Mercurius is an Italian fintech startup that aims at assetizing sports betting markets through the usage of artificial intelligence and machine learning technologies. Founded in 2018 it released Tradr in 2019 delivering positive results to its users since then.

When there is a revolution in your vicinity, you have the choice to either stand passively on the sidelines or get involved and embrace the subsequent risk versus reward scenario using services like our value betting bot for Betfair exchange. Most revolutions are inevitable, as is the case with the use of Big Data and Artificial Intelligence’ (AI) in the worlds of finance and sports betting.

In this case, it is perhaps better to see it as an evolution, and it has already begun. CB Insight’ research found that banks invested $118 billion in FinTech between 2013 and 2017. This seemingly towering figure is only a drop in the ocean, however. According to research agency, Gartner, AI will generate $2.9 trillion in business value and recover over six billion hours of worker productivity per annum by 2021.

Whether you are involved in financial or sports trading/betting, the road towards profitability is the same. In both industries, it is all about good decision making. If you have access to an edge or information that others don’t have, you can make better decisions. In most cases, you gain an advantage by finding signals missed by others, or rapid analysis of a vast amount of data.

AI’s ability to crunch data without bias or emotion offers us a significant edge, as does the technology’s capacity to analyse billions of data points, seemingly in the blink of an eye. As vital as data analysis is in finance and sports betting, speed is even more crucial. The best market price, whether it relates to the cost of a stock or the odds available on a football match, doesn’t last long!

Artificial Intelligence in Finance and Big Data - Achieving a New Gold Standard

Before we proceed, it is a good idea to clear up any confusion about Big Data and AI; two terms often erroneously used interchangeably. While there is a close relationship, they are NOT the same thing!

Big Data is raw input that we must clean, structure, and integrate before it is useful. AI is the output, the intelligence we receive from the processed data. The pair work exceptionally well together. AI needs the information, and Big Data is of little use when you have no idea how to implement it.

Aside from being a defunct monetary system, the term ‘Gold Standard’ denotes something of superior quality when compared to other things in its field. When it comes to investing, nothing compares to the proper implementation of Big Data and AI.

Here are four ways that Big Data and AI are already changing the world of finance.

  1. Better Analysis 

    As we have just mentioned, Big Data won’t change your life if you have no clue how to use it. Two pieces of flint were once a revolutionary method of keeping yourself warm, but woe betide the poor caveman who didn’t know what to do with them!

    In the ‘old days,’ financial analytics involved price analysis and their behaviour. These days, you must also consider geopolitics and social trends. You can use Big Data in predictive models to estimate the likely return rate on an investment. Investors have a greater understanding of risk, which leads to more efficient decisions.

    One thing about Big Data is the fact it takes time to collate. In finance, seconds are precious, which may lead to certain traders eschewing the data in a rush to make their investment. Fortunately, an increasing number of investors understand the value of what Big Data provides them.

  2. Machine Learning - (AI) 

    As we have discussed, Big Data works best with AI and vice versa. Machine learning is the process of AI using past information to make better decisions in the present and future. While humans often make the same mistake more than once, a well-designed program makes the error once at most and learns how to avoid making it forever.

    Through this process, AI can analyse markets, social media, and other Big Data it comes across, to analyse market sentiment in a manner that is impossible for a human. It was long assumed that AI was incapable of ‘thinking like a human.’ However, recent developments in the field (as we discuss later with Go and Poker), have laid this myth to rest.

  3. Real-Time Analytics 

    Successful trading and investing require the following:

    • Finding the best value ‘bet.’
    • Acting at a specific time.
    • Making as few mistakes as possible.

    It is here that a combination of Big Data and AI works beautifully. With real-time analytics, you can use the tech to browse the markets to find immediate opportunities. Imagine a situation where you could gather, ‘clean up,’ and utilise an extraordinary amount of information to make a profitable trade. This dream is now a reality!

    Big Data will gather all the necessary details. Then, AI finds the best value proposition at the correct time. As the machine has no human fallibilities such as being led by emotion, it won’t make the same mistakes as you would.

  4. Risk Assessment 

    In finance, the analysis of risk is a foundation block. A prime example involves using a person’s Credit Score to determine if they should receive a credit card. With AI, a financial institution can quickly find data about a person’s:

    • Loan repayment habits
    • Number of loans currently active
    • Number of existing credit cards

    The machine then decides the level of risk and computes an accurate interest rate.

    As far as investing is concerned, you use the combination of Big Data and machine learning to determine the risk associated with any investment. Then, you can decide if the potential ROI is worth that risk.

Artificial Intelligence in Sports Betting and Big Data - Changing the Game

Today, the fact that the combination of Big Data and AI is primed for use in sports betting seems obvious. However, it is only recently, due to the extraordinary advances in technology, that we are finally in a position to offer you the first Artificial Intelligence for Betfair exchange.

As you probably know by now, AI algorithms are learning structures based on gigantic databases. When you ‘feed’ them the data, they continue to learn and have the ability to make increasingly accurate predictions. What you probably DON’T know is that we use a theorem that is more than 250 years old as one of the chief learning methods!

In the mid-18th century, Reverend Thomas Bayes used conditional probability to provide an algorithm. In statistics and probability theory, Bayes’ Theorem describes an event’s probability, based on existing knowledge of conditions that could be related to the event. His work was published in 1763, two years after his death.

Pinnacle neatly broke down the theorem into a simple formula:

"Probability of A Given B equals Probability of A times Probability of B Given A divided by Probability of B."

Sadly, it wasn’t until the implementation of computer technology that Bayes’ Theorem received the appreciation it deserves. One of the biggest enemies of any sports bettor is their adherence to a particular outcome even when the circumstances change.

Thanks to machine learning, we can now finally use Bayes’ Theorem in the manner in which it was intended. Instead of allowing emotions to run amok, AI uses the cold hard data you feed it to provide an accurate prediction. You can also use it to calculate the true probability of any outcome.

The Evolution of the Revolution

Throughout history, some men and women have found innovative (and sometimes illegal) methods of beating the market. Going back a century or so, merely having access to a calculator could have been the difference between success and failure!

There were some quaint and ingenious methods of earning a crust through sports betting. In the early 1900s in the United States, for example, the results of horse races were communicated from the track to the bookmaker via telegraph. In this scenario, the receptionist who received the initial message knew the outcome before the bookie, and could place a winning bet each time!

In June 1975, a trainer named Barney Curley organised what became known as the ‘Yellow Sam Coup.’ After ensuring that his horse had a favourable handicap rating, Curley sent associates around Ireland to place bets at odds of 20/1. Usually, such wagering action would bring down the price of the horse.

However, the course, Bellewstown, only had one telephone box. One of Curley’s team pretended to be on the phone to a dying aunt in a phantom hospital. As a result, the bookmakers were unaware of the betting situation! Fortunately for Curley, Yellow Sam won, and he profited to the tune of £300,000!

Multiple Regression

When Michael Lewis published Moneyball in 2003, it lifted the lid on the remarkable tale of how the Oakland Athletics baseball team massively overperformed on a tiny budget. In the book, Lewis outlines how the team, coached by Billy Beane, used analytics to create a high-quality squad that belied its financial strength.

Subsequently, teams in different sports started to adopt this practice. Soon, football clubs followed baseball’s lead in terms of analysing potential recruits using Big Data. The analysis extends to the training regimes of players, including what they eat, how much they rest, the optimal level of training, and even how they interact with fans.

Although Moneyball heralded a revolution in how player performance was measured and analysed, the strategy of using this type of data was almost a century in the making.

Karl Pearson is credited with creating the discipline of mathematical statistics. It is he who founded the world’s first university statistics department at University College, London, in 1911. He probably coined the term ‘multiple regression’ in 1908. Its purpose is to learn more about the relationship between several independent or predictor variables and a dependent (or criterion) variable.

In sports betting, regression analysis is used to establish the most crucial factors likely to influence the outcome of an event. Multiple regression is based on the premise that you must know the past to understand the future. For instance:

  • Liverpool beat Everton in six of their last eight games.
  • However, Everton recently beat Liverpool 3-1.
  • Naby Keita played in the most recent game (and the other one Liverpool did not win) and has never won against Everton.

In that situation, you could determine that Keita’s presence is the key variable, and decide to back Everton or lay Liverpool if Keita plays. This, of course, is a VERY simplistic example.

Bolton and Chapman attempted to use multiple regression in horse racing to determine how various factors impact the ability of each horse to win. Ultimately, the duo only researched 200 races and looked at nine factors. Despite having the idea in 1970, they only began in 1981 due to a lack of data, and it was published five years later. How times have changed!

The Midas Touch

The creation of the Internet enabled prospective researchers to gather data more easily than Bolton and Chapman. By the 1990s, traders were able to produce simple predictive models. While this practice was primitive by today’s standards, those who implemented it were extremely successful because bookmakers continued to lag.

In the intervening years, the bookies have been forced to invest heavily but still find it tough to keep up with those using the most advanced technology available.

In 2000, Cantor Gaming, a subsidiary of Cantor Fitzgerald, hired Andrew Garrood, a former trader for a Japanese bank. It proved to be a spectacularly good decision. Within a decade, Garrood helped develop the Midas Algorithm, which generated the best forecasts in the industry. Cantor Gaming became the 2009 market leader and defeated the betting syndicates that had previously run amok.

In many ways, it was a pivotal moment in the history of trader versus bookmaker. Professionals who were once on the fence about Big Data and AI understood that its implementation was essential if they hoped to win in the medium to long-term.

Sports Betting 4.0 - Beating the Bookmakers at Their Own Game

While we are not saying that winning against the bookies over a long period is impossible without technological help, it is becoming complicated. The main issue is that of account closures. High-quality bettors are either restricted to bets worth cents or kicked off the site altogether.

As a result, they are forced to seek the haven of betting exchanges where you won’t receive a ban for winning. However, in making this journey, you are pitted against hundreds of thousands of rival bettors, many of whom possess the benefits of AI.

In some ways, we can thank the growing video game industry for helping us to develop sophisticated algorithms capable of handling more data. The powerful graphics cards that keep up with video game demand offered a route to better and cheaper AI hardware.

We see the results in the astonishing performance of AI against human players in games such as Go and Poker. As the structure of these games requires an immense human component, it was long assumed that a machine couldn’t beat the world’s best.

That attitude changed when a program called AlphaGo thrashed Lee Sedol, recognised by many as the planet’s best Go player, 4-1 in a five-game series. Recently, a program called Pluribus made a profit of $48,000 after a marathon 12-day poker session against many of the world’s top players. It faced five different players each day across over 10,000 hands of Texas Hold ‘em.

Analysing Every Aspect of Performance

How long do you think it would take to try and work out practically every aspect of a football match, for example? The process involves finding out:

  • Every touch taken by a player.
  • The angle and position of each shot.
  • Which body part the player used to strike the ball.
  • How the goal was scored.
  • The likelihood of scoring a goal based on the factors above.
  • Expected goals data.

By the way, you also have to determine which data is useful and which is not! Next, you have to decide on the probability of an outcome, such as who wins the match. Finally, you have to calculate whether or not a bet is a value one or not! Alternatively, you can allow Big Data and AI to do the work for you, assuming, of course, that you can afford what is costly technology.

Betting AI Revolution is Here - Do You Want to Join?

As we said at the beginning, you can either get involved or stay on the sidelines. If you choose the latter, you are competing against machines capable of extraordinary decision-making and calculations at a speed that you can barely imagine.

You can try to invest the money in Big Data and AI, and learn as much as you can about the process. One final option is to invest in Mercurius, a company that already has the technology and a team that knows how to use it. We get our data from our partner, Wyscout, who collects an incredible amount of sports information.

We feed our sophisticated AI program this data, and it allows us to determine the true probability of an outcome. Then, we can decide if there is a value bet available.

Although high-quality professional gamblers will continue to find ways to win, it is nerds and their data that has changed the way we bet and invest. As the legendary bettor, Bill Benter said:

"It wasn't as though streetwise Las Vegas gamblers figured out a system; success came when an outsider, armed with academic knowledge and new techniques, came in and shone light where there had been none before."

We are in the midst of a revolution. Which side of the fence do you want to stand on?