United States Politics: ∆τ_i*ε*φ*m_i<0 where ε=4 and φ=0.25

[PhoenixDark]

I don't know why you would think he wouldn't win Iowa, he's at ~55% and Haley hasn't cracked 20% and there's barely any momentum compared to soaking up the people dropping out instead of DeSantis. NH is the closer one and less favorable Trump territory to begin with.

Polling caucuses is more difficult, sure, but not a 35 point margin difficult:
2020: 23/19/17/16 Sanders/Biden/Buttigieg/Warren on RCP vs. 25/21/19/15 Sanders/Buttigieg/Warren/Biden result.
2016: 26/24/18 Trump/Cruz/Rubio aggregate on 538 vs. 28/24/23 Cruz/Trump/Rubio result.
2012: 23/22/16/14 Romney/Paul/Santorum/Gingrich on RCP vs. 25/25/21/13 Santorum/Romney/Paul/Gingrich result.
2008: 31/29/26 Obama/Clinton/Edwards on RCP vs. 38/30/29 Obama/Edwards/Clinton result.
2008: 30/27/12/12 Huckabee/Romney/McCain/Thompson on RCP vs. 34/25/13/13 Huckabee/Romney/Thompson/McCain result.
The polls also showed the momentum for Buttigieg, Warren, Santorum, Rubio and Huckabee before away from the national frontrunners.

We probably have to go back to 1984 or earlier before finding a candidate polling at 50% or with a 20+ point lead in Iowa.

Looks like it would take a major polling science collapse to get this wrong. The weather just might be the thing to do that though, which is why I can't call it.