Read Ralston's prediction column a bit more. He's pretty much banking on the Reid Machine carrying Harris across the finish line in NV, along with some commonly held Democratic beliefs that are really just assumptions. A couple excerpts from his post...
I can see from top to bottom that races could go either way, but I have decided to trust the Reid Machine that has not lost for four consecutive presidential cycles and will somehow get enough ballots turned in during the next few days to do what it always does. All of this falls apart if indies don’t go for Vice President Kamala Harris and if the machine can’t get enough ballots returned – not only would Trump win but there will be upsets down-ballot.
The key to this election has always been which way the non-major-party voters break because they have become the plurality in the state. They are going to make up 30 percent or so of the electorate and if they swing enough towards Harris, she will win Nevada. I think they will, and I’ll tell you why: Many people assume that with the GOP catching up to the Democrats in voter registration that the automatic voter registration plan pushed by Democrats that auto-registers people as nonpartisans (unless they choose a party) at the DMV had been a failure for the party. But I don’t think so. There are a lot of nonpartisans who are closet Democrats who were purposely registered by Democrat-aligned groups as nonpartisans. The machine knows who they are and will get them to vote. It will be just enough to overcome the Republican lead – along with women motivated by abortion and crossover votes that issue also will cause. I know some may think this reflects my well-known disdain for Trump, heart over data. But that is not so.
Models keep trending toward a -PNA pattern toward the end of the run, maybe the last 10 days of November. This is a pretty strong signal developing. 12z GEFS hr384. I kind of agree that you want -NAO help this year, at least early on.