 |
 |
Some 40 new state head-to-head polls have been released since my previous analysis was posted last Friday. Many of the new polls have been taken partially or fully after the death of Supreme Court Justice Ruth Bader Ginsburg, an event that could possibly change the dynamics of the Presidential race between Democrat Joe Biden and Republican Donald Trump. If so, it has not affected the expected outcome of a (hypothetical) election held now.
After 100,000 simulated elections, Biden wins 100,000 times and Trump wins 0 times. Biden received (on average) 380 to Trump’s 158 electoral votes. In an election held now, Biden would have a greater than 99.9% probability of winning and Trump would have a less than 0.1% probability of winning.
Last week’s analysis gave Biden 389 and Trump 149 electoral votes (on average) so, for the moment, Biden has stopped gaining electors, although it is too early to say that Trump has the momentum going forward. With newly released polls, Trump made modest gains in Nebraska’s 2nd congressional district (13% to 39% probability of winning that district’s elector) and Texas (22% to 48% probability of winning the state). He made smaller gains in Georgia (12% to 15%), North Carolina (23% to 29%) and South Carolina (92% to 96%). At the same time, Trump lost modest ground in Alabama (100% to 77%) and Louisiana (100% to 87%) and lost a bit of ground in Iowa (66% to 63%).
Here is the distribution of electoral votes [FAQ] from the simulations:
Ten most probable electoral vote outcomes for Biden (full distribution here):
- 369 electoral votes with a 2.55% probability
- 360 electoral votes with a 2.11% probability
- 413 electoral votes with a 2.05% probability
- 389 electoral votes with a 2.00% probability
- 407 electoral votes with a 1.78% probability
- 375 electoral votes with a 1.78% probability
- 397 electoral votes with a 1.68% probability
- 385 electoral votes with a 1.66% probability
- 406 electoral votes with a 1.57% probability
- 395 electoral votes with a 1.54% probability
The long term trend in this race can be seen from a series of elections simulated every seven days using polls from 23-Sep-2019 to 23-Sep-2020, and including polls from the preceding month (when possible…see the FAQ).
Returning to the simulation for the current time period, after 100,000 simulations:
- Biden wins greater than 99.9%, Trump wins less than 0.1%.
- Average (SE) EC votes for Biden: 380.4 (24.0)
- Average (SE) EC votes for Trump: 157.6 (24.0)
- Median (95% CI) EC votes for Biden: 380 (335, 421)
- Median (95% CI) EC votes for Trump: 158 (117, 203)
Each column of this table shows the electoral vote total aggregated by different criteria for the probability of winning a state (Safe=100%, Strong=90%+, Leans=60%+, Weak=50%+):
| Threshold | Safe | + Strong | + Leans | + Weak |
|---|---|---|---|---|
| Safe Biden | 262 | |||
| Strong Biden | 46 | 308 | ||
| Leans Biden | 61 | 61 | 369 | |
| Weak Biden | 38 | 38 | 38 | 407 |
| Weak Trump | 0 | 0 | 0 | 131 |
| Leans Trump | 33 | 33 | 131 | |
| Strong Trump | 56 | 98 | ||
| Safe Trump | 42 |
This table summarizes results by state. Click on the poll count to see the individual polls included for the state.
| 1 | 1 | EC | # | Total | % | % | Biden | Trump | |
|---|---|---|---|---|---|---|---|---|---|
| 6 | 2 | Votes | polls | Votes | Biden | Trump | % wins | % wins | |
| AL | 9 | 1* | 552 | 47.8 | 52.2 | 22.7 | 77.3 | ||
| AK | 3 | 1* | 831 | 46.8 | 53.2 | 10.4 | 89.6 | ||
| AZ | 11 | 15 | 10997 | 52.5 | 47.5 | 100.0 | 0.0 | ||
| AR | 6 | 1* | 799 | 48.9 | 51.1 | 33.3 | 66.7 | ||
| CA | 55 | 3 | 2370 | 64.4 | 35.6 | 100.0 | 0.0 | ||
| CO | 9 | 4 | 2675 | 55.5 | 44.5 | 100.0 | 0.0 | ||
| CT | 7 | 2* | 1494 | 62.5 | 37.5 | 100.0 | 0.0 | ||
| DE | 3 | 1* | 675 | 61.0 | 39.0 | 100.0 | 0.0 | ||
| DC | 3 | 0* | (100) | (0) | |||||
| FL | 29 | 19 | 25599 | 52.1 | 47.9 | 100.0 | 0.0 | ||
| GA | 16 | 9 | 7863 | 50.9 | 49.1 | 85.0 | 15.0 | ||
| HI | 4 | 1* | 829 | 65.9 | 34.1 | 100.0 | 0.0 | ||
| ID | 4 | 1 | 565 | 36.6 | 63.4 | 0.0 | 100.0 | ||
| IL | 20 | 0 | (100) | (0) | |||||
| IN | 11 | 1 | 950 | 42.4 | 57.6 | 0.1 | 99.9 | ||
| IA | 6 | 2 | 1354 | 49.4 | 50.6 | 36.9 | 63.1 | ||
| KS | 6 | 2* | 1879 | 46.2 | 53.8 | 1.1 | 98.9 | ||
| KY | 8 | 1 | 1117 | 39.6 | 60.4 | 0.0 | 100.0 | ||
| LA | 8 | 1 | 540 | 46.7 | 53.3 | 13.2 | 86.8 | ||
| ME | 2 | 4 | 2964 | 59.0 | 41.0 | 100.0 | 0.0 | ||
| ME1 | 1 | 4 | 1648 | 64.8 | 35.2 | 100.0 | 0.0 | ||
| ME2 | 1 | 5 | 1691 | 51.8 | 48.2 | 86.4 | 13.6 | ||
| MD | 10 | 1* | 765 | 62.6 | 37.4 | 100.0 | 0.0 | ||
| MA | 11 | 1 | 745 | 69.5 | 30.5 | 100.0 | 0.0 | ||
| MI | 16 | 14 | 11682 | 54.0 | 46.0 | 100.0 | 0.0 | ||
| MN | 10 | 10 | 6551 | 54.6 | 45.4 | 100.0 | 0.0 | ||
| MS | 6 | 1 | 540 | 44.4 | 55.6 | 3.5 | 96.5 | ||
| MO | 10 | 3 | 2435 | 45.5 | 54.5 | 0.2 | 99.8 | ||
| MT | 3 | 3 | 1680 | 46.5 | 53.5 | 2.6 | 97.4 | ||
| NE | 2 | 0 | (0) | (100) | |||||
| NE1 | 1 | 1* | 376 | 48.9 | 51.1 | 39.7 | 60.3 | ||
| NE2 | 1 | 1 | 368 | 51.1 | 48.9 | 60.9 | 39.1 | ||
| NE3 | 1 | 0 | (0) | (100) | |||||
| NV | 6 | 1 | 407 | 52.3 | 47.7 | 74.8 | 25.2 | ||
| NH | 4 | 1 | 387 | 51.7 | 48.3 | 67.8 | 32.2 | ||
| NJ | 14 | 1 | 495 | 59.4 | 40.6 | 99.8 | 0.2 | ||
| NM | 5 | 1 | 1044 | 58.0 | 42.0 | 100.0 | 0.0 | ||
| NY | 29 | 1* | 977 | 66.3 | 33.7 | 100.0 | 0.0 | ||
| NC | 15 | 19 | 16099 | 50.3 | 49.7 | 71.3 | 28.7 | ||
| ND | 3 | 2* | 942 | 40.3 | 59.7 | 0.0 | 100.0 | ||
| OH | 18 | 4 | 5083 | 50.3 | 49.7 | 62.1 | 37.9 | ||
| OK | 7 | 1 | 459 | 37.0 | 63.0 | 0.0 | 100.0 | ||
| OR | 7 | 1 | 452 | 56.6 | 43.4 | 98.2 | 1.8 | ||
| PA | 20 | 17 | 15649 | 52.4 | 47.6 | 100.0 | 0.0 | ||
| RI | 4 | 0 | (100) | (0) | |||||
| SC | 9 | 2 | 1648 | 46.8 | 53.2 | 3.6 | 96.4 | ||
| SD | 3 | 0 | (0) | (100) | |||||
| TN | 11 | 2* | 1407 | 43.6 | 56.4 | 0.1 | 99.9 | ||
| TX | 38 | 5 | 7830 | 50.1 | 49.9 | 51.8 | 48.2 | ||
| UT | 6 | 1 | 880 | 39.8 | 60.2 | 0.0 | 100.0 | ||
| VT | 3 | 1 | 516 | 63.6 | 36.4 | 100.0 | 0.0 | ||
| VA | 13 | 1 | 637 | 58.1 | 41.9 | 99.8 | 0.2 | ||
| WA | 12 | 1 | 471 | 61.8 | 38.2 | 100.0 | 0.0 | ||
| WV | 5 | 1* | 495 | 33.3 | 66.7 | 0.0 | 100.0 | ||
| WI | 10 | 19 | 13799 | 54.0 | 46.0 | 100.0 | 0.0 | ||
| WY | 3 | 0 | (0) | (100) |
* An older poll was used (i.e. no recent polls exist).
Details of the methods are given in the FAQ.
The most recent analysis in this match-up can be found from this page.
It’s getting a little late to get your gear and equipment together, so remember the first thing about surviving what happens on November 4 is to stay away from crowds.
How is Wisconsin dark blue but Oregon and Washington are not?
2,
# of polls.
Being only one each in WA and OR.
@3 I thought maybe the crayon broke.
This “simulation” is not looking good Perfessa. Florida and Arizona are waaaaaaaaaaaaaay off!
WA Post/ABC News has Florida much different. https://www.westernjournal.com/trump-gains-4-point-lead-biden-fl-likely-voters-latest-abc-wapo-poll/?utm_source=site&utm_medium=protrumpnews&utm_campaign=can
What else is slanted here? Huh?
@ 5
No slants. Everything out in the open, including Darryl’s willingness to answer specific questions and argue specific points of contention.
In this analysis, Texas is in Biden’s column with a 50.1-49.9 advantage, although only one of five polls – the largest, by far – favored Biden over Trump. The others favored Trump.
NYT/Siena, reported today but completed on the 21st, has Trump up 3 in TX. Interestingly, NYT reports it as Trump and Biden “effectively locked in a statistical tie“.
Flip 38 EVs from Biden’s column to Trump’s and things change, a bit. Maybe Trump’s chances grow to 0.5%.
Hey Darryl, if you correct the FL GQR poll result currently showing Biden with a 47 point lead, does it flip the state to red? /sarc
FL St Pete poll ending 9/20:
Difference between a registered voter and an “active” one (voted in the last 4 years), by party:
Something to keep in mind when looking at poll weighting.
Good.
[Deleted — off topic]
What else is slanted here? Huh?…
No slants. Everything out in the open, including Darryl’s willingness to answer specific questions and argue specific points of contention.
i.e. SCIENCE.. Teh babblin’ butthole HATES SCIENCE. Science concludes the earth is BILLIONS of years old..
Teh babblin butthole believes what 5, 6 thousand years old? What a credulous kult-addled dummy!
What an imbecile!
@6,
Yeah…I saw that typo this morning. It doesn’t make a huge difference because the poll was only 800 folks out of a total of 25,599 polled individuals. Here is the result with that poll fixed:
The results aren’t totally compatible with yesterday’s analysis because one poll (that happened to have Biden up by 7% or so) “aged out” of the “current poll” set. Together, dropping the old poll and fixing the typo, drops Biden’s probability of winning Florida to only 99.8%.
Scott @2,
How is Wisconsin dark blue but Oregon and Washington are not?
As @3 pointed out, it is because there is only one poll (total) in Oregon, and only one “current poll” in Washington. Both polls are really small (452 and 471, respectively). This means that there is a lot of uncertainty in the outcomes.
In the Washington race, for example, there were 100,000 simulated elections. Biden won 99,964 of them and Trump won 36 times.
One of the arbitrary things about these simulations is that I must pick a “cutoff” for when polls are too old to include. I am using a one month cutoff right now. If I used a larger window that included the SurveyUSA poll from July as well, the sample of polled individuals would be much larger and Biden would likely win all of the simulated elections.
For Oregon, however, there are no earlier polls.
Godwinha @6,
Flip 38 EVs from Biden’s column to Trump’s and things change, a bit. Maybe Trump’s chances grow to 0.5%.
Given that Trump is winning Texas about 1/2 the time now, it will only change the results by about half of Texas’s 38 EVs.
So…lets try. I created a fake poll of 10,000 people with 99% going for Trump. Note that there are 23 new polls that were not included in yesterday’s analysis, including one new Texas poll (N = 653, Trump 46%, Biden 43%), so the results aren’t strictly comparable, but you’ll get the idea.
What happens:
Yep…about 1/2 of 38 plus a couple extra for the new poll.
Biden still won all 100,000 simulated elections. Here are the top 10 most probable electoral college outcomes:
375 electoral votes with a 3.92% probability
369 electoral votes with a 3.82% probability
359 electoral votes with a 3.05% probability
360 electoral votes with a 2.81% probability
365 electoral votes with a 2.70% probability
357 electoral votes with a 2.64% probability
350 electoral votes with a 2.53% probability
374 electoral votes with a 2.52% probability
368 electoral votes with a 2.51% probability
364 electoral votes with a 2.48% probability
I wouldn’t have considered most of these outcomes a “landslide”, per se, but in 2016, I learned that 306 electoral votes constitutes a “massive landslide victory”.
@ 13
Figured you would re-run simulations. Thanks for doing it, and particularly for including this sentence:
Given that Trump is winning Texas about 1/2 the time now, it will only change the results by about half of Texas’s 38 EVs.
I was thinking all-or-none, which I now realize isn’t correct.