2018 Midterm Election Odds: Using FiveThirtyEight’s House Model to Find Betting Value

Nate Silver and Co. at ABC News' FiveThirtyEight.com have released their model predicting the 2018 House of Representatives race in the November midterms.

All 435 seats are up for grabs this fall, and the party that hits the 218 mark will have majority.

PredictIt, a politics betting market, not only has odds on which party will have a majority after the midterms, but it also allows you to wager on a variety of specific Congressional races — 51 of the 435, to be exact.

FiveThirtyEight predicts each race, so we can compare the lines from the two sites to see if there's any betting value.

But first, let's look at the broader House picture. Here's PredictIt, FiveThirtyEight and Smarkets (another politics speculation site) on whether the Democrats will win back control of the House:

Odds Democrats will win back control of the House in 2018 midterms

Overall, 538's classic model is the most bullish on the Democrats' chances of taking back the House, which is very interesting when you look at the odds of specific races in their model and on PredictIt.

I pulled every House race available for wagering on PredictIt and compared the probable outcomes to FiveThirtyEight's model (all odds as of Aug. 16):

Finding the best betting values in 2018 House of Representative races

One quick note on the table above: The way PredictIt works is you buy shares at a certain price, say 46 cents for Republicans in WV-03. If the Republican, Carol Miller, wins, that share is redeemed at $1 — good for a 54-cent profit (you can obviously bet more than cents, but you get the idea).

In the table above, if there is betting value (FiveThirtyEight's model is more bullish than PredictIt's), but it's still not above 50% probability to win, I did not classify it as "worth a bet."

While FiveThirtyEight is more bullish on the overall chances for Democrats, the best betting value is on Republicans, especially in districts that are historically very red.

Both points could be true: While districts that are historically close may swing toward the opposition party in November, it's possible the age of extreme partisanship could still keep the deeply red districts quite red.

>> Sign up for The Action Network's daily newsletter to get the smartest conversation delivered into your inbox each morning.

In West Virginia's third district, for example, PredictIt has it at 46% that Miller (a Republican) will win, while FiveThirtyEight has her at 93.9%.

It's certainly an interesting race: Democrat Nick Rahall represented the district from 1993-2015, and Republican Evan Jenkins is the current incumbent.

The district has voted Republican in recent Presidential elections, notably voting 73% for Donald Trump vs. 23% for Hillary Clinton in the 2016 election. Mitt Romney also dominated the vote share in 2012 over former President Barack Obama.

FiveThirtyEight writer Nathaniel Rakich wrote a piece on the "25 Districts That Could Decide the House in 2018," and none of the three districts with the biggest discrepancy between PredictIt and FiveThirtyEight made the list, which were largely districts with single-digit Republican partisan leans.

Again, it's possible the betting value could be in very red districts that are likely to retain Republican control — even though Democrats are likely to win back the House overall.