Stuart Gentle Publisher at Onrec

Will Algorithms and AI Put Betting Tipsters Out of a Job?

If you have a little bit of experience with sports betting, say beyond an annual flutter on the Grand National, then you quickly learn to be sceptical about anyone offering guaranteed winning tips.

There is always a sense of the snake oil salesman when you see posts on social media offering betting systems and unbeatable tips.

But on the other side of the coin, there is a bubbling industry of sports tipsters, ranging across all sections of the media. From newspapers to YouTube, Facebook to specialist sports websites, you’ll see plenty of columns and videos from tipsters simply plying their trade. The best are honest about their success and failures, and punters tend to appreciate a “show your work” attitude over someone who claims to have a crystal ball.

And yet, we wonder whether the columns and videos will disappear one day soon. As with many industries, sports predictions and betting analysis face challenges from digital alternatives. Algorithms that can analyse millions of data sets, and computers that use AI to make predictions smarter and faster, are coming to the world of sports.

Data already embraced by bettors

There is already a bigger focus on data within sports betting. If you visit a top UK sports betting site, you will find loads of statistics and data-based predictions as part of the service. But in the auxiliary media that complements sports betting, there has always been a role for the human element – the sagacious tipster.

Over in the United States, for example, the world of Daily Fantasy Sports (a type of sports betting for all intents and purposes) is hugely popular, with around 60 million players. IBM, however, is trying to get in on the game, promoting its famous supercomputer WATSON as the answer for sports predictions. The company has partnered with ESPN to create an app offering fantasy sports predictions.


And yet, we should say that there is a long road ahead for the AIs. WATSON has had some success, but there are also some flaws. Other AIs and algorithms, too, have had mixed results. We know that computers can analyse vast amounts of data, but there is sometimes something lacking in how data – and other elements – is perceived by computers.

To illustrate, we will give two scenarios:

Scenario A: England Captain Harry Kane will have a high-scoring season in the Premier League. This is based on his record from the previous season when he won the Golden Boot.

Scenario B: England Captain Harry Kane will have a poor season. This is due to the fact he has been subject to transfer speculation regarding a move to Manchester City, which has caused a breakdown in his relationship with Tottenham Hotspur fans.

As you might expect, Scenario A is best understood by computers. It’s only a rough example, but the computer could crunch millions of bits of data to arrive at an analysis of Kane’s prospects. This is called structured data. Scenario B, however, is better understood by humans. It’s called unstructured data. The latter scenario is more speculative, of course, but it’s where the shrewd tipster comes in, marrying the data with common sense.

The problem for tipsters, however, is that computers are improving all the time, and that means evolving to better understand unstructured data. Perhaps ironically, supercomputers like WATSON can scan 1000s of news articles, which includes columns that give betting and fantasy sports advice. So whereas the average punter will read one column previewing a game, the AI can read an infinite number and arrive at an aggregate from the analysis.


Better to embrace digital alternatives

Whether it’s through traditional journalism or social media and affiliate programmes, giving tipping and betting advice is, of course, only a very minute part of the labour market. But it’s just another example of how machine learning poses a challenge to job roles in certain sectors. It won’t happen overnight, but computers will better understand the “common sense” side of sports predictions.

But it does not necessarily mean those pundits will be out of a job. As with all types of digital disruption, it’s arguably better to embrace it than try to fight it. Already we see websites linked to sports betting advice offering algorithm-based betting predictions.

Will the tipsters’ job disappear one day? It’s hard to say. If the science of sports betting prediction was perfect, then everyone would be doing it. It’s not perfect, and it probably never will be. But the computers will outstrip humans’ ability to analyse sports, and some might claim they already have. Anyone who plies their trade making betting predictions will be better off if they embrace, rather than dismiss, this new way forward.