i'm telling you the blunt reality about the midterms. this is not a general election. this is a series of completely unrelated experiments that are happening. further, there is no existing system in canada or the uk that has any passing resemblance to congressional elections in the united states - no senate and no comparable house. voters in kansas are being asked to pick a democrat as the more godly candidate. voters in maine are being asked to vote with their guts instead of their brains. there's cultural reasons they keep electing a republican. these hyperlocal races about specific issues on the ground cannot be measured with national polling, which includes states like california and new york that aren't even having elections. it's a completely fucking wrong way to look at solving this problem.
of the roughly 10 independent experiments that will be performed on tuesday, none of them have an obvious outcome. these are all very uncertain experiments. even the races that are being presented as having more certain results, like georgia, are not clear. 60% is not very good odds when there's only two outcomes. that means that the opposite thing happens four times every time you run ten experiments.
the mathematical theory - and it has a lot of empirical support - is that you just multiply the probabilities together. probability theory is an analytical theory with empirical support. it works, but it's axiomatic. it's actually more of an art than a science. probability theory is not objectively true, it's a way to approximate an idea that isn't entirely quantifiable. but empirical tests show it's roughly accurate.
the theory makes a pretty clear prediction - as no single one of these events is very likely, multiplying them together very quickly decreases the likelihood that several of them will happen simultaneously. that's not a "political analysis" and it's not my opinion (these are two different ways to describe the same thing). it's just a simple application of high school mathematics. it is the correct way to look at solving this problem.
that being said, i could give you my gut feeling, too. it's not worth much. opinions are like assholes, right? i'm used to being an asshole, here. if you want my opinion, instead, here you go.
my gut tells me that the democrats will win north carolina and the republicans will win michigan. nothing else will change.
the result is no shift in power, except in states like texas and kentucky (and south carolina), where trump replaces a critic with a crony.
that's my gut.
but the math says we have a number of independent experiments being conducted and that the democrats are not very likely to succeed at any one of those experiments. that makes it exceedingly unlikely that they will succeed at many or all of those experiments and overwhelmingly likely that the republicans hold the senate. that's the math.
you can listen to the math or you can listen to the arts majors with their opinions, or you can do both, and weight them according to their relative value. the opinions presented by the poli sci majors should not be considered very valuable, in this hierarchy. but it's up to you. the math doesn't care, in the end.