Nonetheless, many people within the banks think the flow traders are dinosaurs, and will eventually go extinct.

Perhaps the greatest threat facing the human trader, though, comes from computerised trading algorithms known as black boxes. Life for many traders has always been nasty, brutish and short, given the vicious competition between them. Survival has depended on their relative endowment of intelligence, information, capital and speed. But the advent and insidious spread of the black boxes has begun squeezing humans out of their ecological niche in the financial world. These computers, backed by teams of mathematicians, engineers and physicists (‘quants’, they are called) and billions in capital, operate on a time scale that even an elite athlete could not comprehend. A black box can take in a wide array of price data, analyse it for anomalies or statistical patterns, and select and execute a trade, all in under 10 milliseconds. Some boxes have shaved this time down to two or three milliseconds, and the next generation will operate on the order of microseconds, millionths of a second. The speeds now dealt with in the markets are so fast that the physical location of a computer affects its success in executing a trade. A hedge fund in London, for example, trading the Chicago Board of Trade, lags at least 40 milliseconds behind the market, because that is the time it takes for a signal, travelling at close to the speed of light, to travel back and forth between the two cities while a price is communicated and a trade executed, and the delays added by routers along the way mean the actual time is considerably longer. Most companies running boxes therefore co-locate their servers to the exchange they trade, to minimise the travel time for an electronic signal.

Many of these boxes are what are called ‘execution-only’ boxes. This type of box does not look for trades, it merely mechanises their execution. At this task, boxes excel. They can take a large block of equities, for example, and sell it in pieces here and there, minimising the effect on prices. They test the waters, looking for deep pools of liquidity, a practice known as pinging, just like a sonar searching the depths. When they find large bids hidden just below the surface of existing prices they execute a block of the trade. In this way they can move enormous blocks of stock without rippling the market. At this trading exercise, boxes are more efficient than humans, faster and nimbler. They do what Martin did when he pieced out of the DuPont trade, only they do it better. Many managers have started to ask why traders spend so much time and effort executing client trades when a box could do it just as well, and never argue over its bonus.

Other boxes do more than execution: they think for themselves. Employing cutting-edge mathematical tools such as genetic algorithms, boxes can now learn. Funds running them regularly employ the best programmers, code-breakers, even linguists, so the boxes can parse news stories, download economic releases, interpret them and trade on them, all before a human can finish reading even a single line of text. Their success has led to an exponential growth in the capital backing them, and boxes already make up the majority of trading by volume on many of the largest stock exchanges; and they are now spreading into the currency and bond markets. Their growing dominance in the markets is one of the most significant changes ever to take place in the markets. I, like many others, find the markets increasingly inhuman, and when I trade now I often have difficulty catching the scent of the market’s trail.

Human traders such as Martin are therefore in a fight for their lives. Unbeknownst to outsiders, every day a battle rages up and down Wall Street between man and machine. Some informed observers believe human traders have had their day, and will meet the same fate as John Henry, the legendary nineteenth-century railway worker who challenged a steam drill to a competition and ended up rupturing his heart.

Others, however, note with optimism that human traders are more flexible than a black box, are better at learning, especially at forming long-term views on the market, and thus in many circumstances remain faster. Evidence of their greater flexibility is found when market volatility picks up after some catastrophic event, like a credit crisis. Then managers at the banks and hedge funds are forced to unplug many of their boxes, especially those engaged in medium- and long-term price prediction, as the algorithms fail to comprehend the new data and begin to lose ever increasing amounts of money. Humans quickly step into the breach.

Something much like this occurred during the credit crisis of 2007–08. Anecdotal evidence and published fund performance statistics give us something like the following scorecard: in high-frequency trading, humans and machines fought to a draw, both making historic amounts of money; in medium-term price prediction, in other words seconds to minutes, humans pulled slightly ahead of the boxes, as flow traders made record amounts of money; but in medium- to long-term price prediction, minutes to hours or days – the boxes engaged in these time horizons are known as statistical arbitrage and quantitative equity – humans outperformed the boxes, because only they understood the implications of the political decisions being made by central bankers and Treasury officials. Thus, in what may have been the first major test of human versus machine trading, humans won, but only just. And so it is that this futuristic battle ebbs and flows.