Nate is a dingus of a commentator but 75% for Hillary was reasonable imo. It put her odds of losing the same as getting tails two coin flips in a row, which seems right given how narrow Trump’s margins were in PA, WI, and MI
Good ol red pilled Silver. Even if he were “right” any statistical model would change every day between now and November making today’s “odds” utterly pointless.
Besides, there is no real “right” in non-replicable experiments. It’s not like a coin or dice. If his model says Trump has a 65% chance of winning, he’d win 65 out of 100 elections. But we’re only having exactly one with current conditions.