From the books you buy online to the Club points you accumulate, algorithims drive modern commerce – thanks to smart minds that once cracked the Stealth's secrets, writes ADAM MAGUIRE
WHAT DOES a Hungarian physicist, who helped to break the potency of the US stealth aircraft programme during the Cold War, do next? According to entrepreneur and public speaker Kevin Slavin, they get a job on Wall Street. At a recent TED Global event in Scotland Slavin told the story of a chance encounter he had with a scientist-turned-algorithmic trading developer.
Early in his career this physicist had played a part in undermining the value of the hundreds of billions of dollars worth of US military research that made stealth aircraft a reality.
Stealth at the time primarily worked, in simple terms, by making an object appear to be many small things rather than one big one. As a result radar was not able to accurately distinguish a flock of birds from an F-112.
The Hungarians’ solution was to develop a machine that could pull in various readings to identify when the radar system was being tricked in such a way. If it found a flock of birds emitting electronic signals it would be fair to assume that there was something more to it.
Suitably impressed, Slavin said he asked the physicist what he did next; the answer was financial services. The physicist’s expertise had brought him to Wall Street to develop complex computations to help companies trade more effectively.
In many ways, as Slavin points out in his talk, he was still doing what he did in Hungary. He devises ways of identifying big objects – in this case share movements – that have been split into many small pieces to avoid detection. Perhaps most strikingly, however, the Hungarian estimated that he was one of around 2,000 physicists working on similar projects for Wall Street companies.
Many of these are likely developing for the other side of algorithmic trading, creating formulae to help split up big trades so they can hide amongst normal market noise. They have been sought out because of their extremely high proficiency in mathematics and the sciences. They all likely took the jobs because of the pay packets and bonuses on offer to those who can give their employers an edge.
However finance is not the only sector that holds algorithms in high esteem.
Many big companies, particularly those online, have complex calculations at the very core of what they do. Equally, such mind-boggling maths can often play a crucial role in making everyday things work, even where the end user may not realise it.
The following are five of the most important algorithms being used today:
THE SCIENCE OF GOOGLE'S SEARCH
Google has mastered the craft of search, making it the first port of call for anyone seeking answers online. How it finds these answers is based on a very nuanced algorithm that considers keywords, popularity, relevance, timing, location and many other factors.
Google’s constantly changing calculation is arguably the most valuable piece of mathematics in the world.
It is not only that it is at the core of the company’s ability to generate massive profits but it also has significant social value, not least because it dictates what users see when they go looking. Google’s algorithm has the power, perhaps unwittingly, to condemn any number of sites to irrelevance. Conversely it can also push others to prominence, guaranteeing them a constant stream of visitors and ad revenue of their own.
TESCOS ROLLING BALL
With thousands of outlets and millions of customers, deciding what to stock in a Tesco store is a delicate science.
To help in this the retailer utilises the skills of data company Dunnhumby, which has created a “Rolling Ball” algorithm around the chain’s ClubCard system.
The equation works by categorising customers’ purchases based on what they were purchased with. So if one thing in the trolley fits the “family” category, then the rank of all other items in that category increases.
There are currently 16 million Tesco ClubCard members and Dunnhumby analyses in excess of five billion pieces of information a week for the company. The resulting data helps to identify the types of customers Tesco has, as well as going towards dictating stock levels and even how products are stocked on shelves.
CORRECTING A COSTLY CREDIT CARD ERROR
Entering the wrong credit card number into an online payment site is a common mistake, but it is also a costly one for those processing the details. Without any way of identifying the error in advance each mistake would require the incorrect details to be sent on, rejected and then sent back. All of this takes time and IT power. To solve the problem, IBM researcher Hans Luhn created the Luhn Algorithm.
This relatively simple piece of maths takes each card’s number, reverses it, doubles every second digit and then adds everything together. If the result is divisible by ten, it is most likely a valid credit card number.
AMAZONS ARMY OF ALGORITHMS
As the default shopping destination of the web, Amazon utilises a number of algorithms to pull buyers in and get them spending money. For a start it has its own search calculation that, like Google, aims to give the most relevant results to requests and keywords. In addition to this it also pulls in data from other customers’ purchases and browsing habits to suggest related products.
However, it is not just Amazon itself that tries to take advantage of algorithms on Amazon.com; regular users selling their own items are even able to use them to set prices. This allows them to ensure that, for example, their book is always priced at a certain ratio below that of a newer version. In one notable case, however, it led to a book about flies ending up with a $24m price-point. The reason? Two sellers had set their separate algorithms to put the price of their book slightly higher than the other, leading to a pricing arms race.
A MODERN DAY ENIGMA
Algorithms are now the driving force of the cryptology industry around the world, with long equations taking the place of traditional encryption methods.
This is no longer the preserve of security companies and state agencies, either, as data encryption has become an important part of every digital medium.
How data is sent over the web, how information on ATM cards is kept safe and even how files are secured on a business computer all involve algorithms that are next to unbreakable. The use of calculations has spread into other areas of personal security too, particularly online. Algorithms are used to decide what emails will be marked as spam and what will be let through to the inbox, for example.
