How the first stock exchange programs were created

In the 1920s, the procedure for stock exchange trading changed dramatically. Brokers and traders began to gather in the stock exchange “pit” and make deals on the purchase and sale of goods, currency and securities there, looking at a large board on the wall, where their offers, quantities and prices appeared. In this way, they saved time on looking at ticker tapes, and then typing and sending their orders on a telegraph ticker. Minutes and even tens of minutes spent on this, in their business, could sometimes cost a lot of money. They gathered at the exchange literally in a pit with a sloping floor so that everyone could see the board. With shouts and conventional gestures, brokers let the “conductor” of this concert, the clerk on duty at the exchange, know exactly what they were ready to buy, sell and how much.

Nowadays, this picture can be seen in the newsreels and films of the last century. In stock exchange slang, it was called “open outcry”, that is, “open scream”, which would probably be the most appropriate literary translation into Russian. And this picture is so impressive that it will probably not be erased from the public consciousness for a long time as a symbol of stock exchange trading. In the 2000s of our century, when stock exchange trading was already largely computerized, an interesting publication of memoirs of very old brokers about their youth in the “stock exchange pit” appeared on the Internet, who agreed that on the New York Stock Exchange, although the ears were blocked by the screams of competitors standing nearby, everything was generally decorous and noble, while on the Chicago one, as if by accident, they could even bruise their sides.

There had been attempts to create a “Big Board” that “programmed” stock exchange trading in the open outcry mode before. The first of them was Samuel Laws' electric indicator with a two-sided dial, which he placed on the windowsill of his Gold Exchange so that its dials were visible both from the exchange and from the street. But it was problematic to see the numbers and letters on it without pressing your nose to the window glass. Therefore, they simply used a slate. However, this made no sense, since while the stock exchange clerk was copying the ticker tape onto it, the broker in his office had time to read it and “tap out” his order on the same ticker telegraph apparatus.

It was not until the advent of the cinematograph that a clearly visible public broadcast of ticker tape became possible, and American Lux Products adapted its reverse movie projector (that is, one that projects an image onto the back of a screen, allowing a movie to be watched in a bright room) for this purpose. In 1923, it demonstrated its first ticker tape projection system, the Movie ticker, on the New York Stock Exchange, where the letters and numbers of the ticker tape ran across a dark silk screen. In 1925, this private company went public on the same New York Stock Exchange, changing its name to Trans-Lux Co., and received several patents for improvements to its projector (see, for example, U.S. Patent No.1771499 1930 with priority from March 6, 1929).

In parallel with the projector version of the “Big Board”, its electric version was developed, more informative, since it did not just read the ticker tape. The data on quotes from it could be placed on a large screen in a more visual tabular version. Usually, historians of stock trading cite as an example the first such board “Big Board” designed by Robert Dane, the inventor and founder of the Teleregister Corporation, which he actually created in 1928 for his invention of the stock exchange “Teleregistrar”.

But his US patent no.1658516 (with priority from September 17, 1927) for “Method and device for indicating price offers or other items” is interesting precisely because in it Mr. Dane generously cites the patents of his predecessors, and not only American ones, but also English, French, Swiss. The problem, as they say, was ripe and was solved by the collective efforts of inventors in different countries. Incidentally, Robert Dane filed the next patent application for “Bulletin board or similar device” in January 1828 together with the Swiss engineer from Neuchâtel René Guier (patent no.1872126). Then in February 1928, an application was filed by Teleregister Corporation engineers for another patent (No.1890878) for the “Electric Indicator System.” And here is the next application for the “System and Device of an Information Board” for stock exchange quotations (US Patent No.1972341 was submitted in May 1930 by engineers from the Western Union Telegraph Company, which, of course, could not pass by such a thing.

Before the start of World War II, there were already more than a thousand of these large electric screens with a diagonal of up to 3 meters on exchanges and more modest sizes in brokerage offices. And in the 1940s, they were replaced by electronic boards. But it is probably worth mentioning another curious patent from the 1920s (US Patent No.1931091 (October 17, 1933, with priority from August 1929) – to the “Talking stock quotation board” of Clyde Smith of Communications Laboratories, a subsidiary of International Telephone & Telegraph (ITT).

As Smith, an engineer, writes in his application for his invention, it “relates generally to an automatic audible price announcement system and, more particularly, to an interactive stock market quotation board. In a broad sense, the invention includes the provision of a price indication system associated with an automatic stock quotation board for loudly announcing the name of a stock, the number of shares sold, and the price of the stock. Another object of the invention is to provide a plurality of intercoms for servicing multiple agencies through the operation of a single announcement system.”

The bank of voice notifications of the talking board Smith consisted of pre-made records of goods and prices, from which the necessary phrase was collected with the help of a rather ingenious electromechanical system and sent to the loudspeaker. It is clear that such a talking robot had no future in those years, especially since it was about something sacred – money. Nowadays, such robots exist, but they are also far from perfect. However, there was enough noise on the exchanges even without a talking board of quotes, the screaming there continued even after the boards became electronic in the second half of the 1940s. It began to gradually subside only in the early 1960s, when two fundamentally new systems for delivering stock exchange quotes to brokerage firms were launched – Quotron I and Stockmaster, which could constantly update the data in their magnetic storage device, which was recorded there from the ticker running line.

Quotron I inventor Jack Scantlin explained the invention in his May 18, 1960 patent application for “Security Quotation Devices”: “It is an object of the invention to provide a device that can accept input data from one or more sources and that can accept data arriving in a random sequence. It is a further object of the invention to provide such a device in which the stored data can be revised from time to time as re-allocation data is received. Another object <…> of the invention is to provide a data storage and distribution device that utilizes a single large storage unit at a central location, which storage unit is accessible to a large number of users upon request by telephone or other means of communication.” U.S. Patent No.3082402 Jack Scantlin received the contract for this device on March 19, 1963, when it was already being sold in hundreds and then thousands in the New and Old Worlds under the trade name Quotron I, and it already had a competitor – Stockmaster,

In his patent application, which the inventor of the Stockmaster, Robert Shinn, filed with the U.S. Patent Office on May 5, 1961, a year after Scantlin, formulated the essence of the invention in very similar words, as if he and Scantlin had written their applications from scratch: “A data storage system is provided for storing information on stock transactions in a form from which the information can be readily selected by an operator at a stockbroker’s office or similar location. Once selected, the information is made available to the operator. A system is provided for maintaining the stored information in an updated form.” Of course, Shinn’s “Data Storage System” (Stockmaster is its trademark) was “a new and improved magnetic drum storage and recording system <…> with a new and improved digital transmission and control system” (U.S. Patent No.3286235 (November 15, 1966, with priority from May 5, 1961). What else can you get a patent for, if not an improved version of something that already exists? Perhaps a cheaper version, but patents are not issued for that. And Shinn really improved it, and improved it significantly.

Scantlin's Quotron I indicated only the last price of a stock, while Stockmaster indicated the opening price of a position, the maximum and minimum for the day, and the volume of shares in the order book. This data could be calculated on a piece of paper, and an experienced broker could estimate in his mind the tendency of the bid/ask spread to increase or decrease, or, in simple terms, the difference between the lowest price asked for an asset and the highest price asked, that is, to determine the most liquid assets. And Stockmaster was immediately noticed in the circles of stock brokers and traders. In response, Jack Scantlin's company Scantlin Electronics (SEI) created for Quotron I owners a network with a memory server of four 12-bit CDC 160-A minicomputers that did the same thing.

But if we don’t go into details, both of these office-sized devices, desktops as they would say now, were essentially the first programmable ticker machines in the modern sense of the word. Only they didn’t have screens yet, electronic screens were still too expensive and would have greatly increased their cost price and, accordingly, the market price of these already expensive machines. They produced a ticker tape, and monitors for such devices appeared later, in the 1970s–1980s, when the PC era began.

These early 1960s tickers were programmed to track every exchange transaction and make the appropriate adjustments to their database, rather than simply printing a new tape each time, like the classic 19th-century tickers. Years later, engineer Howard Beckwith of Jack Scantlin’s team recalled, “Multiprogramming, multiprocessing, batch control, time sharing—we didn’t think in those terms, and most of us had never even heard of them. But we really believed we were breaking new ground; and it was that belief, more than any other factor, that made the work exciting and the time fly.”

And while engineers were enthusiastically constructing hardware compatible with computers and simultaneously mastering the basics of a new profession of programmer, their colleagues from the computer industry, already skilled in programming, were mastering a new area for them, creating software for stock exchange trading. And here two main directions of programming immediately emerged.

The first was aimed at depriving brokers in the stock exchange pit of a time head start of at least a few minutes or even seconds compared to brokers and traders “remotely”. After all, in a couple of seconds you can easily raise your finger in a conventional gesture and shout: “I'll take it!” Ideally, information from all world exchanges should be distributed in real time. It was also desirable to pick it up, sort it, bring it together in one place and show it in a summary table with the minimum delay. Computer technologies even in the 1960s showed that such a problem was quite solvable, and its solution was just around the corner.

The second hypostasis of stock exchange programming looked more abstruse from the outside, but it was precisely in it that software solutions were easier to find, because it was pure mathematics, that is, a chain of calculations, which is what computers were initially, essentially, designed for since the time of Charles Babbage. It was about the algorithmization of stock exchange trading, that is, buying/selling not on the subjective intuition that an experienced trader generates from a natural neural network in the “black box” — his brain, but in strict accordance with the objective laws and formulas of mathematical statistics and probability theory.

The first mathematical model for such stock trading was created in the late 1940s by Richard Donchian, a securities analyst (many self-respecting brokerage firms kept such specialists on their payroll, paying them quite decent money – as long as, of course, their advice brought profit). This model was called the “Donchian Channel” and was a chart of three curves, built on the basis of calculating a moving average. The upper line reflects the dynamics of the highest price of the traded asset for a fixed period of time. The lower line – its lowest price for this period. Inside the banks of this channel float other prices and the middle curve passes, which represents the average value of these two extreme values.

A simple picture at first glance, but to an experienced eye it spoke volumes. The configuration of the upper curve shows the degree of the market's “bullish” sentiment towards the asset, the configuration of the lower curve shows the “bearish” sentiment. The width of the channel (the area between the upper and lower curves) is an indicator of the asset's volatility. And most importantly, you can set a stop-loss order (an order to close the deal) slightly below the lower curve if you hold a long position (you are playing for an increase), and slightly above the upper one if you hold a short position (you are playing for a decrease). This way, losses are minimized when the Donchyan channel is broken. And the place to set an order to fix the profit is also clear: for cautious traders, slightly above the middle line, for risky ones, on the upper line.

But it was only possible to calculate the formulas and draw the Donchyan channel manually on paper using a slide rule and a calculator, which was quite a job, long and fraught with trivial arithmetic errors. Now it could be programmed on a computer. Just as it became possible to program the calculation of an investment portfolio using the formulas of Harry Markowitz, who published them in 1952, and in 1990 received the Nobel Prize in Economics for them.

In the 1960s, a young mathematician Edward Thorp, under the watchful eye of Professor Markowitz, was the first to use a computer for so-called arbitrage trading. In economics, arbitrage is the extraction of profit from the difference in prices of identical or related assets on different markets (exchanges) or on the same exchange at different times. A high-speed computer gave Thorp, who, by the way, worked at a hedge fund, the opportunity to calculate the profitability of a deal and buy/sell while other traders were considering whether it was worth doing. The effect was stunning, Thorp was pompously called “the mathematician who hacked the stock exchange.”

Later, in the PC era, many applications for optimizing stock trading were launched that were different from traditional time series analysis. And in the same 1960s, the New York Stock Exchange acquired its own computer programmed with the Market Data System I algorithm for automating quotations and trade reporting, providing 100 thousand transactions per day. In 1967, the Instinet trading system appeared on Wall Street, which was conceived as a digital alternative to the New York Stock Exchange, providing communication via computer channels between banks, funds and other financial institutions.

In February 1971, Bunker Ramo Corporation introduced the first version of the National Association of Securities Dealers automated quotation system — NASDAQ. Today, this is the name everyone knows about the stock exchange, the second largest after the New York Stock Exchange and the first to offer its clients trading exclusively online. It used the Intermarket Trading System (ITS), an electronic network that linked trading platforms and allowed communication and trading between them in real time. Only the classic exchanges were not yet allowed to use ITS.

But in the same year of 1971, the board of directors of the New York Stock Exchange reviewed a report prepared for them by William McChesney Martin, Jr., former chairman of the Federal Reserve System, member of the board of directors of the Rockefeller Foundation, etc., etc., who himself was the president of this exchange in 1938-41, in general a highly respected financier, in which he wrote that electronic communications “offer the means for radically improving the functioning of the markets.” Not immediately, but eventually common sense prevailed. In 1981, the New York Stock Exchange allowed the NASDAQ automated execution system, which processed securities listed on the New York Stock Exchange, to connect to the ITS. In 1984, the world's main exchange introduced the “Super DOT” system, which allowed orders to be sent electronically. And then, as they say, it all started.

Software for various exchange instruments and different types of exchanges poured out like a horn of plenty. There is no point in listing its varieties. Those who are, as they say, in the know, know everything themselves. And for those who are simply interested, it is easier to read the book by Michael Lewis “Flash Boys” (2014), where the author described in simple language without IT terminology, perhaps, the most striking episode in the recent history of exchange programming – the creation of a high-frequency trading algorithm. It is available in Russian translation online, although so far only for money.

As for the general trend, economics is far from being an exact science, the human factor is too important in it, and programmers, faithful to the precepts of the professors of mathematics Donchyan and Markovits, strive to minimize this factor in stock trading. So far, they have succeeded to a certain extent, and perhaps in the future, stock trading will really be subject to exclusively strict mathematical formulas. Only one goal is definitely unachievable in principle: an algorithm for guaranteed profit. This is prevented by the most fundamental law of our universe – the law of conservation of Mikhail Lomonosov, which in this case sounds like this: if someone has more stock assets, then the other will lose exactly the same amount of assets.

Drawing for registration of 5 trademarks.

Useful from Online Patent: