My 2020 Presidential and Senate Simulator/Now Cast.

[ProgressiveModerate]

This was a project I was working on in spreadsheets
It simulates the election, and compiles a now-cast
Click the checkbox in C3 to simulate an election
Positive numbers mean a Trump victory, negative mean Biden victory
Change the average lead in Nation Average (B4) and the simulator will change the average NPV
There's two maps; one shows the election you just simulated and the other shows the compiled average
G1 and H1 show Biden and Trump's chances
A1 and B1 show how many EVs Biden and Trump win in a random simulation
To reset the compiled average, copy n re-paste M11-M68 as well as K1-N1, this will reset all the data
Experiment with different PVs and see how often a canidate would win

Would love feedback!

Link: https://docs.google.com/spreadsheets/d/18OOoAo_bU_dWA6kMfdI90MJhH-7koaEAdxUPDvv3vw8/edit?usp=sharing

Link:

[SaneDemocrat]

Looks pretty good

[ProgressiveModerate]

Looks pretty good

Thanks!

[ProgressiveModerate]

After 100 simulations with the average PV at 9.6% in Biden's favor:

Biden won 95.3% of the time

States:

100% Safe Trump (Never flipped in any simulation): ID, WY, ND, SD, NE-AL, NE-3, OK, AR, TN, AL, KY, WV

100% Safe Biden: WA, OR, CA, CO, IL, NJ, MD, DC, MA, NY, VT, RI, CT

States that flipped in at least one simulation:

Trump-Leaning States

MS: 99.9%
LA: 99.6%
NE-1: 99.1%
KS: 98.9%
MT: 98.4%
IN: 98.2%
UT: 97.5%
SC: 96.6%
MO: 94.4%
AK: 89.5%
ME-2: 60.9%
TX: 59.7%
IA: 55.5%

Biden leaning states:

DE: 99.9%
NM: 99.9%
VA: 99.8%
ME-AL: 98.7%
MI: 97.8%
NV: 97.8%
MN: 97.4%
WI: 95.8%
PA: 95.4%
NH: 95.3%
FL: 93.8%
NE-2: 88.5%
AZ: 77.5%
NC: 77.1%
GA: 64%
OH: 60.6%

Again, this is a NOWCAST and not a forecast. Still, not good news for Trump.

[ej2mm15]

Not sure if this is just a problem for me or not, but it reloads the numbers a bunch of times and then crashes some of the times I open it. Also some of the cells have error values, not sure if those are intentional.

[ProgressiveModerate]

Not sure if this is just a problem for me or not, but it reloads the numbers a bunch of times and then crashes some of the times I open it. Also some of the cells have error values, not sure if those are intentional.

Yep those are intentional errors since some states don't have polling, but the model deals with it. Also; if you spam the checkbox at C3, it will eventually crash, especially if your computer doesn't have a lot of RAM and/or CPU power. Fortunately, my laptop usually runs pretty smoothly, but if I click it 100 times at a rate of 10 times a second it'll crash, or just stop working. Lot's of long decimal numbers that computer needs to calculate for each simulation.

[Fmr. Gov. NickG]

After 100 simulations with the average PV at 9.6% in Biden's favor:

Biden won 95.3% of the time

States:

100% Safe Trump (Never flipped in any simulation): ID, WY, ND, SD, NE-AL, NE-3, OK, AR, TN, AL, KY, WV

100% Safe Biden: WA, OR, CA, CO, IL, NJ, MD, DC, MA, NY, VT, RI, CT

States that flipped in at least one simulation:

Trump-Leaning States

MS: 99.9%
LA: 99.6%
NE-1: 99.1%
KS: 98.9%
MT: 98.4%
IN: 98.2%
UT: 97.5%
SC: 96.6%
MO: 94.4%
AK: 89.5%
ME-2: 60.9%
TX: 59.7%
IA: 55.5%

Biden leaning states:

DE: 99.9%
NM: 99.9%
VA: 99.8%
ME-AL: 98.7%
MI: 97.8%
NV: 97.8%
MN: 97.4%
WI: 95.8%
PA: 95.4%
NH: 95.3%
FL: 93.8%
NE-2: 88.5%
AZ: 77.5%
NC: 77.1%
GA: 64%
OH: 60.6%

Again, this is a NOWCAST and not a forecast. Still, not good news for Trump.

It’s a cool spreadsheet, I’ll have to dig into it a little more.

But Biden winning 95.3% of time time conditional on winning the popular vote seem awfully low to me.  At that PV margin, he should win easily more than 99% of the time.  Can you show an example of a simulation where he wins by this margin but loses the EC?

[ProgressiveModerate]

After 100 simulations with the average PV at 9.6% in Biden's favor:

Biden won 95.3% of the time

States:

100% Safe Trump (Never flipped in any simulation): ID, WY, ND, SD, NE-AL, NE-3, OK, AR, TN, AL, KY, WV

100% Safe Biden: WA, OR, CA, CO, IL, NJ, MD, DC, MA, NY, VT, RI, CT

States that flipped in at least one simulation:

Trump-Leaning States

MS: 99.9%
LA: 99.6%
NE-1: 99.1%
KS: 98.9%
MT: 98.4%
IN: 98.2%
UT: 97.5%
SC: 96.6%
MO: 94.4%
AK: 89.5%
ME-2: 60.9%
TX: 59.7%
IA: 55.5%

Biden leaning states:

DE: 99.9%
NM: 99.9%
VA: 99.8%
ME-AL: 98.7%
MI: 97.8%
NV: 97.8%
MN: 97.4%
WI: 95.8%
PA: 95.4%
NH: 95.3%
FL: 93.8%
NE-2: 88.5%
AZ: 77.5%
NC: 77.1%
GA: 64%
OH: 60.6%

Again, this is a NOWCAST and not a forecast. Still, not good news for Trump.

It’s a cool spreadsheet, I’ll have to dig into it a little more.

But Biden winning 95.3% of time time conditional on winning the popular vote seem awfully low to me.  At that PV margin, he should win easily more than 99% of the time.  Can you show an example of a simulation where he wins by this margin but loses the EC?

So if I set the average NPV to 9.6% in Biden's favor, it means that sometimes you'll get a NPV of only 0-5%, which allows for a Trump EC win. So he's not winning the NPV by that much every simulation, that's just the average margin of victory.

[ProgressiveModerate]

Added the senate simulation, which correlates with the presidential results. Still need to code for GA(S) though. Click the checkbox at C3 once to run a simulation. Spamming it will result in crash crash. To change the national environment, change the number in cell B4. To reset the compiled averages, copy and repaste M1-R68. Link: https://docs.google.com/spreadsheets/d/1Zfbis37_u_WJQSh5Mqtnsl93GmhLk-P91sOEIbGHuFw/edit?usp=sharing

[ProgressiveModerate]

Finished coding the potential runoffs for senate. Now it shows you how often Ds and Rs win a majority in the senate, and how many seats each side won in any given simulation. Also, did some slight tweaking to make the senate simulation results better: https://docs.google.com/spreadsheets/d/1tEyWiu4ArLtXNiakYz7Z1oDy7ZVPepno4GJhnrDpaeg/edit?usp=sharing

[Amenhotep Bakari-Sellers]

Why do we need simulations,  its gonna come down to MI, PA and WI and AZ, CO, GA, IA, MT and NC senate. Dems are winning 3 to 4 seats not 12 and expected to lose AL
  We dont need no more simulations.

[ProgressiveModerate]

Why do we need simulations,  its gonna come down to MI, PA and WI and AZ, CO, GA, IA, MT and NC senate. Dems are wimmimg 3 to 4 seats not 12 and expected to lose AL
  We dont need no more simulations.

It's just to show a range of possibilities. You can also change the average NPV by entering a new value in cell B4 if you want to see a different sort of national environment.

[ProgressiveModerate]

Added a forecast mode. In this mode, the simulations are less realistic, but they compile to make a much more realistic forecast. You can now also change the weight of polls, depending upon how much you trust polling thus far. You can also the change the national environment to where you think it'll be on election day. Also, tweaked the senate races a bit more.

Link: https://docs.google.com/spreadsheets/d/1xUCMXizUXvq2DZmNPG1NGJ351mN-rNlbJmDhQjqnBwI/edit?usp=sharing

[Cassandra]

When I click the checkbox on C3, nothing happens. Do I need to change any settings?

Thanks for making/releasing this!!

[ProgressiveModerate]

When I click the checkbox on C3, nothing happens. Do I need to change any settings?

Thanks for making/releasing this!!

Try using this link for the updated model: https://docs.google.com/spreadsheets/d/1xUCMXizUXvq2DZmNPG1NGJ351mN-rNlbJmDhQjqnBwI/edit?usp=sharing

Now the checkbox is in C4. I didn't give anyone editing perms on the linked document so you'll need to go to file-make a copy to play around with it. It should work then, as long as C4 isn't spammed (It can take a few seconds for the simulation to complete). If it still doesn't work let e know because then maybe there is an issue I need to fix. Thanks for checking it out!

[ProgressiveModerate]

Added some more interactive features and fixed some random senate outcomes where SD, ID, and NE went blue. When you click the link go to file-make a copy, and check C4 to simulate an election once.

https://docs.google.com/spreadsheets/d/1wbHWDL4LCLlbEqGI1QTvM2ujPja4hJ6jBQ2cZLJQZYc/edit?usp=sharing

[electoralfox]

Added some more interactive features and fixed some random senate outcomes where SD, ID, and NE went blue. When you click the link go to file-make a copy, and check C4 to simulate an election once.

https://docs.google.com/spreadsheets/d/1wbHWDL4LCLlbEqGI1QTvM2ujPja4hJ6jBQ2cZLJQZYc/edit?usp=sharing

Great work! You've quickly turned a simple model into a much more complicated one on a Google spreadsheet. The chance of a Biden victory in the Electoral College seems a bit high but other than that, it's good work.

[MR DARK BRANDON]

Damn how do you people do this?

[ProgressiveModerate]

Added some more interactive features and fixed some random senate outcomes where SD, ID, and NE went blue. When you click the link go to file-make a copy, and check C4 to simulate an election once.

https://docs.google.com/spreadsheets/d/1wbHWDL4LCLlbEqGI1QTvM2ujPja4hJ6jBQ2cZLJQZYc/edit?usp=sharing

Great work! You've quickly turned a simple model into a much more complicated one on a Google spreadsheet. The chance of a Biden victory in the Electoral College seems a bit high but other than that, it's good work.

Thanks for the positive feedback! It' currently set to default to the forecasts; the now cast has Trump's chances at around 3%. The forecasts has Trump at about 13% overall. The margin of error will shrink as we get closer to election day, since there will be less and less time for things to happen that shake up the race.

[ProgressiveModerate]

Damn how do you people do this?

It seems scary but once you start working on the model, you kind of figure it out as it goes along. You begin to find a set of functions that are helpful in different situations. =norminv(rand() can be helpful when making a simulator.  I'm no expert at google sheets/excel; I really only started getting into it a year ago, and even then, I haven't done many intense projects. I'm no math genius; I'm going to be a sophomore in high school next year and still have a lot more to learn. It really isn't that hard to make if you dedicate some time to it.

[ProgressiveModerate]

https://cdn.discordapp.com/attachments/693691041353826314/731903506583257199/Simulator_Version_1.1.xlsm

Switched over to Google Spreadsheets. Now you can click the buttons to run 1 simulation, 30000 simulations, or just reset. 30000 simulations may cause lag and can take a bit to fully complete. Even if the screen isn't updating, it is still running. Added some more variables such as fundraising for senate races and looking at polls not only by margin but also canidate %s. Do note that in order to run this, you will need to have the full version of Excel with macros enabled. Working on creating a forecasts on sheet 2 based on the simulations.

[Amenhotep Bakari-Sellers]

278 map, Ds can win AZ, CO, IA and ME, IA and AZ split their votes and GA special goes to runoff for majority.  But IA can be replaced by NC or MT 2 to 5 seats

[American2020]

https://twitter.com/USA_Polling/status/1282354594551562242

[Amenhotep Bakari-Sellers]

Well, they need to start releasing Senate numbers aside from GA, NC and AZ. If we get polls from KS and MA Senate and more polls from MT and IA, we can see if it's a wave or Tsunami

[ProgressiveModerate]

Yay! Finally got the map to pretty much where I want it:



Does this look about right?