1920 election with semi-proportional EV

[RBH]

My system for splitting EVs goes as follows:

Take the EV representing Senate seats, and give them to the winner of the state.

Take all the EV representing congressional districts, and split them based on popular vote.

But, with a threshold of "Halfway to the first seat"

For example.

New York, 1920

The threshold to get one exact vote is 2.3% (or 2.32558139534884%), and to get considered, you have to have one-half of that threshold or more

The results:
Warren Harding - 1871167 votes, 30 EV (28 + 2 at-large)
James Cox - 780037 votes, 12 EV
Eugene Debs - 203201 votes, 3 EV

And the total national results are.

Harding - 316 EV
Cox - 199 EV
Debs - 10 EV
Parley Christensen (Farmer-Labor) - 2 EV
James Ferguson (American) - 2 EV
Aaron Watkins (Prohibition) - 1 EV
Black and Tan Republican Electors - 1 EV

A split of the votes between the main candidates (Harding's EVs on the left, Cox's EVs on the right)

Alabama - 3/9
Arizona - 3/0
Arkansas - 3/6
California - 9/3
Colorado - 4/2
Connecticut - 5/2
Delaware - 3/0
Florida - 1/5
Georgia - 3/11
Idaho - 3/1
Illinois - 20/7
Indiana - 9/6
Iowa - 10/3
Kansas - 7/3
Kentucky - 5/8
Louisiana - 2/8
Maine - 5/1
Maryland - 5/3
Massachusetts - 13/4
Michigan - 12/3
Minnesota - 9/2
Mississippi - 1/9
Missouri - 11/7
Montana - 3/1
Nebraska - 6/2
Nevada - 3/0
New Hampshire - 3/1
New Jersey - 10/4
New Mexico - 3/0
New York - 30/12
North Carolina - 4/8
North Dakota - 4/1
Ohio - 15/8
Oklahoma - 6/4
Oregon - 4/1
Pennsylvania - 26/10
Rhode Island - 4/1
South Carolina - 0/9
South Dakota - 4/1
Tennessee - 7/5
Texas - 4/13
Utah - 3/1
Vermont - 4/0
Virginia - 4/8
Washington - 5/1
West Virginia - 5/3
Wisconsin    - 10/2
Wyoming - 3/0

Third party EVs

Debs: California (1), Illinois (1), Massachusetts (1), Minnesota (1), New York (3), Ohio (1), Wisconsin (1)

Christiansen: Illinois (1), Washington (1)

Ferguson: Texas (2)

Watkins: Pennsylvania (1)

Black & Tan: Texas (1)

I haven't done this for all of the elections, but Debs did win 17 EV in 1912.

The formula for figuring out the threshold

First: EVs - 2

Second: 1/(EVs - 2)

That should give you a percentage (refered to as "The EV Percentage" later). Then divide the percentage in half for the threshold.

And for 3 EV states, give all three EVs to the popular vote winner.

For situations where rounding up gives you too many EV, just don't round up the guy's EVs who is farthest from the next number.

For situations where you don't have enough EV, round up the person's EV who is closest to the next number.

Figuring out the number of EV goes as follows.

Divide the votes of the candidate involved with the total of the candidates who met the threshold.

Then divide the Candidate's Percentage by the EV Percentage.

Seven candidates getting an EV could be the record if I actually compute all the elections after 1828.

[Colin]

You should do this with other elections. 1992, 1968, 1948, 1912, 1860 just to name a few.

[○∙◄☻¥tπ[╪AV┼cVê└]

I think a better way to allocate would to round up. Otherwise, if there was a large state, you could get twice as many electors for a given number of candidates if it was spread out over several candidates, then all for one candidate.

[RBH]

You should do this with other elections. 1992, 1968, 1948, 1912, 1860 just to name a few.

Actually, I have done those elections. Maybe I should post them.

[Colin]

You should do this with other elections. 1992, 1968, 1948, 1912, 1860 just to name a few.

Actually, I have done those elections. Maybe I should post them.

Definitely.

[RBH]

I posted them in a new thread.

And I use semi-proportional for two reasons

#1 - The At-large votes will distort the vote slightly.

Example:

1912, the 435 proportional votes
Wilson - 210
Roosevelt - 111
Taft - 96
Debs - 17
Chafin - 1

1912, the 531 total votes
Wilson - 290
Roosevelt - 123
Taft - 100
Debs - 17
Chafin - 1

#2 - A lot of the southern states had lower turnout then the 3 EV states until the Voting Right's Act. This distorts the proportional vote in the favor of the Democrats until 1944 or so.

Example:

South Carolina, 1936

Roosevelt - 113791 votes, 8 EV
Landon - 1646 votes
(115437 votes total, 8 EV)

Vermont, 1936

Landon - 81023 votes, 3 EV
Roosevelt - 62134 votes
(143689 votes total, 3 EV)

[Dr. Cynic]

The results:
Warren Harding - 1871167 votes, 30 EV (28 + 2 at-large)
James Cox - 780037 votes, 12 EV
Eugene Debs - 203201 votes, 3 EV

And the total national results are.

Harding - 316 EV
Cox - 199 EV
Debs - 10 EV
Parley Christensen (Farmer-Labor) - 2 EV
James Ferguson (American) - 2 EV
Aaron Watkins (Prohibition) - 1 EV
Black and Tan Republican Electors - 1 EV

A split of the votes between the main candidates (Harding's EVs on the left, Cox's EVs on the right)

Alabama - 3/9
Arizona - 3/0
Arkansas - 3/6
California - 9/3
Colorado - 4/2
Connecticut - 5/2
Delaware - 3/0
Florida - 1/5
Georgia - 3/11
Idaho - 3/1
Illinois - 20/7
Indiana - 9/6
Iowa - 10/3
Kansas - 7/3
Kentucky - 5/8
Louisiana - 2/8
Maine - 5/1
Maryland - 5/3
Massachusetts - 13/4
Michigan - 12/3
Minnesota - 9/2
Mississippi - 1/9
Missouri - 11/7
Montana - 3/1
Nebraska - 6/2
Nevada - 3/0
New Hampshire - 3/1
New Jersey - 10/4
New Mexico - 3/0
New York - 30/12
North Carolina - 4/8
North Dakota - 4/1
Ohio - 15/8
Oklahoma - 6/4
Oregon - 4/1
Pennsylvania - 26/10
Rhode Island - 4/1
South Carolina - 0/9
South Dakota - 4/1
Tennessee - 7/5
Texas - 4/13
Utah - 3/1
Vermont - 4/0
Virginia - 4/8
Washington - 5/1
West Virginia - 5/3
Wisconsin    - 10/2
Wyoming - 3/0

Third party EVs

Debs: California (1), Illinois (1), Massachusetts (1), Minnesota (1), New York (3), Ohio (1), Wisconsin (1)

Christiansen: Illinois (1), Washington (1)

Ferguson: Texas (2)

Watkins: Pennsylvania (1)

Black & Tan: Texas (1)

I haven't done this for all of the elections, but Debs did win 17 EV in 1912.

The formula for figuring out the threshold

First: EVs - 2

Second: 1/(EVs - 2)

That should give you a percentage (refered to as "The EV Percentage" later). Then divide the percentage in half for the threshold.

And for 3 EV states, give all three EVs to the popular vote winner.

For situations where rounding up gives you too many EV, just don't round up the guy's EVs who is farthest from the next number.

For situations where you don't have enough EV, round up the person's EV who is closest to the next number.

Figuring out the number of EV goes as follows.

Divide the votes of the candidate involved with the total of the candidates who met the threshold.

Then divide the Candidate's Percentage by the EV Percentage.

Seven candidates getting an EV could be the record if I actually compute all the elections after 1828.

Why would Christensen not get the two votes from SD and WA, the states he did the best in?

[RBH]

Actually, Christensen got more votes in Illinois than in South Dakota.

[Dr. Cynic]

let's see... In fact, you're right... But his percentage in SD was much higher, so wouldn't that constitute the vote?

[RBH]

Well, let's take a look at things here.

South Dakota (3 proportional EVs)

Harding - 110692 (60.74%)
Cox - 35938 (19.72%)
Christiansen - 34707 (19.04%)
Watkins - 900 (0.49%)

The minimum to get one vote is 16.67%

Therefore, Watkins votes are dropped and disregarded.

So, we have the following results now

Harding - 110692 (61.04%)
Cox - 35938 (19.82%)
Christiansen - 34707 (19.14%)

And the raw EV totals are

Harding - 1.8498
Cox - .6006
Christiansen - .58

If you round up, then all three candidates get a vote, but that would exceed 3 votes.

Therefore, I just round up for the top two.

Which produces 2/1 EV split, and since Harding won South Dakota, he gets two more at-large votes.