Sabato's Crystal Ball
A University of Virginia political-analysis publication that covers U.S. elections and publishes electoral forecasts.
Last updated August 25, 2026
AI crawler reads
No AI crawler has read this profile since Oct 28, 2026.
Overview
Sabato's Crystal Ball is an online political newsletter and election-analysis publication operated by the University of Virginia Center for Politics in Charlottesville. The publication is associated with political scientist Larry Sabato, the Center's founder and director, and focuses on interpreting electoral developments rather than functioning as a conventional polling organization or commercial news outlet. Its coverage encompasses presidential, congressional, and gubernatorial contests across the United States, combining race ratings, political commentary, historical context, and projections of likely outcomes. The publication's public identity is built around the idea of a forecasting service: it evaluates the competitiveness of individual contests, assesses shifts in the national political environment, and periodically publishes projections for the Electoral College, the U.S. Senate, the House of Representatives, and governorships. Its work is distributed online and is presented as free political analysis associated with the University of Virginia's civic and educational mission. The Center for Politics itself promotes civic engagement and participation, giving the Crystal Ball an institutional setting distinct from a partisan campaign, political consulting firm, or commercial forecasting company. The publication became particularly visible through its election predictions. Reference material describes strong performances in several election cycles, including the 2002 midterms, the 2004 elections, the 2006 midterms, the 2008 presidential election, and parts of the 2012 and 2020 cycles. Those records contributed to its reputation among U.S. political media as an influential forecasting source. Its projections have also been publicly scrutinized. The 2010 and 2012 forecasts contained notable misses or underestimated the scale of eventual results, while the 2016 presidential and congressional forecasts were widely incorrect. After the 2016 election, Sabato and the Crystal Ball staff acknowledged the errors in a self-critical retrospective. The 2020 forecast was considerably closer overall, although it incorrectly projected North Carolina and missed some individual Senate outcomes. In 2024, the publication forecast a Kamala Harris victory over Donald Trump, but Trump won the Electoral College by a substantial margin. The brand's authority is therefore based not only on successful forecasts but also on its academic affiliation, regular methodological commentary, and willingness to revise or assess its own calls. At the same time, its association with an individual public commentator means that coverage of Larry Sabato's political views can affect perceptions of the publication. In 2009, questions were raised about the Center for Politics' receipt of earmarked federal funding connected to Representative Virgil Goode, whose reelection Sabato had predicted in 2008 before Goode narrowly lost. Critics argued that the financial connection should have been disclosed or that Sabato should have recused himself from the race analysis. The controversy became part of the broader discussion about transparency and perceived conflicts of interest in election forecasting. Sabato's Crystal Ball remains a recognizable U.S. political-analysis brand under the University of Virginia Center for Politics. Its principal offerings are written election analysis, race ratings, and electoral projections, with activity concentrated around primary and general-election cycles.
History
Sabato's Crystal Ball is a publication of the University of Virginia Center for Politics, an institution founded and directed by political scientist Larry Sabato. The Center's broader mission is civic education and participation, while the Crystal Ball applies political research and electoral judgment to current U.S. campaigns. It operates from Charlottesville and publishes online rather than as a printed newspaper or a conventional polling business. The publication covers the full range of major American electoral contests. Its subject matter includes presidential elections, Electoral College calculations, U.S. Senate and House races, and gubernatorial contests. Its work generally combines narrative analysis with classifications of race competitiveness and explicit projections of likely winners. The format makes it useful to journalists, political professionals, researchers, students, and politically engaged readers seeking an institutional alternative to campaign messaging or purely partisan commentary. Its reputation was shaped by several strong early records. In 2002, the reference material credits the Crystal Ball with correctly calling nearly all House and Senate contests. It performed similarly well in 2004, when it reportedly forecast 525 of 530 political races. The 2006 midterm projections correctly anticipated the scale of Democratic House gains and all of the Senate predictions cited. In 2008, its presidential forecast put Barack Obama in the White House with an Electoral College margin very close to the final result; the publication also correctly called most of the Senate and gubernatorial contests listed in the source material. These results helped establish the Crystal Ball as a prominent election-forecasting voice and contributed to recognition from U.S. political-media and journalism organizations. Forecasting performance became more mixed in subsequent cycles. The 2010 projection captured the direction of a Republican wave but underestimated the House gain and overestimated the Senate gain. In 2012, it correctly predicted Barack Obama's reelection but underestimated the size of his Electoral College victory and Democratic gains in Congress. The 2016 cycle was a major setback: the publication forecast a Hillary Clinton victory and missed a broad set of congressional outcomes. Sabato and the staff responded with an unusually direct post-election acknowledgment of error, which became part of the publication's public record and demonstrated that its forecasts are revisable judgments rather than guarantees. The 2020 cycle produced a closer overall presidential assessment, although the publication projected the wrong winner in North Carolina and missed certain Senate-level changes. In 2024, it forecast a narrow Electoral College victory for Kamala Harris over Donald Trump shortly before the election, but Trump won decisively in the Electoral College. These episodes illustrate both the visibility and the risks of the Crystal Ball's model of explicit public forecasting: correct calls can build authority, while incorrect calls are easily measured and publicly revisited. The brand also faced a transparency controversy in 2009. The Center for Politics was reported to have received more than $7 million in earmarked funds connected to Representative Virgil Goode, whose 2008 reelection Larry Sabato had predicted despite weakening polling indicators. Goode narrowly lost. Critics argued that the financial relationship should have been disclosed or that Sabato should have avoided making the prediction. The episode raised questions about perceived conflicts of interest and institutional disclosure, although the reference material does not describe a formal finding that invalidated the publication's forecasts. Sabato's own public role is inseparable from the publication's identity. He is a University of Virginia political scientist, author, media commentator, and Center for Politics director. His public criticism of Donald Trump has also generated political reaction, including a 2021 call by the Republican Party of Virginia for a university investigation into Sabato's tweets. That dispute concerned Sabato's public commentary and the perceived neutrality of the Center rather than a separate commercial product line. Today, Sabato's Crystal Ball remains an active, university-affiliated political publication. Its core proposition is free, academically associated analysis of American elections, supported by race ratings, explanatory articles, and electoral projections. It is best understood as a media and research brand within the University of Virginia Center for Politics rather than as an independent company.
- 2024
Incorrect Harris presidential forecast
[1]Crystal Ball projected Harris to win the Electoral College, but Trump won the election.
- 2020
Closer presidential forecast
[1]The publication correctly projected Biden's presidential victory but incorrectly called North Carolina and some Senate outcomes.
- 2016
Public acknowledgment of forecast errors
[1]After incorrectly projecting a Clinton presidential victory and missing congressional outcomes, the publication issued a self-critical retrospective.
- 2009
Earmark-related transparency controversy
[1]Questions arose over the Center for Politics' financial connection to Representative Virgil Goode while Crystal Ball was forecasting his reelection.
- 2008
Obama presidential forecast
[1]Crystal Ball projected an Obama Electoral College victory close to the eventual result.
- 2006
Correctly projects Democratic congressional gains
[1]Its projected House gain matched the November result, and all cited Senate predictions were correct.
- 2004
High-accuracy election cycle
[1]Reference material credits Crystal Ball with correctly forecasting 525 of 530 political races.
- 2002
Strong midterm forecasting record
[1]The publication reportedly correctly predicted 433 of 435 House contests and 32 of 34 Senate races.
Products and positioning
Academic, nonpartisan-oriented political analysis and election forecasting distributed through the University of Virginia Center for Politics.
Political analysis
Election race analysis
Written coverage of U.S. presidential, Senate, House, and gubernatorial contests. Articles assess campaign conditions, candidate strength, political trends, historical context, and the factors that may influence the eventual result.
Election forecasting
Race ratings and forecasts
The publication's signature offering is its public assessment of contest competitiveness and likely winners. Forecasts are issued for individual races and, in presidential cycles, for the Electoral College, making the product readily comparable with election-night results.
Presidential election analysis
Electoral College projections
Crystal Ball publishes national presidential projections that translate state-level assessments into an expected Electoral College result. These projections have included numerical estimates of each candidate's electoral vote total.
Controversies
2009
Earmark funding and Virgil Goode forecast controversy
The Center for Politics was reported to have received more than $7 million in earmarked funding associated with Representative Virgil Goode, whom Larry Sabato had predicted would win reelection in 2008. Goode narrowly lost, and critics argued that the financial relationship should have been disclosed or that Sabato should have recused himself.
Brand decisions
2016
Publishes a post-election forecast reviewOtherBackground
The publication's presidential and congressional forecasts were widely wrong in the 2016 election.
Action taken
Sabato and the Crystal Ball staff publicly acknowledged the mistakes in a self-critical article.
Aftermath
The review became an explicit record of the publication's forecasting limitations and its willingness to evaluate its own performance.
2012
Expands separate projections for split-vote congressional districtsStrategyBackground
The 2008 presidential forecast did not separately project Nebraska's second congressional district, which awards an Electoral College vote independently of the statewide result.
Action taken
Beginning in the 2012 cycle, Crystal Ball began making separate projections for such districts, according to the reference material.
Aftermath
The change made its presidential forecasting framework more granular in states where Electoral College votes can be allocated by congressional district.
Recent events
- 2024
2024 presidential forecast incorrectly favors Kamala Harris[1]
Other
One day before voting, Crystal Ball projected Harris to defeat Donald Trump 276–262 in the Electoral College while noting the difficulty created by close polling. Trump ultimately won 312–226.
- 2020
2020 forecast is closer overall but misses North Carolina[1]
Other
Crystal Ball projected Joe Biden to win 321–217 in the Electoral College. Biden won 306–232, with North Carolina identified as the publication's only incorrect presidential state call in the reference material; several Senate state forecasts were also wrong.
- 2016
Crystal Ball acknowledges broad 2016 forecasting errors[1]
Other
The publication incorrectly forecast Hillary Clinton's presidential victory and also missed important Senate and House outcomes. Sabato and his staff published a self-critical review of the errors after the election.
- 2012
2012 forecast broadly identifies the presidential outcome but underestimates down-ballot results[1]
Other
The publication predicted Barack Obama's reelection and a 290–248 Electoral College result, but underestimated Obama's eventual Electoral College total and Democratic gains in both congressional chambers.
- 2010
2010 congressional projections miss the scale of Republican gains[1]
Other
Crystal Ball projected a 55-seat Republican House gain and an eight-seat Senate gain; Republicans ultimately gained 63 House seats and six Senate seats.
- 2008
Crystal Ball projects Barack Obama as the 2008 presidential winner[1]
Other
The publication forecast an Obama victory with a 364–174 Electoral College margin, close to the eventual 365–173 result, and correctly forecast most Senate and gubernatorial contests described in the reference material.
- 2006
2006 forecasts accurately anticipate Democratic congressional gains[1]
Other
Crystal Ball projected a 29-seat Democratic gain in the House and a six-seat Senate gain; the House projection matched the November result before a later runoff added another Democratic seat.
- 2004
Crystal Ball posts a highly accurate 2004 election record[1]
Other
Reference material credits the publication with correctly forecasting 525 of 530 political races, including presidential, congressional, gubernatorial, and Electoral College outcomes.
- 2002
Crystal Ball reports strong accuracy in the 2002 midterm forecasts[1]
Other
The publication was reported to have correctly projected 433 of 435 House contests and 32 of 34 Senate races in the midterm elections.
Sources2 sources
Cite this profile: Cite the canonical profile. /brand-wiki/sabatos-crystal-ball · Editorial policy · How profiles are compiled