100+ Simulations until election day: Day 4

 So, I did 10 more simulations to add to the pile.

Simulation 61

So I'm going to admit, this one popped up right when I was fixing my simulator after removing New Jersey from the list of swinf states, and I didnt even resize the map yet. But I've been wanting this one to come up so I grabbed it. This is the most likely trump win, if trump overperforms the polls. he wins all 7 swing states, but falls short of taking anything else. This has a very high probability of being the REAL map. This is also the only not entirely random one of the bunch. it came up randomly, but I wanted it to come up at some point, so when i saw it, i screenshotted it immediately, even though I wasn't doing this exercise at the time.

Simulation 62

Another harris loss, nothing special here. As I said a lot of these are boring and I have little to nothing to add.

Simulation 63

Another Harris win, just with Arizona instead of Wisconsin. 

Simulation 64

Cant win without Wisconsin here. Another Trump win.

Simulation 65

Wisconsin and Michigan defect again...another loss.

Simulation 66

Another weirdo combo, where Harris only wins like 2 states and Trump wins. Would you believe all of the scenarios posted are to my knowledge unique?

Also, red minnesota again...

Simulation 67

Another weird map with red minnesota and new hampshire.

Simulation 68

WI/MI/PA/GA except VA defects. Well, they did just remove people from their voter rolls so...who knows? it can happen.

Simulation 69

Basically what my prediction is now. 

Simulation 70

Blue Ohio and Kansas but red Minnesota, another oddball map. 

Tally so far..

48 Trump, 21 Harris, 1 tie

Idk what it is with this sample being so different than my normal samples which are a lot more even. For this exercise I'm getting far more trump wins than this simulator normally gives me. When I do 100 simulations and record the results, I'm using this exact simulator, with the exact data I'm using now. And while yes, the data is changing day to day as polling is bonkers right now, the overall probabilities of things arent radically changing.

It's like if I do a small sample of 10 it gives me trump outcomes but if i do a larger sample overall i get more harris outcomes. it's really weird. Part of me suspects the random number generators arent actually fully random, although any attempt to study their randomness actually gives me fairly uniform results over the long term. But in the short term i get more red outcomes. it's weird. Idk why. 

Anyway that's today's batch. I'll keep going like this until i hit 100 then I'll use the other spreadsheet with the modded variables to push out some of those more pro trump or pro harris type maps that you'd get if the candidates overperform but don't come up because this simulator seems to mostly settle upon outcomes that are in the middle. As you can tell, very few outcomes exceed the 226-312 or 319-219 thresholds of both candidates and the ones that do tend to be in the top 2% of outcomes for each candidate respectively. To compensate for this, I would like to eventually do some maps with a 4% overperformance for either candidate. But again, doing 100 normal simulations at first.