Skip to main content

Posts

FDL Analysis: Lessons from the 2025 Democratic Sweep

The clean Democratic sweep in the 2025 elections came as no surprise to FDL Review: On November 3, we correctly predicted that Democrats would win the races for New Jersey governor and for Virginia governor, lieutenant governor, and attorney general. But we underestimated Democrats’ performance by an average of 6.25 percentage points. This article assesses where we went right and wrong in our 2025 predictions, concluding with recommendations for how Republicans can fare better in 2026. New Jersey: Our model predicted that Democratic nominee Mikie Sherrill would defeat Republican nominee Jack Ciattarelli by 2.55 percentage points. Current results show that we were directionally correct but that we underestimated the magnitude of Sherrill’s win by 10.96 points—the single worst error in FDL modeling history. Two factors contributed to our Garden State blunder. First, the governor-elect massively outperformed pre-election polls. Bad polling is nothing new for New Jersey, which a recent W...
Recent posts

FDL Prediction: Democrats Sweep 2025 New Jersey, Virginia Races (11/3/25)

Two Democratic-leaning states—New Jersey and Virginia—are holding statewide elections in 2025. The FDL Review election model predicts that Democrats will sweep these elections, retaining the governorship of New Jersey and flipping the governorship, lieutenant governorship, and attorney generalship of Virginia. The races for New Jersey governor and Virginia attorney general are, however, exceedingly close. Our model considers public polling (adjusted for past accuracy), state fundamentals, and voter enthusiasm for each candidate. New Jersey: Our model predicts that Democratic Congresswoman Mikie Sherrill will narrowly defeat Republican former Assemblyman Jack Ciattarelli in the Garden State’s gubernatorial race, winning by 2.55 percentage points. Boosting Sherrill’s chances of victory, New Jersey’s fundamentals favor Democrats. It is a solidly blue state where 38% of voters are Democrats and only 25% are Republicans. Indeed, it last elected a GOP governor in 2013 and hasn’t voted for ...

Final Scorecard: FDL Model's 2024 Election Predictions Almost Perfect

The clean Republican sweep in the 2024 elections came as no surprise to FDL Review: On November 4 , we predicted that Republicans would win the House, Senate, and presidency, even though betting markets thought there was only a 36% chance of this happening. Now that the results are certified, here’s our final scorecard. Presidential Election: 96% Right We accurately called 48 of the 50 states in the presidential election, where President Donald Trump managed the best result for a Republican since George H. W. Bush in 1988. The only states we got wrong were Michigan and Nevada. Across the ten battleground states, our final projection was off by an average of 1.65 points and overestimated Vice President Kamala Harris by 0.66 point. We were vindicated in our prediction of a Trump victory in the national popular vote, which betting markets thought he had only a 26% chance of winning. He outdid our final estimate of a 0.73-point victory, winning the popular vote by  1.47 . Our sub-arg...

Final FDL Prediction: Trump 291–247; GOP 52–48 in Senate, 225–210 in House (11/4/24)

The FDL Review election model predicts that former President Donald Trump will win reelection, that the GOP will retake the Senate, and that Republicans will retain their House majority. Trump is expected to win the Electoral College 291–247, taking the battleground states of Arizona (by 3.81 percentage points), Georgia (1.50), North Carolina (2.45), Pennsylvania (0.18), and Wisconsin (5.21). Additionally, our model predicts that Trump will win the national popular vote by 0.73 percentage point. Vice President Kamala Harris, meanwhile, is expected to take Michigan (by 1.45 percentage points), Minnesota (4.28), Nevada (0.73), New Hampshire (2.85), and Virginia (5.18). Giving Harris a boost at the margins, her voters are a hair more excited than Trump’s   about participating in the election. And  recent polls indicate that the vice president has momentum, winning over “late deciders” by double digits. Our model also predicts that Republicans will flip the Senate, netting three...

Trump Electoral College Advantage Buckles Despite His Popular Vote Lead Doubling; GOP Holds Steady in Senate, Expands House Lead: FDL Model (10/26/24)

If the 2024 elections were held today, the FDL Review model finds that Republicans would retake the Senate and presidency while holding the House of Representatives. This is unchanged from our findings last week. Over the past week, former President Donald Trump’s advantage in the Electoral College shrank from 44 votes to six. This is because Vice President Kamala Harris captured a 0.007-point advantage in Pennsylvania, which Trump led last week by 0.11. Hurting Trump at the margins, polling indicates that Harris’s supporters are marginally more enthusiastic and certain to vote than the former president’s. This could be the result of Trump’s lackluster ground-game operation. That said, Trump’s 272 votes are a safer bet than Harris’s 266. Our model indicates that he leads in all ‘his’ states by at least 1.46 percentage points, whereas two of Harris’s states—Pennsylvania and Michigan—are within a percentage point. Trump would win the national popular vote by 1.39 percentage points if t...

FDL Model (10/19/24): Republicans Leading 291–247 in Electoral College, 52–48 in Senate, 226–209 in House

If the 2024 elections were held today, the FDL Review model finds that Republicans would retake the Senate and presidency while holding the House of Representatives. According to our model, GOP presidential nominee Donald Trump would win 291 electoral votes to Vice President Kamala Harris’s 247. This would make Trump the first president to win a second nonconsecutive term since Grover Cleveland in 1892 . Our model finds that Trump would win Arizona (by 2.75 p.p.), Georgia (1.35), North Carolina (1.86), Pennsylvania (0.11), and Wisconsin (4.85) if the election were held today. Harris would prevail in Michigan (by 0.34 p.p.), Minnesota (4.02), Nevada (0.86), New Hampshire (4.98), and Virginia (4.79). Further, our model finds that Trump would win the popular vote by 0.75 percentage point, powered by what would be the tightest presidential margins in California and New York since 2004 and 1988 , respectively. Trump would lose California by 22.83 points, compared to 29.16 in 2020 . In the ...

FDL Electoral Model (9/10): Biden 318, Trump 220

If the 2020 presidential election were held today, the FDL Review election model predicts that former Vice President Joe Biden would narrowly win the Electoral College. Per our model, Biden commands 318 electoral votes to 220 for President Donald Trump. Biden has added 21 electoral votes since our previous forecast. However, the Trump-Biden race is still up in the air. When states where neither candidate commands 50% of the vote are shifted to the "tossup" column, Biden's electoral tally drops to 232 and Trump's to 126. The FDL Review model identifies the following states as tossups: Arizona (48.29% Biden/47.14% Trump), Florida (48.83% Biden/47.55% Trump), Georgia (48.64% Trump/46.95% Biden), Iowa (49.91% Trump/46.54% Biden), Minnesota (49.79% Biden/45.81% Trump), Nevada (49.64% Biden/44.71% Trump), North Carolina (48.41% Trump/48.19% Biden), Ohio (49.9% Trump/46.23% Biden), Pennsylvania (49.94% Biden/47.07% Trump), Texas (49.89% Trump/46.77% Biden), and Wisconsin (48...