When Gottfried Wilhelm Leibniz arrived in the English capital for the first time in 1673, a crowd of people was already awaiting him at London Bridge. They were, however, less curious about the German scholar himself, than about any information regarding the ongoing Dutch-English war which he may have picked up during the course of his journey. As Leibniz was still reporting on what he had heard, a man broke loose from the crowd and was immediately surrounded by a couple of barefooted, jumping boys:

This fictional scene, taken from Neal Stephenson’s book Quicksilver, may appear like some kind of game to an uninformed observer but what we witness here is driven by economic intention. Valuable information is being exchanged for currency at the London Stock Exchange. Leibniz had indeed reported gunshots, yet he had also added – long after the boy had run off – that they were not actions of war but rather “coded data, speeding through the fog so opaque to light but so transparent to sound.”

On March 2nd 1791, about eighty years after Leibniz’s death, the Frenchman Claude Chappe invented the first functioning system to optically transmit complex messages over a distance. “Si Vous reussissez, vous serez bientot couvert de gloire.” (“If you succeed, you will be covered in glory.“) was the text of the first message to be sent. The apparatus was named the ‘télégraphe’. By 1830 there was already a network of roughly 1000 optical telegraph towers across Europe, allowing messages to be carried from Paris to Amsterdam, or from Brest to Venice. At its peak in 1850, the French system alone involved 29 cities, 556 stations, and spanned across roughly 4800km. Good visibility was critical to the functioning of the system. If it was foggy, nothing could be reported.²

Today it is common practice to conduct business transactions under obscured conditions. Trading algorithms are constantly racing through global networks of underground cable systems. Stocks of varying trade value are bought cheaply in one commercial venue and sold at a higher rate in another until the differences equilibrate. Sometimes, the interdependencies between stocks are direct: if the price for crude oil increases, it is very probable that the value of oil companies will do so as well. If an algorithm wants to earn money, it has to be quicker than its fellow algorithms. Easy thus far. Things get more complicated if the algorithm not only runs as fast as possible, but watches out for rivals and attempts to hide itself from them, all while trying to find out their patterns of behavior.

Digital trading algorithms have long been more than just simple sets of rules for buying and selling stocks. They scan their environment for correlations – driven by the urge to adapt, they become quicker and more elaborate.

We find them beautiful while sensing that with every movement of the line millions of dollars change hands.

Some processes are only understood after they have ceased to function properly – or have collapsed. The international race of trading algorithms, usually invisible to laymen, came to light on May 6th 2010, in the form of a ‘flash crash’. The algorithmic sale of 75000 so-called “E-mini” futures triggered one of the most drastic drops in the history of the American stock market. Within minutes the market had recovered, but for the first time, high-frequency trading was linked to a stock market crash. Even though the event itself had only lasted a few minutes, numerous private and state actors were kept busy for three years to follow while trying to figure out the cause of this phenomenon.

Not all unusual trading activity is immediately visible. When algorithms trade, they make the market oscillate in millisecond staccati. It is therefore quite common that such unusual activities pass unnoticed at the time of their occurrence. The company Nanex has specialized in scrutinizing historical trading data. As part of these investigations they discover and collect unusual examples of curious algorithmic trading sequences which they name and present on their website: Castle Wall, To the Moon, Alice!, Sunshowers, Broken Highway…

The graphs remain inexplicable for now. If they fail in their attempt to provide information they may still act as dadaist poems, where form itself comes to the fore: the rhythm, the shape, the color. Guided by their titles, the graphs become pictures – frenzied dashes slowly form castle walls. Only, that there are no walls, they are merely a product of our perception, our desire to understand. Maybe, the graphs give us an inkling of an algorithm’s world. Where, driven by their urge to succeed, they pace through the fog; nightly birds become thieves and monsters, transformed by their paranoid gaze. Just like us, they create order and coherence, images and delusions. If today, a human and an algorithm were to lay in the grass and to look up into the sky, they both would likely see faces where there are only clouds. A market crash–it seems–is not only a financial but a semiotic accident.

We regard the graphs but don’t understand them. We find them beautiful while sensing that with every movement of the line millions of dollars change hands. What these images make visible is a cascade of events deciphered in retrospect. Tempted by their beauty, this book as tried to amplify their effect. Stripping away all information which may point to context, time, value, what remains, are abstract images which present only color, expanse, and limitation. On several hundred double pages, 75000 Futures collects these species as an atlas of curious phenomena. Identification or classification are not paramount, this atlas mainly serves to exhibit the life of algorithms.

We regard these graphs once more and now they speak differently to us. We no longer see activities or functions in these images, rather we find landscapes, tools, journeys. We find 75000 futures.