Message from @meglide
Discord ID: 776276398452113438
but it seems ... promising?
¯\_(ツ)_/¯
another lead
https://www.youtube.com/watch?v=TZ3EgGeFzNQ watching this now
@RobertGrulerEsq dunno how busy you are, you might find this interesting
I actually covered that last night I believe 🙂
wow nice
musta spaced out
Kind of quick, but I covered his flowchart which I thought was well done. Interesting project.
watching it now or listening to it while I generate a simulation of Dr. Shiva's data
am curious what you are doing but don't interrupt 😛
finish that simulation
it's a spreadsheet and here's a screenshot
essentially I assumed something like 5 percent of Dems vote for Trump and 30 percent of Reps don't vote for Trump, all Reps vote for Reps downline and no Dems vote for Reps downline
zoom on chart
I need to watch his video again
what is the x axis and y axis measuring
above/below what?
the above/below is percent above or below 1.0 ratio
that's the y-axis
x-axis is percent Republican
got it
my point is that with just some simple assumptions about human voting patterns I can get a graph to mimic what he plotted
mmm
I could dynamically change the percent of Reps that defect from Trump based on how Republican the precinct is, etc. ... maybe adjust the Dems voting for Trump etc. ... probably get any kind of curve you with just a few tweaks here and there
Again just some simple assumptions about folks voting in that particular county could explain Dr. Shiva's data
@DrSammyD did a big data grab and looked for Dr. Shiva's pattern on a county-by-county basis for every state and didn't see that pattern BUT my point is that a local level, within a county say if you just make some simple assumptions about they way folks vote in that county then you can get some peculiarities that over a larger dataset you would not expect
You have to do some min max to figure out how many people didn't vote for the top level candidate
So the diff between the lowest number vote in that party, and the highest is the minimum possible total # of none down ballot votes.
But non-down ballot votes could (and almost certainly do) exceed this number, and they mix/match to come with the numbers they have.
The problem is that aggregating the data at the county level can never show what Shiva says. If the syphon happens at the precinct level, that simply lowers the one dot you have for a county.
You'd have to scrape every precinct.
If all of the counties have this syphon, then every dot just looks like it was shifted down equally.
Seriously, Michigan non-down ballot voters are suppressed by a minimum of 50% next to MN, IL, and WI https://bitcadia.github.io/DownBallot/Outputs/CountyDownBallotDiffsByStraightMI.html
Nessel is targeting conservative websites I hear
Suppression sounds boilerplate