Artificial intelligence in sports betting

Era Eldorado (1990s)

Let us remind you that in the 1990s, when the Internet was still developing, the so-called "Eldorado" prevailed in bookmaking establishments. Bookmakers did not have a system that would directly compare their offer to each other, which resulted in so-called "arbitration situations". You could bet one bookmaker on one event and the other on the opposite event and regardless of the result you could earn a certain percentage of the sum of the bets.

Over time, software was developed to compare odds (BetRadar 2001) and integrated with the systems of individual bookmakers so that they received notifications about the occurrence of an arbitrage situation. Players who used this method were limited as to the size of their stakes, which in the era of accounts set up via the Internet was not a problem, and in ground premises there were similarities of "who not to serve". Some players set up accounts with others in order to continue to benefit from this proven strategy. However, this is becoming increasingly difficult.

Mathematicians gang and Hole-in-One gang

Some players preferred to take a little risk rather than enjoy a certain profit. Using AKO coupons or system bets they could gain a huge advantage over the bookmaker by putting aside those matches where they mistakenly estimated odds. In this way, the so-called "Mathematicians Gang" almost smashed the Toto-mix in the UK the "Hole-in-One Gang" smashed the bookmakers into over £500,000 betting on golf.

Era Big Data (2001-2015)

With the development of technology, the computational power of computers has developed, and so, for example, in 1997 Garry Kasparov was defeated in a chess match by the supercomputer Deep Blue.

With the advent of the Big Data era, bookmakers and players were given access to a huge amount of data that could be processed statistically. This was not yet an ideal approach, but relying on raw data gave a very solid basis as a starting point for estimating the chances of victory of individual teams / players.

The first programs of artificial intelligence for bookmakers appeared already in the years 2000-2010, of which I liked Odds Wizard the most.

They were the first step towards gaining an advantage over the bookmaker. Some of the projects have failed, but the ones that survived are worthy of interest.

The Era of Artificial Intelligence (2015-)

DeepMind was founded in 2010 and acquired by Google in 2014. In 2015, DeepMind took advantage of its enormous computing power to create the AlphaGo Zero project, which uses artificial intelligence to defeat Europe's champion Fan Hui in the game "Go". Go is considered more complicated than chess.

In December 2017, the AlphaZero project discovered StockFish 8's most powerful chess engine, based on its experience with AlphaGo. Although the StockFish engine had quite limited RAM, there is no doubt that in a relatively short time it was possible to create a very powerful tool that performs equally well, and even better than tools that do not use artificial intelligence.

In 2017 another Artificial Intelligence project - Libratus - defeated the best poker players in the world!

In 2019 bookmakers still use Big Data to estimate odds, but they have weaknesses because they do not take into account "unusual" factors, such as weather, suspicion of an important player's injury, etc. The bookmakers are not able to estimate the odds. However, they are very good at dealing with typical matches.

Artificial intelligence can have problems with learning unusual patterns, although it is possible with good approximation. In 10-30 years perspective I think that there will be algorithms taking into account such factors and professional players will have great difficulties to find a value against a bookmaker using artificial intelligence.

However, there will remain a betting exchange, where everyone can act as a bookmaker and still earn money on betting. You just have to keep up with the "arms race" to adapt to the new conditions on the battlefield.

Summary

We live in a world of rapid technological progress. The development of artificial intelligence has never before been as fast as it was after 2015. According to the ideas presented in the Factfulness book, we are witnessing big changes that are progressing slowly. The world and knowledge from 10-20 years ago is already outdated.

It is only a matter of time before artificial intelligence will be applied to betting, chess, poker and in everyday life, e.g. vacuum cleaners, and maybe in the future humanoid robots, whose vision was presented in the excellent film "I, robot".

The use of artificial intelligence in life is only limited by our imagination. Fortunately, Google did not agree to use it for the purposes of the Pentagon. This will slightly delay the development of artificial intelligence towards more effective killing.

If you know how to use artificial intelligence, it is probably high time to use these skills. If not, perhaps it's worth learning more?