NEW: 2020 map
Biden 350, Trump 188

OH-Outlier D+5. Only poll since June, need confirmation. Hard to believe
Finally polls in GA, TX, IA. All close
FL ticks under 2%. Tilt D
WI stays strong. Likely D
AZ ?s on SCOTUS. Joe needs to keep right leaning voters. Lean D

Quinn banned
1/
NC remains close, but clear Biden lead. Tilt D

SC now safe given Biden's weakness in the South. GA, FL, NC have swung 6% or less from 2016. SC swinging 15% isn't feasible

PA/MI no chg

NV finally a poll, strong Biden. Stays likely D

MT/AK/MO/KS possible reaches

Explained👇
2/
Ohio showing Biden+5 in Fox is a clear outlier. OH is fundamentally hard

Needs 8.3 swing from 2016 (higher than any state now)

Small swing in 2018
R+4.8 (House) and R+3.7 (gov) (3.5/4.6 swing)
PA D+8.2, MI D+9.1, MN D+10.1 swing

Rural OH largely Appalachia. Cities are small
3/
Also Rs DRASTICALLY beat polls in 2016 and 2018
2016 polled R+1.9, voted R+8.3 (6.4 off)
2018 gov D+2.4->R+3.7 (6.1 off)
2018 sen D+11.9->D+6.8 (5.1 off)

Blue OH possible, but given the fundamental difficulty, outlier nature of FOX and lack of other polls, I'm leaving tilt R
4/
IA had D+3, tie, R+3

Needs even bigger swing from Trump+9.4. It's a very swingy state!

Bush+0->Obama+9->Obama+5->Trump+9

Fundamentals are better. Biden making big gains with WI/MI/MN rural whites

Trump approval low for how R it is

Could overcome 2018 R+3, but hard. Tilt R
5/
GA had tie/tie/R+2. However far more D undecideds (black, young, indy). That pushes to Biden <1

Early VBM is amazing for Dems. 25% of requests from 2016 nonvoters, a heavy Biden group usually not included in polls. High in metro ATL and for nonwhites. Return data soon

Tilt D
6/
TX saw R+3. Similar to GA with heavy undecideds among D voters (20% among AAs?!)

But, other polls this month R+2, R+2, R+1, D+3, so narrow but real R lead

VBM limited to 65+, taking away low propensity D voters who use VBM in other states. Although half of VBM is 65+ anyway

7/
Also, SCOTUS could eat at Bidens R and R leaning indy support. That is why I moved to tilt R right when RBG died. TBD if it happens

Fundamentally TX is great for Biden. HEAVILY urban/suburban, and (historically) DEEP R suburbs are swinging fast to Dems. And exploding in pop

8/
From 2012->2016->2018, the burbs sprinted left. Insane. Will continue. Trump is so toxic

Dallas: R+41->R+33->R+21
Houston: R+33->R+25->R+18
Austin: R+19->R+10->DEM+5

Core urban ones too

But heavy nonwhite vote

So no guarantee the swing is enough. For now Trump +1-2. Tilt R
9/
FL: outlier R+4 poll. Still, clear Biden lead on avg, albeit smaller than last week (1-2%). VBM is extensive and starts soon. Tilt D

NC: Still close, D+1-2%. VBM being returned and very strong for Dems. Tilt D

WI remains very strong. With MI/IA/MN too, not a fluke. Likely D
10/
AZ: Lead of 3-5% not as large as some others, but very consistent. Even before summer, 2.5-3%. No tightening like other states. Rock solid

However, Biden relies on some right leaning indies and Rs. With SCOTUS, TBD if he retains them. Probably, but can't be certain. Lean D

11/
NV polling is unreliable and sparse, but today's Biden+11 from FOX is a good sign that NV isn't seriously competitive

Was D+2 in 2016 and hard to imagine the national vote being 5% more D and NV swinging right 7%

Universal VBM should help young/ minority turnout. Likely D

12/
For likely R, take out SC. It was R+15 in 2016, but the South is where we're seeing the smallest shifts. NC, GA just 5-6% and FL under 3%. A 15% move is out of the question

AK/MT/KS/MO remain plausible. AK is a weird state, KS is trending left. MO, MT *nearly* went Obama

13/
NH polling also sparse but we have D+8 avg, and given ME seems unlikely to be closer. Likely D

MI/PA, no chg. PA still worst D state in rust belt. Likely D/lean D

Finally, there is no sign that MN, VA, CO, NM or ME are at all close. If that changes I'll update. All safe D
14/
PS I'm banning @QuinnipiacPoll polls. Long time coming. They routinely show ridiculous numbers like

Aug - KY 9% and McGrath down 4%, SC just 5%, senate tied
Sep - SC just 6%, senate tied. Maine+21%, Gideon+12%
July - FL Biden+13%
June - Nat'l Biden+15%

But it gets worse
15/
Outliers are okay, a pollster that leans D usually is okay

BUT they frequently have a HUGE issue with minorities, esp Latinos. Frequent, egregious, irreconcilable

Today TX: Biden+8
Sep Nat'l: Biden+20
Aug FL: Trump+2
July Nat'l: Biden+10
May Nat'l: Biden+7
Apr FL: Biden+8
16/
Also today's Texas poll Trump at 19% with blacks

Fix? In the TX poll if Biden wins Latinos by 32% and blacks by 80%, then overall result is Biden+3 (because Trump lead among whites so low)

But you can't do that. Maybe the Rs they poll are Latino/black, so the Ds are white?
17/
Thus Trump way too high with Latinos/blacks but too low with whites. So can't just add Dem margin to the Latino/black #s . They're intertwined

But if 50% of the Latinos they poll are GOP/indy, that's a trash sample that doesn't represent TX. Similar if the whites are too Dem
18/
Could explain some of the ridiculous results in white states like Trump+9 KY or Biden+21 ME

Anyway, there's no solution in my view other than to throw them out entirely. I'd love to say "if Latino # is decent, use," but that's cherrypicking

SO, henceforth, they're banned

/fin
You can follow @rainbow_jeremy_.
Tip: mention @twtextapp on a Twitter thread with the keyword “unroll” to get a link to it.

Latest Threads Unrolled: