If predictions are not only for Fantasy Hockey Leagues etc., where you only need to pick a winner, but also for Sports Betting, where you need to beat the odds, then if the analysis says that the home team has a 51% chance of winning, and the away team has a 49% chance of winning, but the home team's odds are so small and the away team's odds are much bigger, then the pick should be 'AWAY WIN'. Even though a home win is slightly more likely, in that scenario, both teams win almost every other time, but with a home win, you don't double your money and end up losing money in the long run, whereas with an away win, you more than double your money and make a profit in the long run. Gathering all this data should be quite easy with so much information available about team news, injuries, game schedules, statistics, history data, etc. Some of these might have APIs, and when all the information is gathered, AI can work its magic and analyze it all (maybe AI can also do most of the work in gathering the data using those sources as a Knowledge Base). Then, you can feed all that information to the AI and ask it to simulate today's games 1000 times with this data, and not only to pick the winners but also to compare the win probabilities with the current odds and suggest the best games where to put money to win the most in the long run (and also the games to stay away from because they don't offer any advantage this time). After the AI analysis, do your own manual check and send your final picks to your happy subscribers. Once again, great job @brief-accountant-68119 ⭐