Two 53% chances that are probably very correlated is probably a bit lower odds than a coin flip, so it's more like "republicans could sweep". Still, Hillary had much better chances on all the models the night before the election in 2016, and well, I hate to tell you she lost.