President: https://www.270towin.com/maps/Db3xgThese ratings are just for now and they'll shift for certain once we have an idea of the theme of the primary, who the front runners are (so not until next February,) and if there are big changes (like a recession or something really convincingly bad from Mueller) then I'll shift some things left.
Senate: http://www.270towin.com/2020-senate-election/lp4GklAgain, these are subject to change once nominees come up. Texas could easily become lean R by 2019, Montana likely R, etc. My eyes are on Kansas as I really think Svaty or one of the party switchers could pull off an upset, but it's solid R for now.
House: http://www.270towin.com/2018-house-election/wpkXExVSome things to explain here other than the currently unknown challengers:
-with districts susceptible to a special election i'm averaging what I think will happen in the special vs what will happen with the current incumbent in 2020. For example, Nate McMurray would win a special in NY-27 but Collins would be favored in 2020 if there's no special. Mark Harris is assumed the incumbent in NC-9.
-i'm putting a big premium on suburban Dems and I don't see the GOP clawing more than 6 of those districts back. The opposite with longtime Dem incumbents in Trump-won friewal districts - Ryan, Bustos, Kind, Loebsack start as likely D and can go either way.
-Notice that outside of PA, there are 7 seats that flipped this year that I rate safe: CA-21, CA-25, CA-48, CO-6, MN-1, MN-8, VA-10. A lot of the prognosticators also like to call FL-7 lean/likely D but Murphy's not losing w/ Trump on the ballot.
-NC ratings are for the current form of the districts and will change once a new map is implemented. I'd probably move NC-2, 6, and 13 to tossup when that time comes.