Getting around to answering Sardia's question about if the recent events have altered the odds of the midterm elections. Well, my two cents, anyways.
The PA election is interesting. Recall elections can be predictive, sometimes. I think this one was more informative than Georgia, if only because the GA situation had a suddenly unusually bad Republican candidate. Democrats can take advantage of such situations when they arise, but they cannot depend on such situations existing everywhere. PA was more about reading the ground, and making sound counter-moves(to include a candidate tailored to appeal to local issues). That's a strategy that *can* be used elsewhere.
Now, Lamb and Saccomb were pretty neck and neck, so it's effectively a 50/50 split. Roughly a 22% swing vs 2016, favoring the Democrats. It doesn't just matter that the democrats won, but how, so this ends up being rather more encouraging. Now, I don't think the democrats hold a 22 point edge or anything in general, it's hard to extrapolate out that much...but it's definitely a result more encouraging to the Democrats than the Republicans. Republicans in the wake of the election tried to justify it by noting that Lamb is, in some respects, relatively friendly to some conservative concerns(such as taking a "we already have plenty of gun control laws on the books, we just need to enforce them" approach typical of republicans). So, some are trying to say that he's basically a Republican. Eh, you always see post-hoc justification of how a defeat really isn't. It wasn't the outcome the party tried for though.
So, if we see a continued pattern of the Democrats providing candidates more in tune with each area's voters, then yeah, they can potentially do quite good indeed. I am not sure this is entirely universal. Some of the most red districts are getting firebrand far-left candidates. In fairness, those folks will likely lose either way, so opportunity cost there is low, and it's still better than running nobody. Look at how Republicans didn't bother to run an establishment candidate for a doomed seat, and a literal Nazi decided to run as a Republican there. That's not the kind of publicity you want, and if you want to maintain good branding for the party, running candidates even for the doomed seats is a good policy. It may not translate to direct, immediate power, but in the long run, it matters. So, that's a second way in which Democratic enthusiasm helps them. It'll contribute to a blue resurgence at some point, even if 2018 isn't great.
Last but not least, we got Trump himself. He's...an odd duck. Generally speaking, I believe while he is very big on promoting himself, he's not as good at promoting others. Therefore, I'm not sure he's good at getting down-ticket support for other candidates even from those voters who like him. You've got the image issue, where he ran as anti-establishment to an extent, and that's...not particularly helpful to many republican incumbents, who have a difficult time portraying themselves as anything but the establishment. And then you've got his priorities, which mostly center around his activities and his image. He's unlikely to do as much on-the-road campaigning for down-ticket races as, say, Obama. He's also much less likely to build a brand around the party, rather than around himself.
Now, that ball of wax is pretty hard to enumerate. Not really sure what methodology to use for a lot of this, but continued eyeballing should give us an idea as it gets closer. I'm curious how relevant others think these factors will be.
Swapping back to hard data for a moment, Jersey and Virginia had about a 3% democratic edge in their 2017 elections. Now, not every state is the same, but given the special election data and the other factors listed above, I feel like there's pretty good odds that this is true elsewhere. May not seem like a ton, but even a coupla percents matter, at least for the house. Predictit's numbers still favor(somewhat) a republican senate and a democratic house, if that's your bar. That said, they also believe, still, there is a really significant change of Trump not being president by year's end. Like last year, I expect this to be a moneymaker for me.