Perfect example. It is a known logical fallacy called Normalization Bias. Because something hasn't happened before (amateur politician winning presidency or mma striker winning in top shelf boxing) means it can't happen. So many experts and polls and conventional wisdom was wrong about Trump's chances. Yet the evidence was there. Polls showed Hillary only 2 points ahead, and polls also suggested that Trump supporters would be hesitant to admit they were voting for Trump.
If you look at the 11 most common fallacies, many sherdoggers in this thread have made the same mistakes. When logic clearly should make us know better. We know that boxers degrade rapidly going into their 40's. We know that Conor has dispatched opponents with true ease striking, and that he was competitive with IBO champ in sparring. We know that, objectively, Mayweathers speech patterns have changed, which is a red flag. We know that despite Normalcy Bias, oddsmakers still give Conor a reasonable shot at winning. We know that Conor will only improve in his boxing leading up to the fight with Floyd. We know Floyd has problems/injuries with his hands, and started already to decline with a couple of majority decisions later in his career. We know Conor is unorthodox and the 3 most problem opponents for Floyd were southpaws.
People resist change, and fight back with vigor against change to the status quo, and conventional thinking is that an mma fighter can't strike with a top boxer. Objectively, even Joe Duffy trying boxing late in his career compiled a 7-0 win record against reasonable competition.
The signs are all there for an upset. Just like Trump beating Hillary, the signs are there. With Trump election, people laughed out loud at Trumps chances, including those who should know better, even President Obama laughed. Normalcy Bias is one of the most powerful of fallacies, blinding people to the obviousness of the truth.