One Reason I'm Not That Good at Predicting Presidential Nomination Winners ...

... as opposed to general election winners, is that there doesn't seem to be as much state-level polling.

Keep in mind that, like the general election, the primaries are not a "national election." They're 50 separate elections, with delegate rules similar to, but not exactly like, the Electoral College system. That is, not every primary or caucus is "winner take all" for delegates, but it's still possible for a nomination candidate to get the most votes nationally yet lose the nomination.

"Nationally," Joe Biden continues to lead the Democratic nomination contest, and to lead in various sub-demographics (black Democrats, female Democrats, "somewhat liberal" Democrats, and "moderate/conservative" Democrats), the only exception being "very liberal" Democrats according to the latest Quinnipiac poll.

Biden dropped from 34% to 32%. Elizabeth Warren is gaining on him versus a week ago (up from 15% to 21%), Kamala Harris is fading (down from 12%  to 7% -- THANK YOU, Tulsi!), Sanders moved from 11% to 14%. Fourteen of the also-rans are at less than 1%.

The reasonably current (last two weeks) state polls at RealClearPolitics also have Biden leading in every state polled -- New Hampshire, Texas, Nevada, South Carolina, and Ohio.

1) It's early days and 2) I'd be more confident if there were 50 state polls every week. Among other things, that much information would make it worth digging into how delegates are distributed, who's doing well with the "superdelegates" (I hear Buttigieg is pushing hard on that front), etc. The state-level snapshots are too few and far between. The last California poll I see has Biden only edging out Harris by 1% there, and it's from nearly three weeks ago.

Based on the available current information, it still looks like a Biden/Warren race for the nomination and my best guess is a Biden/Warren ticket. But there's no way to be really confident about that yet.