2020 Election forecast (subject to change on Nov.2)

Richard Charnin: Oct.27, 2020

2020 Election forecast (subject to change on Nov.2)
Assume 150 million votes and a fair count

Party ID (Gallup voter affiliation): Rep 28%, Dem 27%, Ind 45%

Electoral Vote: Trump 356, Biden 182

Trump 51.7%, Biden 46.4%, Other 1.9%

Votes (millions) Trump 77.5, Biden 69.6, Other 2.8

Note: I forecast the electoral vote exactly in each of the last three elections. Will I do it again in 2020? I doubt it. It takes skill and luck.

Assumptions: Trump wins 10% of Dems, 93% of Repubs and 51% of Independents. In 2016 Trump had 8%, 89%, 46%.

Trump increases his a) Dem% due to gains among Blacks, Hispanics and Dem Walkaways, b) Rep% due to an energized base, c) Ind% due to fear of socialism.

Assume equal voter % turnout. This is conservative. Reps always turn out at a higher rate than Dems.

	Votes	Party-ID	Biden	Trump	Other
Dem	40,500	27.0%	89.0%	10.0%	1.0%
Rep	42,000	28.0%	6.0%	93.0%	1.0%
Ind	67,500	45.0%	46.0%	51.0%	3.0%
	Total	100.0%	46.4%	51.7%	1.9%
	Votes	150,000	69,615	77,535	2,850
	Margin	7,920

Sensitivity Analysis

		Trump	%Rep
Trump	91.0%	92.0%	93.0%	94.0%	95.0%
% Ind			Trump%
55%	52.9%	53.2%	53.5%	53.8%	54.1%
53%	52.0%	52.3%	52.6%	52.9%	53.2%
51%	51.1%	51.4%	51.7%	52.0%	52.3%
49%	50.2%	50.5%	50.8%	51.1%	51.4%
47%	49.3%	49.6%	49.9%	50.2%	50.5%
			Biden%
55%	45.2%	44.9%	44.6%	44.3%	44.1%
53%	46.1%	45.8%	45.5%	45.2%	45.0%
51%	47.0%	46.7%	46.4%	46.1%	45.9%
49%	47.9%	47.6%	47.3%	47.0%	46.8%
47%	48.8%	48.5%	48.2%	47.9%	47.7%
		Pct	Margin
55%	7.8%	8.3%	8.9%	9.4%	10.0%
53%	6.0%	6.5%	7.1%	7.6%	8.2%
51%	4.2%	4.7%	5.3%	5.8%	6.4%
49%	2.4%	2.9%	3.5%	4.0%	4.6%
47%	0.6%	1.1%	1.7%	2.2%	2.8%
		Vote	Margin
55%	11,640	12,480	13,320	14,160	15,000
53%	8,940	9,780	10,620	11,460	12,300
51%	6,240	7,080	7,920	8,760	9,600
49%	3,540	4,380	5,220	6,060	6,900
47%	840	1,680	2,520	3,360	4,200

Model calculations https://docs.google.com/spreadsheets/d/1yh0YkXBVctdJwt0_XHrpCC6f7K30oEz0Vr13MiU5qT4/edit#gid=297739797

 

Trump projected Vote share range and corresponding Electoral Votes

Richard Charnin

Oct.26, 2020

The following 2020 Election Model tables show six forecast scenarios of Trump vote shares and corresponding electoral votes. Trump’s margins range from a 5.6 million loss with 47.4%, a 3.4 million win with 50.4% and a 9.4 million win with 52.4%.

	Turnout	Party-ID	Biden	Trump	Other
Dem	41,753	27.8%	90%	9%	1.0%
Rep	43,299	28.9%	7%	92%	1.0%
Ind	64,948	43.3%	47%	51%	2.0%
	Share	100%	48.2%	50.4%	1.4%
	Votes	150,000	72,227	75,635	2,138
	Elect Vote	538	182	356

Pop Vote	52.4%	51.4%	50.4%	49.4%	48.4%	47.4%
EVote	382	356	356	325	291	214
AK	3	3	3	3	3	3
AL	9	9	9	9	9	9
AR	6	6	6	6	6	6
AZ	11	11	11	11	11
CA
CO	9	9	9	9	9
CT
DC
DE
FL	29	29	29	29	29
GA	16	16	16	16	16	16
HI
IA	6	6	6	6	6	6
ID	4	4	4	4	4	4
IL
IN	11	11	11	11	11	11
KS	6	6	6	6	6	6
KY	8	8	8	8	8	8
LA	8	8	8	8	8	8
MA
MD
ME	4	4	4	4
MI	16	16	16
MN	10	10	10
MO	10	10	10	10	10	10
MS	6	6	6	6	6	6
MT	3	3	3	3	3	3
NC	15	15	15	15	15
ND	3	3	3	3	3	3
NE	5	5	5	5	5	5
NH	4	4	4	4	4	4
NJ	14
NM	5	5	5
NV	6	6	6	6	6	6
NY
OH	18	18	18	18	18	18
OK	7	7	7	7	7	7
OR
PA	20	20	20	20
RI
SC	9	9	9	9	9	9
SD	3	3	3	3	3	3
TN	11	11	11	11	11	11
TX	38	38	38	38	38	38
UT	6	6	6	6	6	6
VA	13	13	13	13	13
VT
WA	12
WI	10	10	10	10
WV	5	5	5	5	5	5
WY	3	3	3	3	3	3

https://docs.google.com/spreadsheets/d/1yh0YkXBVctdJwt0_XHrpCC6f7K30oEz0Vr13MiU5qT4/edit#gid=737914131

 

Probability of 1,2 or 3 Covid deaths in a group of 200

Richard Charnin Oct. 25

What if you know 200 people in all age groups. What is the approximate probability P that 1, 2 or 3 died from Covid? Assume CDC reported deaths.

The probabilities are: P(1)=1/100; P(2)= 1/18,000; P(3)=1/5 million

The calculation function: P(n) = poisson (n, N*R*T, false) where n=number of deaths among N= 200 people, R=.00007= mortality rate based on the total U.S. population, T=.75 year (9 months).

Since 6% of Covid deaths had ZERO comorbidities, the approximate probabilities of pure Covid deaths would be much lower than above.

 

3rd parties in the 2016 election cost Trump the recorded popular vote.

Richard Charnin

Oct. 21, 2020

Consider 3rd party adjustments to the 2016 vote. Clinton won the recorded popular vote by 2.8 million 65.7-62.9 (48.2-46.1%). Trump won by 304-232 electoral votes.

Other candidates votes (millions): Johnson (Libertarian- 4.48); Stein (Green- 1.45); McMullin (Independent- 0.72). In 2015-16, McMullin was Chief policy director for the House Republican Conference.

Assume Trump would have won 90% of the Johnson and McMullin votes and Clinton won 90% of the Stein votes. Then Trump wins the popular vote by 68.6 million (50.34%) to Clinton’s 67.6 million (49.66%). Trump wins the electoral vote by 325-213 and these states he lost in 2016: MN(10), NH (4), NV (6)

https://docs.google.com/spreadsheets/d/1R9Y3ae2uyW8SUxVUnnOt9ZyvheAxa0fAhesAw_nhciM/edit#gid=1200059719

 

Exposing the MSM myth that Biden leads Trump by 10% nationally

Richard Charnin

Oct. 18, 2020

The Lamestream Media keeps repeating the myth that Biden is leading Trump nationally by 10%. Let’s take a closer look at 10 polls. Biden leads the average by 50.7-42.1%. Average Party-ID: Dem 38.1%, Rep 33.4%, Ind 28.4%.

If we apply Gallup national voter affiliation weights (28R-27D-45I) in each of the polls, Biden’s 8.6% margin is cut to 3.2% (46.5-43.3%).

Republicans always turnout at a higher rate than the Democrats which would increase Trump’s total vote share. But if we cannot trust the MSM Party-ID weights, why should we trust the vote shares? Could they have also been inflated for Biden?

View row 20 in the 2020 Election Model: https://docs.google.com/spreadsheets/d/1yh0YkXBVctdJwt0_XHrpCC6f7K30oEz0Vr13MiU5qT4/

 

YouGov poll: Biden leads 46-38%. After allocating undecided voters: Trump leads 49.6-46.8%

Richard Charnin

Oct. 15, 2020

YouGov poll: Biden leads 46-38%. After allocating undecided voters: Trump leads 49.6-46.8%

In 2016 Trump had 57% of whites but only 45% in the poll. Trump is doing better among Blacks and Hispanics than he did in 2016

YouGov	2016 Census	Biden	Trump	Other	Undec
White	73.3%	40%	45%	2%	5%
Black	12.4%	75%	8%	0%	5%
Latino	9.2%	52%	24%	0%	18%
Other	5.0%	45%	42%	1%	5%
Calc		45.7%	38.3%	1.5%
Poll		46%	38%	2%
Adjusted	2016 Census	Biden	Trump	Other
White	73.3%	40%	57%	3%
Black	12.4%	82%	13%	5%
Latino	9.2%	54%	42%	4%
Other	5.0%	45%	47%	8%
Adj Calc		46.8%	49.6%	3.6%
Poll		46%	38%	2%

RV	Party-ID	Biden	Trump	Other
Dem	36.0%	93%	4%	1%
Rep	32.0%	7%	91%	0%
Ind	32.0%	46%	37%	5%
Calc	100.0%	50.4%	42.4%	2.0%
Poll	93.0%	50.0%	41.0%	2.0%

LV	Gallup	Biden	Trump	Other
Dem	27.1%	96%	1%	0%
Rep	28.7%	2%	95%	1%
Ind	44.2%	32%	42%	6%
Calc	100.0%	40.7%	46.1%	2.9%
Poll	96.0%	51.0%	43.0%	2.0%

RV	Gallup	Biden	Trump	Other
Dem	27.1%	93%	4%	1%
Rep	28.7%	7%	91%	0%
Ind	44.2%	46%	37%	5%
Calc	100.0%	47.5%	43.5%	2.5%
Poll	93.0%	50.0%	41.0%	2.0%

 

Effect of Party ID change in YouGov Poll to Gallup on Election share

YouGov vs. Gallup party-ID: Effect on YouGov Election forecast. Another example of how over-weighted Democrat samples skew the polls.

Party-ID: YouGov: Biden 45.2-38.5%; Gallup: Trump 42.7-39.0%. Approximately 6% undecided.

YouGov	Party-ID	Biden	Trump	Other
Dem	37.6%	89%	1%	0%
Rep	27.4%	2%	92%	1%
Ind	35.0%	32%	37%	4%
RV Calc	100%	45.2%	38.5%	1.7%
RV Poll	100%	46%	38%	2%

YouGov	Gallup	Biden	Trump	Other
Dem	27%	89%	1%	0%
Rep	28%	2%	92%	1%
Ind	45%	32%	37%	4%
Adj. Calc	100%	39.0%	42.7%	2.1%

pg21 https://docs.cdn.yougov.com/1qb5j6pdqd/econTabReport.pdf

 

Another rigged poll from NBC: Biden leading 53-39%.

Richard Charnin

Oct. 5, 2020

LINKS TO  BLOG POSTS

Another rigged poll from NBC: Biden leading 53-39%. As usual the poll overweights Dems and understates Trump’s shares.

Let’s compare the poll to my calculation which shows Trump leading using the latest Gallup voter affiliation survey. Trump had a recorded 8% of Dems in 2016. He will do better in 2020 since he is gaining black and Hispanic voters. He will also do better among Republican voters (he had 89% in 2016) and Independents (he won by 46-42%).

NBC	Party-ID	Biden	Trump	Other
Rep	27%	9%	84%	7%
Dem	35%	93%	5%	2%
Ind	38%	47%	38%	15%
	Poll	53%	39%	8%
RC	Gallup	Biden	Trump	Other
Rep	29%	7%	91%	2%
Dem	30%	87%	10%	3%
Ind	41%	42%	46%	12%
	Poll	53%	39%	8%
	Calc	45.3%	48.3%	6.4%

 

Debate infection Probabilities for Democrats and Republicans

Richard Charnin

Oct. 5, 2020

LINKS TO  BLOG POSTS

Assume 100 people attended the Barrett nomination and/or the debate, 50 Dems and 50 Repubs. The probability of infection is 2%.

The probability that a) ZERO Dems would be infected is 0.364 and b) TEN Repubs would be infected is 1 in 21 million.

	Binomial distribution
n	1 in	Prob
0	2.75	0.364
1	2.69	0.372
2	5	0.186
3	16	0.061
4	69	0.015
5	366	0.003
6	2,392	0.000
7	18,645	0.000
8	169,976	0.000
9	1,784,753	0.000
10	21,329,970	0.000

n	Poisson Prob (n of 50)	prob
0	89.39%	8.94E-01
1	10.03%	1.00E-01
2	0.56%	5.62E-03
3	0.02%	2.10E-04
4	1 in 1 million	5.90E-06
5	1 in 10 million	1.32E-07
6	1 in 1 billion	2.47E-09
7	1 in 100 billion	3.97E-11
8	1 in 10 trillion	5.56E-13
9	1 in 1000 trillion	6.93E-15
10	1 in 100,000 trillion	7.77E-17

 

What are the odds that FOUR Republicans, SIX Trump/Amy Barrett-related people and ZERO Democrats would be infected within a few days of the debate and the upcoming SC vote?

Richard Charnin

Oct.3, 2020

LINKS TO  BLOG POSTS

What are the odds that FOUR Republicans, SIX Trump/Amy Barrett-related people and ZERO Democrats would be infected within a few days of the debate and the upcoming SC vote?…. President Trump, Senators Ron Johnson, Mike Lee and Tim Tillis.

The following 6 were also infected:Rep. Louis Gohmert got the virus last month (he recovered with HCQ). Others include Bill Stepien, President Trump’s campaign manager; Melania Trump, first lady; Hope Hicks, Trump advisor; John I. Jenkins, president of the University of Notre Dame; and Kellyanne Conway, former top White House adviser.

Given: 7 million out of 300 million infected. The probability that FIVE of 250 Republicans would be infected in 1 week: P= 1 in 7.5 million where P= poisson (5, 250* 7/300* 1/52, false)

https://en.wikipedia.org/wiki/White_House_COVID-19_outbreak#Amy_Coney_Barrett_nomination_event,_September_26

https://www.nytimes.com/live/2020/10/02/us/trump-vs-biden?fbclid=IwAR2qNJTfN2YdTJhQ0hTH3tprmeJteeT5kR4WE8pVZu57eoSTTfJLB7tOp_k