Why You Should Be Wary of Supercomputer Sports Betting Picks

Technology is going to change sports betting. That statement is unequivocal. Already, the use of data and algorithms has ameliorated many sports betting strategies, and it is only going to become more prevalent as technologies like AI and machine learning become more prevalent. Even today, you will see websites and social media accounts offering betting advice based in AI and supercomputer predictions. Some of it is useful for sure, although we would suggest to be careful and watch out for scams.

In general, though, we see a trend toward algorithmic and AI picks for sports betting. As mentioned, they can be useful, but only up to a point. The best frame of mind to have is that they can be a companion to sports betting but not a substitute for your own intuition and experience. It doesn’t matter if it is horse racing betting for a big race like the Kentucky Derby or intricate props betting on a college football match; a blend of data analysis and your subjective views will serve you better than leaning too heavily upon either strategy.

Crunching numbers only goes so far

To explain, it’s worth noting how computers analyze data compared to humans. An algorithm, for example, is simply a formula trained on a data set. Artificial intelligence, for all its intricacies, is basically the same format – training on a data set. AI does not reason or think; it is simply a formula. Tools like ChatGPT do not understand what they are saying to you; they are simply “predicting” a bunch of words in relation to each other. You need to keep that in mind when you are told of the omnipotence of supercomputers.

Again, that’s not to say they are not useful. For sports betting, they are able to analyze vast amounts of data, much of it unintelligible to the human eye and unattainable for human endeavor. As such, they can tell you in seconds that the Patriots are 13% less likely to win if it’s raining or that the Warriors score 3.5 pts per game less when Steph Curry is injured. This is what we call structured data. It is data that a machine can understand.

Unstructured data is not so easy for AI to understand. What is unstructured data? Essentially, it is stuff that cannot be easily boiled down to raw numbers. It is what you see with your eyes in a sports stadium. For example, when a player plays badly but still manages to put up good numbers or when a horse simply doesn’t look right despite winning. As humans, we can analyze this stuff, whereas computers cannot. You, as a bettor, should combine what you see with the structured data of quantifiable stats.

Sports betting is about going against consensus

One other area of consideration is this: Some of the machine learning that dictates sports predictions is often based on the same concept that the sportsbooks use themselves for odds setting. What’s more, you will see articles about supercomputer predictions for a sports event, but you will find that the data is produced by sports betting affiliates, not geniuses from Google or IBM. The reason that this is important is that it builds a consensus and that consensus is reflected in the odds.

And valuable sports betting is about going against consensus. Anyone can tell you that the hot favorite has a good chance of winning the Preakness Stakes or that the Kansas City Chiefs may repeat in 2025, but it’s about seeing what others miss, including sportsbooks in their odds-setting. That takes a mixture of structured data analysis and the evaluation of unstructured data. Supercomputers will get better at the latter, eventually. But for the moment, there is a limit to what they can do.