Washington
RCP shows Washington State with 2 toss up races…District #5 & #8.
I don't know much about Washington, except that the Seattle area is blue, and the rest of the state is reliably red.
District #5 is on the eastern side of the state. Red territory. The race is between incumbent Cathy Rodgers (R) and Lisa Brown (D). RCP has no polling data listed. In experience, an incumbent has an advantage over a challenger, and that area of the state normally votes R. I chalk this one up as leaning R.
District #8 is the southeastern suburbs of Seattle. The race is between Dino Rossi (R) (a current state senator) and Kim Schrier (D). According to the polling data, the race is a dead heat. Rossi has more name recognition.
When you dig a little into the poll data you notice that they talked to 505 people, but they called over 40,000 people. I'm pretty sure the calls are only to people with landlines. How many people still have landlines in 2018? Also, the info on the district listed on the NY Times site states that the district has always a Republican represent it in the House. There is no incumbent in this race, as the sitting Republican Congressman for this district is not running this time. I chalk this one up as leaning R as well.
Only one state into my amateur analysis and already 2 toss-ups to add to the R's count. And that's without access to the underlying poll data to see if they over sampled D's or other obvious poll skewing techniques. Remember how wrong the polls were in 2016? Midterms are more about who voter turnout…not who they like when talking to a pollster.