Betting odds tend to be the best prediction method for 2 reasons; the first is there is no other good prediction method, and the second is that people who spend time doing analysis will bet on one side if that side’s odds are better than the market is giving it credit for. As an example, my stats professor in 2016 bet 1000$ on Trump, not because he wanted Trump to win, but because he saw the betting market was only giving Trump like 30% odds, whereas my prof estimated it was closer to 35 (or something). For the record, the polls had Trump at like 2% during this time.