How Accurate Was Toss-Up Central?

Votes are still being counted in races all across the country. However, we have a pretty good idea of how each race will go. Joe Biden is the projected winner of the 2020 Presidential Election, winning 306 Electoral Votes to Trump’s 232. This is the first election cycle Toss-Up Central has covered. It’s a good exercise to look at my accuracy rate, examine the races I miscalled, and compare them to other election outlets.

One aspect of my site I’m proud about is that I leave no race in Toss-Up going into Election Day. I think it’s cowardly to leave people in suspense about close races. There are three other major bi-partisan election websites that call every race, Larry Sabato’s Crystal Ball, CNalysis, and Elections Daily. Throughout this article, I’ll be comparing my accuracy to theirs.

The Electoral College

There are 56 separate contests in the Electoral College, all 50 states, Washington DC, Maine’s 2 congressional districts, and Nebraska’s 3 congressional districts. Of these 56 contests, I correctly called 53, for 94.6% accuracy. I predicted a Biden victory of 334 Electoral Votes. It looks like he’ll end up with 306, for a Biden overestimation of 28 Electoral Votes.

Biden’s improvement among urban working class voters and in his native eastern Pennsylvania didn’t catch me off-guard, nor did Biden’s improvement among suburban voters and seniors in the Phoenix area.

My only three miscalls were in Florida and North Carolina, favoring Trump, and in Georgia, favoring Biden. In Florida, I underestimated how much Cubans and Venezuelans would swing towards Trump. The Latinx swing towards Trump was astonishing. The results in Georgia and North Carolina were bizarre. North Carolina has voted to the left of Georgia for a long time, and the two states share striking similarities in geography, demographics, and the population in urban areas compared to rural areas. But metro Atlanta swung much farther left then metro Charlotte or Raleigh. Some people credit this to Stacey Abrams, a former gubernatorial candidate turned activist, who was able to produce Democratic votes more than Democrats were able to in North Carolina. Others credit this to the growing film industry in Atlanta, causing some to dub the city as “Hollywood East”.

Website	Toss-Up Central	Sabato’s Crystal Ball	CNalysis	Elections Daily
Electoral College Accuracy	94.6%	98.2%	92.9%	92.9%
Biden Overestimation by Electoral Votes	+28	+15	+83	+29

Kudos to Sabato’s Crystal Ball for nailing the Presidential forecast. Their only miscall was North Carolina. As I mentioned earlier, I don’t blame them for missing that one. After their debacle in 2016 and the hiring of popular election mapmaker and analyst J. Miles Coleman in 2019, they worked hard analyzing all of their resources to repair their Presidential Forecast.

As a novice elections analyst who’s only had a website for one election cycle, who doesn’t yet have the connections, time, and resources that come with seniority, I’ll gladly accept 2nd place on the Presidential level, especially considering that, spoiler alert, I place higher down-ballot.

The Senate

There were 35 Senate seats up for election Tuesday night. Of those, the two in Georgia will be decided by a runoff election on January 5, 2021, between the top-two vote getters in each race. If Democrats win both races, then they’ll gain control of the Senate. Expect Georgia to become the mecca of American politics. I plan on releasing ratings for the Georgia runoffs, as well as for the other 33 and counting Senate races in 2022, next week.

Of the 33 races where the result is settled, I correctly called 31, for 94.0% accuracy. The rating change I was most proud of wasn’t a rating change at all: I never put the Iowa Senate race in Toss-Up, even as others moved it to Lean Democratic. There was much hype in Democratic circles around Teresa Greenfield, their Senate candidate in Iowa, as poll after poll showed Greenfield up. I stubbornly kept the race in Leans Republican. I just didn’t think Greenfield was that much better of a candidate than Joni Ernst, the incumbent Republican. Republicans would come home to Trump and Ernst in the end. In Iowa and elsewhere, I was right to be skeptical of Democrat efforts to expand the map beyond the Lean Biden states.

My two miscalls were in Maine and North Carolina. In Maine, I underestimated the cross-over appeal Susan Collins still had in blue Maine. Even the most optimistic Republicans thought Collins was done for and that Democrat Sara Gideon’s superior campaign would send her to DC. Turns out there’s still a lot of split-ticket voters out there. North Carolina was the state that gave me the most trouble this cycle. There were a number of people that voted for Joe Biden, but couldn’t pull the lever for Cal Cunningham after his extramarital affair.

Website	Toss-Up Central	Sabato’s Crystal Ball	CNalysis	Elections Daily
Senate Accuracy	94.0%	94.0%	94.0%	94.0%

Not only did all four sites tie for 1st, all of us also got the same two states (Maine, North Carolina) wrong. The election cycle isn’t over. Those two Georgia runoffs will also decide the accuracy contest between the four websites.

The House

Republicans outperformed expectations in the House by a lot. Of the 428 House races where the outcome is certain, I miscalled ten, for 97.7% accuracy. All of my miscalls favored Republicans. Eight races remain uncalled, so my accuracy could still drop to as low as 96.0%. Let’s look at the other websites.

Website	Toss-Up Central	Sabato’s Crystal Ball	CNAlysis	Elections Daily
# of Races Miscalled	10-18	16-24	21-29	13-21
Accuracy	96.0-97.7%	94.5-96.3%	93.3-95.2%	95.2-97.0%

Toss-Up Central is the clear frontrunner in the race for House accuracy. Throughout the entire election cycle, I wasn’t convinced that Democrats would, after a blue wave in 2018 that was likely their high-water-mark, gain that many more House seats. I found it more likely that there would be no significant net change in the House, and that was reflected in my final predictions, where I predicted Democrats would gain just 2 House seats, compared to Elections Daily, Sabato’s Crystal Ball, and CNalysis, which predicted D+9, D+10, and D+14 respectively.

Turns out we were all wrong. Many more voters than expected voted for Biden, but also for Republicans down ballot. There were a lot of voters that didn’t like Trump, but also preferred Republicans to Democrats when it came to fiscal policy. The GOP improved by a lot among Hispanic and Asian voters. House Republicans defied expectations and are now looking at a gain of about 9 seats, +/- 3, cutting their current deficit by at least half.

In House predictions, I missed the bulls-eye, but I at least hit the target. The same can’t be said for other election websites.

Thank you for reading Toss-Up Central. I hope you enjoyed my articles this cycle as much as I enjoyed writing them. If you want more Toss-Up Central, visit my home page.