AtlasIntel- was the most accurate presidential poll for the 2020 presidential election. They just dropped their 2024 presidential poll. 
This is very bad news if you're a Hyena-nette.
 Trump: 50.7% (+3.6)
 Harris: 47.3%
 West: 0.2%
 Stein: 0.4%
 Oliver: 0.1%
 Undecided: 1.1%
 1,775 Likely Voter | ±2 p.p | 09/11/2024 - 09/12/2024 |
 Dem 32.4, Rep 33.5, Ind 34.1%
From On Point Politics:
I decided to create an "A-PLUS" Version of my National Average, which only includes my A Graded Pollsters! That includes these pollsters with their 2020 Margins:
@Rasmussen_Poll
 - Rasmussen Reports (Biden+5.25)
 @Peoples_Pundit
 - Big Data Poll (Biden+4.1)
 @EmersonCollege
 - Emerson (Biden+5)
 @atlas_intel
 - AtlasIntel (Biden+4.7)
When we include these four polls into an A Graded version of my average, weighing to an R+1, this is the result:
 Trump: 49.38% (+2.86)
 Harris: 46.52%
 Kennedy: 0.11%
 West: 0.67%
 Stein: 0.61%
 Undecideds
When I use these pollsters and include them in average from October 2nd to November of 2020, as well as weighing them to the 2016 Electorate, we get:
 Biden: 51.0% (+5.90)
 Trump: 45.1%
 Undecided: 4.0%
This is more accurate than both the RCP and 538 Average in 2020, which had Biden winning by +7.2 and +8.4 respectively:
So, if we do the math:
 2020 Average - Biden+5.9
 2024 Average - Trump+2.9
 Shift = R+9
 Biden+4.5 -----> Trump+4.5 (Landslide In Electoral College)
 Electoral College Based On Trump+4.5 Popular Vote
Trump- 319 EV
Harris- 219 EV 
My 296 EV prediction for Trump is looking rather pedestrian.
Mean tweets 2024!