ANALYSIS — Votes are still being counted, a handful of House races are still uncalled, and control of the Senate hasn’t yet been decided, but it’s not too early to start reflecting on what I got right, and wrong, this election cycle.

‘Expect record turnout’

“It’s too early to declare a winner in the 2020 presidential race, particularly without knowing the Democratic nominee. But as long as Trump is on the ballot and part of the conversation, expect another turnout record to be shattered.”

Dec. 6, 2018

In the wake of record-breaking turnout in the midterm elections and with President Donald Trump driving turnout among Republicans and Democrats, I projected record turnout nearly two full years before the 2020 elections and well before the pandemic. That’s exactly what happened. While votes are still being tallied, it looks like turnout will be about 66 percent (which I mentioned in the Roll Call piece) with more than 155 million people casting ballots, according to the hard work by our friends at The Cook Political Report.

‘The Senate is in play’

“With President Donald Trump struggling to recreate his 2016 Electoral College victory, control of the Senate should be regarded as in play. Republicans are still more likely than not to maintain control of the Senate, but Democrats have a legitimate path to control, particularly if they win the White House.”

Oct. 11, 2019

Considering Trump lost reelection and control of the Senate hinges on two runoff races in Georgia on Jan. 5, this turned out to be a pretty solid look into the future more than a year before the elections. Over the course of the following year, as November 2020 drew closer, I thought Democrats were more likely than not to gain control of the Senate. That is still a possibility, even if the path is more narrow.

‘Democrats will be united’

“I still believe Democrats will be united in November. I know it’s hard to believe after a debate like Wednesday night, but I think we are consistently underestimating Trump’s ability to unify and energize the Democratic Party.”

Feb. 19, 2020

In the wake of a brutal Las Vegas debate, I reiterated my long-held belief that Democrats would be united in the general election. In 2016, just 89 percent of Democrats voted for Hillary Clinton. But after Trump won, the party started approaching elections with a new seriousness, and 95 percent of Democrats voted for Democratic congressional candidates in the midterms. And this year, 94 percent of self-identified Democrats voted for Biden, according to the exit polls.

House Democrats will add seats

“In the House, Democrats are poised to expand their majority. With a projected gain of 14-20 House seats, Democrats are on track to repeat the 2006 and 2008 cycles, when the party built on significant midterm gains to seal the capture of two entire branches of government two years later.”

Oct. 28, 2020

Undoubtedly my biggest whiff of the cycle, and maybe ever. Rather than a double-digit Democratic gain in House seats, Republicans are poised for a double-digit gain, depending on a handful of uncalled race results. While we accurately projected that Democrats would maintain their majority, we obviously totally missed the overall trajectory of the House elections.

The clear majority of preelection data (public and private, partisan and nonpartisan, including Democratic and Republican polling) showed Trump underperforming in key districts around the country and proving to be a liability for many GOP candidates down the ballot. Now we know there was a widespread underestimation of the president’s support and he ended up being an asset, by doing about as well as he did in 2016 in battleground districts. But to be clear, GOP strategists did not believe this scenario was particularly likely, or they would have been shouting from the rooftops that they’d come within a handful of seats of the majority.

Don’t get me wrong, I take responsibility for our projections, and I’m still a little sick about them. Not because we were rooting for a party, but because we take it seriously. This is not a game to us. I think people misunderstand that there’s literally no incentive for us to skew our ratings against what we think is going to happen, since we’re just trying to get it right. I’m still frustrated and am going to be thinking about our 2020 House projection for a long time.

It might be dangerous to say, but I’m not convinced public opinion polling is permanently broken, considering our 2018 midterm projections (and projections going back decades) were pretty darn good. I am skeptical of polling in elections when Donald Trump is on the ballot. If that ever happens again, I’d have to consider punting individual race ratings to the next cycle.

RIP, election night

“It’s time to retire the term ‘election night.’ The semiannual national tradition of staying up a few hours past bedtime to know who will control our government is over. From close races to voting by mail to human error, it’s becoming clear that counting votes no longer fits neatly into prime-time television windows. Reporters and politicos should prepare to practice patience when handling and digesting the results. … Even if Arizona delivers a clear Senate winner in November, knowing who controls the Senate may not be decided until a potential January runoff for one of Georgia’s seats.”

Feb. 21, 2010

Even before voting access was dramatically expanded to account for the pandemic, it was clear that vote counting would never be the same. This projection ended up being absolutely correct. Biden wasn’t declared the president-elect until four days after Election Day and control of the Senate won’t be known until two months after Nov. 3.

Susan Collins the underdog

“The senator is mired in the low 40s and the ranked-choice process isn’t likely to be friendly to the longtime incumbent. She probably has to clear 50 percent on the first ballot, and that looks nearly out of reach. Of course, the race isn’t over, and Collins could still win, but this doesn’t look like a Toss-up. Move to Tilt Democratic.”

Oct. 1, 2020

Swing and a miss. We knew Collins was in the fight of her life for the entire cycle and had the Maine race rated as a Toss-up for most of the last two years. But in the final weeks, it looked like she was going to fall short of 50 percent and lose under the state’s ranked-choice system. Not only did the senator win, but she received 51 percent and defeated state House Speaker Sara Gideon by 9 points. That’s just a stunning margin of victory, considering all the preelection analysis, Biden’s 9-point victory at the top of the ballot and the lack of ticket-splitting nationwide.

Biden more likely to win

“Critics will claim that Inside Elections has counted Trump out and that we’ve declared the president can’t win. That is not true. Just like the race wasn’t over when neither candidate was projected to win more than 270 votes, the race isn’t over when current projections put Biden over 270. But the fear of projecting the race incorrectly shouldn’t cause us to ignore the preponderance of data which show a Biden win is more likely than a Trump win at this stage of the race.”

July 17, 2020

While we considered the president to be the underdog for most of the race, this was the first time our state projections pushed Biden across the 270 Electoral College vote threshold. Trump ended up winning two states that we had in the Tilt Democratic category in our final preelection ratings (Florida and North Carolina), and the president swept the Toss-up states. But our overall projection of a Biden victory turned out to be correct (including Georgia going blue) with a projected 306 Electoral College votes, which makes 2020 significantly different than 2016, when Clinton was projected to win.

Shalala ‘solid’ for reelection

“In early September, Republicans released a poll, which showed former TV news anchor Maria Elvira Salazar leading the congresswoman by 3 points. But that poll was either an outlier or the race has shifted dramatically back to something close to the district Hillary Clinton carried by nearly 20 points in 2016. Republicans are going to have a hard time taking this district away from Shalala. Solid Democratic.”

Oct. 16, 2020

“Just a bit outside,” as Bob Uecker said in “Major League.”

While there were a number of close races that were not on our list of competitive districts, this was the only race rated as Solid that switched party hands. It wasn’t a complete shock, considering the preelection narrative about Biden’s troubles with Hispanic voters in Miami-Dade County, but there was limited data to support the scenario. Now we know the GOP optimism was real, not just in the neighboring 26th District, but here, buoyed by Trump’s strong performance at the top of the ballot.

Once the final tallies are in and the Georgia runoffs are over, there will be more opportunities to analyze what we got right and what we missed. I’m sure you all will hold me to it.

Nathan L. Gonzales is an elections analyst with CQ Roll Call.