538 and Iowa: What a crock.

[Quixote1818]

8. That was correct based on polling on Jan. 10th when Biden was leading. It narrowed a lot

and and had Sanders ahead by the end of January. It's just like during a football game. If one team gets up by a couple of touchdowns then their chance of winning goes way up in the middle of the game based on past models of how many teams that were up by that amount at that time ended up winning but it's just a model and as the polling changes the model changes. It's why polling firms really try to get the last poll they take right because that is the one they will be judged by NOT the one 20 days in advance.

So yes, a lot can happen in the next few months but their models are based on past elections and what occurred after someone won or came close in Iowa and others under performed. Currently betting markets favor Sanders to win most of the states but they also give him a less than 50% chance of getting the nomination. So there are a lot of variables and they know a lot can happen as time passes. How many times has it rained when there was only a 20% chance of rain? Well, probably about 20% of the time it does even though there was an 80% chance it wasn't going to rain. They are only educated guesses based on modeling.