Since I published Nothing But Numbers’ original election preview back in September, the election has been quite the roller coaster. At the time of the original post, Trump had cut so significantly into Clinton’s lead that the Presidential race was a virtual tie. Soon after, Clinton had an exceptionally strong first debate, which was then followed by the leaked audio of Trump saying he “grabs women by the p***y” – essentially admitting to sexual assault – resulting in a drastic plunge in the polls for Trump. At this point in the race, Nothing But Numbers’ probabilistic model had Clinton as high as a 97% favorite to win the Presidency.
Still, as time elapsed between the release of the lewd audio, Trump slowly began to gain in the polls. This hit a crescendo the night before Halloween, when FBI Director James Comey announced that the FBI was reviewing new emails in the Hillary Clinton private email server case. And while Comey has since sent a letter to Congress saying that the FBI’s original decision not to prosecute Clinton still stands – corresponding to a slight bump in the polls for Clinton as the election nears – the original announcement still had a pronounced effect in making the Presidential race more competitive. Now, before we delve into our final Election Prediction, let’s go over a few changes to the model since the original election preview in September:
- The Facebook “Likes” Model Now Adjusts for Ineligible Voters: The previous iteration of my model went purely based on Facebook Likes, which is fine as long as everyone who’s on Facebook is eligible to vote. That’s not the case, and in states with large populations of undocumented immigrants like California, Texas, and Nevada, undocumented immigrants who are ineligible to vote may have been skewing the results towards Clinton. In the updated version of the model Trump has gotten a boost in these states. Likewise, Clinton has received a bump in states with smaller than average undocumented immigrant populations.
- In Most States, 20% of the Model is Now Based on Demographics: While Facebook Likes were an excellent predictor of the voting results during the Democratic primaries, the Facebook “Likes” data itself is a bit outdated, with the data being from late April. As a result, I felt it necessary to add in a demographic component of the model to make sure the model isn’t over-reliant on old data. The Demographics model is calculated by looking at The Upshot’s Estimates of each state’s demographic 2012 voting history, and adjusting the 2012 vote for demographic shifts and changing vote preferences. For example, while an estimated 68% of Hispanics supported Barack Obama in 2012, our model estimates that Hillary Clinton will receive 78.5% of the Hispanic vote in 2016. Whites without a college degree, on the other hand, have shifted an estimated 10 points towards Trump since 2012.
- In States Where the Demographic and Facebook “Likes” Models Disagree by 6 Points or More, Different Weights Are Used For the Two Models: If the models disagreed so profoundly to the point where they disagreed by 6 points or more (which only occurred in 11 states), Nothing But Numbers looked at each state on a case by case basis to decide which model represented the state better. In some states, for example, The Facebook “Likes” Model seemed to be too pessimistic on Clinton’s chances relative to the demographic model simply because Bernie Sanders received the preponderance of Facebook “Likes” in the state, not Clinton. This effect was particularly pronounced in the typically blue states of Hawaii, Maine, New Mexico, and Vermont where Bernie Sanders performed well during the Democratic Primaries. As a result, for these states the Facebook “Likes” Model was reduced from 80% to 50% of the state-wide projection. On the other hand, in other states the Demographic Model seemed to miss obvious changes in voter opinion since 2012. In Utah, for example, the Demographic Model didn’t pick up on the well-documented abandonment of Trump from Utah’s Mormon population, and projects Trump to win by 50 points in Utah based on it’s largely white population and strong showing for Mitt Romney in 2012. As a result, in Utah (and Arkansas) the Demographics-based Model was reduced from 20% to 10% of the state-level projection.
- Third Party Candidates Are Now Accounted For: While adjusting for third parties makes a small difference in each candidates’s chances of winning the Presidency, it makes a noticeable difference in states like Utah and New Mexico, where third party candidates Evan McMullin and Gary Johnson are polling well.
- Nothing But Numbers Now Runs 10,000 Simulations of the Election to Provide a Probabilistic Forecast: Back in September, the model simply identified a “tipping point” state as a barometer for how the election would turn out. Now, Nothing But Numbers runs 10,000 simulations of the election every time there is a slight change in the projected popular vote to provide the most accurate election prediction.
Now, without further ado, let’s release the final iteration of my predictions, including the Facebook “Likes” Model, the Demographic Model, and Nothing But Numbers’ Official Projections, a hybrid of the two that combines the best aspects of both models into an accurate prediction:
| State | Projected Trump Vote (FB Likes) | Projected Clinton Vote (FB Likes) | Projected Trump Vote (Demographics) | Projected Clinton Vote (Demographics) | Projected Trump Vote (No Third Parties) | Projected Clinton Vote (No Third Parties) |
|---|---|---|---|---|---|---|
| Alabama | 66.5% | 33.5% | 65.2% | 34.8% | 66.2% | 33.8% |
| Alaska | 59.7% | 40.3% | 54.9% | 45.1% | 58.7% | 41.3% |
| Arizona | 52.0% | 48.0% | 49.3% | 50.7% | 51.5% | 48.5% |
| Arkansas | 54.2% | 45.8% | 64.6% | 35.4% | 55.3% | 44.7% |
| California | 29.4% | 70.6% | 33.7% | 66.3% | 30.2% | 69.8% |
| Colorado | 43.9% | 56.1% | 44.1% | 55.9% | 43.9% | 56.1% |
| Connecticut | 44.1% | 55.9% | 38.7% | 61.3% | 43.0% | 57.0% |
| Delaware | 49.1% | 50.9% | 40.8% | 59.2% | 47.5% | 52.5% |
| District of Columbia | NA | NA | 7.5% | 92.5% | 7.5% | 92.5% |
| Florida | 52.9% | 47.1% | 47.1% | 52.9% | 51.8% | 48.2% |
| Georgia | 55.6% | 44.4% | 56.4% | 43.6% | 55.8% | 44.2% |
| Hawaii | 42.8% | 57.2% | 21.7% | 78.3% | 32.2% | 67.8% |
| Idaho | 62.8% | 37.2% | 65.7% | 34.3% | 63.3% | 36.7% |
| Illinois | 39.0% | 61.0% | 40.0% | 60.0% | 39.2% | 60.8% |
| Indiana | 56.6% | 43.4% | 56.4% | 43.6% | 56.5% | 43.5% |
| Iowa | 41.7% | 58.3% | 47.3% | 52.7% | 42.8% | 57.2% |
| Kansas | 53.7% | 46.3% | 60.4% | 39.6% | 57.0% | 43.0% |
| Kentucky | 60.4% | 39.6% | 63.6% | 36.4% | 61.1% | 38.9% |
| Louisiana | 65.2% | 34.8% | 62.5% | 37.5% | 64.6% | 35.4% |
| Maine | 51.5% | 48.5% | 42.6% | 57.4% | 47.0% | 53.0% |
| Maryland | 37.0% | 63.0% | 37.1% | 62.9% | 37.0% | 63.0% |
| Massachusetts | 32.4% | 67.6% | 35.3% | 64.7% | 33.0% | 67.0% |
| Michigan | 48.6% | 51.4% | 46.4% | 53.6% | 48.2% | 51.8% |
| Minnesota | 40.6% | 59.4% | 45.1% | 54.9% | 41.5% | 58.5% |
| Mississippi | 65.9% | 34.1% | 60.3% | 39.7% | 64.7% | 35.3% |
| Missouri | 56.1% | 43.9% | 56.4% | 43.6% | 56.1% | 43.9% |
| Montana | 58.8% | 41.2% | 56.8% | 43.2% | 58.4% | 41.6% |
| Nebraska | 51.4% | 48.6% | 60.0% | 40.0% | 55.7% | 44.3% |
| Nevada | 47.0% | 53.0% | 43.3% | 56.7% | 46.3% | 53.7% |
| New Hampshire | 47.1% | 52.9% | 46.3% | 53.7% | 47.0% | 53.0% |
| New Jersey | 45.1% | 54.9% | 42.8% | 57.2% | 44.7% | 55.3% |
| New Mexico | 45.5% | 54.5% | 36.0% | 64.0% | 40.8% | 59.2% |
| New York | 36.7% | 63.3% | 33.0% | 67.0% | 36.0% | 64.0% |
| North Carolina | 54.1% | 45.9% | 53.0% | 47.0% | 53.8% | 46.2% |
| North Dakota | 60.1% | 39.9% | 59.6% | 40.4% | 60.0% | 40.0% |
| Ohio | 53.5% | 46.5% | 50.4% | 49.6% | 52.9% | 47.1% |
| Oklahoma | 62.8% | 37.2% | 66.2% | 33.8% | 63.5% | 36.5% |
| Oregon | 42.1% | 57.9% | 42.4% | 57.6% | 42.2% | 57.8% |
| Pennsylvania | 53.5% | 46.5% | 47.5% | 52.5% | 52.3% | 47.7% |
| Rhode Island | 47.5% | 52.5% | 33.4% | 66.6% | 40.5% | 59.5% |
| South Carolina | 60.8% | 39.2% | 58.8% | 41.2% | 60.4% | 39.6% |
| South Dakota | 57.5% | 42.5% | 58.2% | 41.8% | 57.7% | 42.3% |
| Tennessee | 61.8% | 38.2% | 62.8% | 37.2% | 62.0% | 38.0% |
| Texas | 49.6% | 50.4% | 52.9% | 47.1% | 50.3% | 49.7% |
| Utah | 52.4% | 47.6% | 73.0% | 27.0% | 54.5% | 45.5% |
| Vermont | 42.5% | 57.5% | 31.6% | 68.4% | 37.1% | 62.9% |
| Virginia | 47.0% | 53.0% | 49.0% | 51.0% | 47.4% | 52.6% |
| Washington | 36.0% | 64.0% | 40.6% | 59.4% | 36.9% | 63.1% |
| West Virginia | 64.9% | 35.1% | 66.0% | 34.0% | 65.1% | 34.9% |
| Wisconsin | 47.9% | 52.1% | 46.4% | 53.6% | 47.6% | 52.4% |
| Wyoming | 65.9% | 34.1% | 70.5% | 29.5% | 66.8% | 33.2% |
| Maine 1st | 44.9% | 55.1% | 36.1% | 63.9% | 40.5% | 59.5% |
| Maine 2nd | 58.6% | 41.4% | 49.8% | 50.2% | 54.2% | 45.8% |
| Nebraska 2nd | 37.9% | 62.1% | 46.5% | 53.5% | 42.2% | 57.8% |
| State | Projected Trump Vote% | Projected Clinton Vote% | Trump Win State% | Clinton Win State% |
|---|---|---|---|---|
| Alabama | 62.4% | 31.9% | 100.0% | 0.0% |
| Alaska | 51.7% | 36.4% | 98.5% | 1.5% |
| Arizona | 47.9% | 45.2% | 64.5% | 35.5% |
| Arkansas | 51.7% | 41.8% | 95.0% | 5.0% |
| California | 28.5% | 65.7% | 0.0% | 100.0% |
| Colorado | 39.8% | 50.7% | 7.7% | 92.3% |
| Connecticut | 40.1% | 53.2% | 2.1% | 97.9% |
| Delaware | 44.3% | 48.9% | 25.3% | 74.7% |
| District of Columbia | 7.2% | 88.1% | 0.0% | 100.0% |
| Florida | 49.6% | 46.2% | 67.7% | 32.3% |
| Georgia | 53.1% | 42.2% | 93.3% | 6.7% |
| Hawaii | 29.8% | 62.6% | 0.0% | 100.0% |
| Idaho | 57.8% | 33.5% | 100.0% | 0.0% |
| Illinois | 37.0% | 57.4% | 0.1% | 99.9% |
| Indiana | 52.4% | 40.3% | 96.8% | 3.2% |
| Iowa | 40.0% | 53.3% | 2.4% | 97.6% |
| Kansas | 51.6% | 38.9% | 97.2% | 2.8% |
| Kentucky | 57.5% | 36.7% | 99.9% | 0.1% |
| Louisiana | 61.5% | 33.6% | 100.0% | 0.0% |
| Maine | 43.4% | 48.9% | 22.9% | 77.1% |
| Maryland | 34.9% | 59.3% | 0.1% | 99.9% |
| Massachusetts | 30.5% | 62.1% | 0.0% | 100.0% |
| Michigan | 44.8% | 48.1% | 32.3% | 67.7% |
| Minnesota | 38.4% | 54.1% | 1.6% | 98.4% |
| Mississippi | 62.2% | 33.8% | 100.0% | 0.0% |
| Missouri | 52.3% | 40.8% | 95.9% | 4.1% |
| Montana | 52.4% | 37.3% | 98.1% | 1.9% |
| Nebraska | 50.7% | 40.4% | 93.8% | 6.2% |
| Nevada | 43.1% | 50.0% | 11.9% | 88.1% |
| New Hampshire | 43.1% | 48.6% | 19.3% | 80.7% |
| New Jersey | 42.8% | 52.9% | 5.7% | 94.3% |
| New Mexico | 35.0% | 50.7% | 1.4% | 98.6% |
| New York | 34.1% | 60.7% | 0.0% | 100.0% |
| North Carolina | 51.7% | 44.3% | 83.6% | 16.4% |
| North Dakota | 54.8% | 36.5% | 99.6% | 0.4% |
| Ohio | 49.5% | 44.1% | 77.3% | 22.7% |
| Oklahoma | 58.9% | 33.8% | 100.0% | 0.0% |
| Oregon | 39.2% | 53.7% | 3.5% | 96.5% |
| Pennsylvania | 49.3% | 45.1% | 74.7% | 25.3% |
| Rhode Island | 37.8% | 55.7% | 0.4% | 99.6% |
| South Carolina | 57.3% | 37.6% | 99.9% | 0.1% |
| South Dakota | 52.3% | 38.4% | 97.6% | 2.4% |
| Tennessee | 58.5% | 35.9% | 100.0% | 0.0% |
| Texas | 47.1% | 46.6% | 52.7% | 47.3% |
| Utah | 35.0% | 29.3% | 74.1% | 13.0% |
| Vermont | 34.5% | 58.7% | 1.5% | 98.5% |
| Virginia | 44.7% | 49.7% | 22.7% | 77.3% |
| Washington | 34.5% | 59.0% | 0.0% | 100.0% |
| West Virginia | 60.8% | 32.5% | 100.0% | 0.0% |
| Wisconsin | 44.8% | 49.3% | 27.5% | 72.5% |
| Wyoming | 60.1% | 29.9% | 100.0% | 0.0% |
| Maine 1st | 37.6% | 55.2% | 0.0% | 100.0% |
| Maine 2nd | 49.6% | 42.0% | 94.8% | 5.2% |
| Nebraska 2nd | 39.1% | 53.5% | 0.1% | 99.9% |
| Tipping Point State | Projected Trump Vote% | Projected Clinton Vote% | Trump Win State% | Clinton Win State% |
|---|---|---|---|---|
| Texas | 47.1% | 46.6% | 52.7% | 47.3% |
And now, the probabilistic model, fresh off of running 10,000 simulations of the election based on Nothing But Numbers’ hybrid model:
| Category | Clinton | Trump | No 270/McMullin |
|---|---|---|---|
| Chance of Reaching 270 Electoral Votes | 79.0% | 19.5% | 1.5% |
| Electoral Votes (Average) | 295.5 | 241.7 | 0.8 |
| Mode EVs (Most Likely) | 303 | 235 | 0 |
| Trump Best Case Scenario EVs | 201 | 331 | 6 |
| Clinton Best Case Scenario EVS | 385 | 153 | 0 |
As you can see, Clinton is a pretty strong favorite to win the election, winning in roughly 80% of the model’s simulations. While that’s not quite the >99% chance given by Princeton Election Consortium, it gives Clinton a slightly better chance than FiveThirtyEight’s polls-only model, which puts Clinton’s chances at 71.4%. Let’s run down the entire map, from Safe Clinton to Safe Trump and everything in between, to get a better idea of how the model arrives at it’s forecast. Keep in mind that 270 electoral votes are required to win:
Click the map to create your own at
Safe Clinton (>85% chance of Clinton victory), 217 Electoral Votes: California, Colorado, Connecticut, District of Columbia, Hawaii, Illinois, Iowa, Maryland, Massachusetts, Minnesota, Nevada, New Jersey, New Mexico, New York, Oregon, Rhode Island, Vermont, Washington, Maine’s 1st Congressional District, Nebraska’s 2nd Congressional District
Most of these are pretty typical blue states, with a few notable exceptions. While many prognosticators have Iowa going in Trump’s column later tonight, both the Facebook Likes and Demographic Model are in agreement about Clinton’s strength there. Remember, Iowa was a very bad state for Trump during the primary season, as Trump finished second in the Iowa caucuses to Ted Cruz with an uncharacteristically low 24% of the vote. The model sees a similar situation in Nebraska’s 2nd Congressional District, a district located in a state where Trump only got 61% of the vote during the primaries despite running completely uncontested (Ted Cruz had already dropped out). Traditional conservative Republicans staying home in Iowa and Nebraska’s 2nd Congressional District is the main reason Clinton is almost guaranteed to win both contests.
In Nevada, both models simply see Trump being overwhelmed by a large Latino population, leading to an easy Clinton victory. For what it’s worth, the early voting data in Nevada largely confirms this prediction. Colorado is similar, except in addition to Latinos, college educated whites also doom Trump in the Centennial State.
Likely Clinton (70% to 85% chance of Clinton victory), 32 Electoral Votes: Delaware, Maine, New Hampshire, Virginia, Wisconsin
The one big surprise here is Delaware, a state most have as Safe Clinton. The Facebook “Likes” model in particular is bullish on Trump’s chances in Delaware, so don’t be too surprised if Delaware comes in closer than expected, with southern Delaware coming in strong for Trump.
Another state to watch is Maine. Usually a safe blue state, Trump’s strong performance with whites without a college degree has put the mostly white state in play. Maine is also a state where the Facebook “Likes” model and Demographics model disagree pretty strongly. The Facebook “Likes” model has Trump pulling off a narrow victory, while the Demographics model has Clinton winning comfortably based off of Obama’s big victory there in 2012. It will be interesting to see which model comes closer to the final outcome. Personally, I think the Facebook “Likes” model is underselling Clinton’s chances in Maine, as I suspect that a lot of former Sanders supporters in Maine have joined team Clinton.
Lean Clinton (55% to 70% chance of Clinton victory), 16 Electoral Votes: Michigan
Michigan is a very interesting state for a number of reasons. While Obama won Michigan comfortably in 2012, this time around both the Facebook “Likes” and Demographic models see Michigan as very competitive. During the primaries, Michigan was where Bernie Sanders pulled off one of the greatest upsets in the history of modern polling, beating Clinton despite being down by more than 20 points in the polls. If most of those Sanders supporters turn out and vote for Clinton, Clinton shouldn’t have a problem winning here. If a large contingent of Sanders supporters stay home and others vote for Trump, Clinton may be in trouble in Michigan.
Tossup (45% to 55% chance of Clinton victory), 38 Electoral Votes: Texas
Perhaps the biggest surprise in our predictions, both models have Texas, typically a deep red state, as highly competitive in 2016. The reason is pretty obvious: 28.1% of eligible voters in Texas are Hispanic, the second highest figure in the country after New Mexico. In addition, the vast majority of these Hispanics are of Mexican descent, a group Trump literally insulted on the first day of his campaign and has continued to insult since. Meanwhile, 12.1% of eligible voters in Texas are African American and 3.6% of eligible voters are Asian, and after including other non-whites, non whites make up an astonishing 48.7% of Texas’ eligible voting population.
So why is Texas not a safe bet for Clinton? A couple reasons. First, Texas’ large Hispanic population rarely turns out to vote in the same numbers as whites and blacks, posting an especially anemic turnout rate in 2012. Second, as one could probably guess, the white voters in Texas are especially conservative. In 2012, despite Texas white voters being the 11th most educated in the country in terms of total percentage with a bachelor’s degree or higher, an estimated 76% of white voters in Texas voted for Romney, compared to only 22% who voted for Obama.
Fortunately for Clinton, Trump is the perfect candidate to quell both of these concerns in Texas. In terms of Hispanic turnout, the prospect of a Trump Presidency is reportedly leading to massive increases in Hispanic turnout based on early voting data in heavily Hispanic states. In addition, Trump has done much weaker than previous Republicans with college-educated whites, which could prove important for Clinton in a state where we estimate that more college educated whites will vote than whites without a college degree. To make matters worse for Trump, many of these college educated whites were never on board with Trump in the first place, as Trump received an anemic 26.7% of the vote in Texas during the Republican primaries.
If Hispanics, African Americans, Asians, and college educated whites turn out en masse against Trump, Clinton could be the first Democrat to win Texas since Jimmy Carter in 1976.
Lean Trump (30% to 45% chance of Clinton victory), 40 Electoral Votes: Arizona, Florida
Arizona and Florida are both critical states for Clinton if she wants to reach the 270 electoral votes required to win the Presidency. If she fails to win Texas, it will likely be one of these two states that puts Clinton over the top.
In Arizona and Florida, the Facebook “Likes” and Demographic models disagree on the winner. In both states, the Facebook “Likes” model (which is weighted more heavily) predicts Trump victories, while the Demographic model forecasts two narrow Clinton wins thanks to a large non-white population in both states. Similar to Texas, the big unknown variable in Arizona and Florida is Hispanic turnout. If Hispanics turnout at similar rates to 2012, Trump should win both states pretty easily. If Hispanic turnout surges because of Trump’s anti-Hispanic rhetoric, don’t be surprised if Clinton pulls off the upset in one or both of these heavily Hispanic states.
Another thing to watch is how Cubans vote in Florida. Unlike Arizona and Texas, who have primarily Mexican Hispanic populations, much of Florida’s Hispanic population is of Cuban descent, a typically Republican leaning group. Will Cubans in Florida back Trump in spite of his rhetoric towards Latinos, or will they instead back Clinton, who they dislike for other reasons?
In addition, I’ll be closely watching working class white counties in Florida like St. Lucie County, Hillsborough County, and Monroe County as bellwethers for the white working class vote. Obama won these counties narrowly in 2012, but my Facebook “Likes” model has Trump winning these counties over Clinton by taking white working class Democratic voters from Clinton. If these counties come in strong for Trump, expect Trump to win Florida.
Likely Trump (15% to 30% chance of Clinton victory), 59 Electoral Votes: North Carolina, Ohio, Pennsylvania, Utah
Both the Demographic and Facebook “Likes” models have Trump winning North Carolina, Ohio, and Utah, so don’t expect those states to be very competitive. In fact, out of these three states, the most interesting state will probably be Utah, as Evan McMullin tries to become the first third party candidate to win electoral votes since George Wallace won five Southern states in 1968 as the candidate of the American Independent Party.
Pennsylvania is the most interesting state out of the four “Likely Trump” states. The polls give Clinton a narrow lead there, and the Demographic Model sees Pennsylvania playing out similarly. The Facebook “Likes” model, on the other hand, is extremely bullish on Trump’s chances in Pennsylvania, and for reasons I laid out two days ago, I’m inclined to agree with the Facebook “Likes” model.
If Clinton manages to pull off the upset in Pennsylvania, it will be because she dominates the Philadelphia and Pittsburgh metropolitan areas and wins the typically Democrat Scranton/Wilkes Barre region in northeastern Pennsylvania. How white working class counties like Erie County, Lackawanna County, and Bucks County vote should prove to be a bellwether for the state as a whole. I expect all three counties to vote for Trump over Clinton.
Safe Trump (0% to 15% chance of Clinton victory), 136 Electoral Votes: Alabama, Alaska, Arkansas, Georgia, Idaho, Indiana, Kansas, Kentucky, Louisiana, Mississippi, Missouri, Montana, Nebraska, North Dakota, Oklahoma, South Carolina, South Dakota, Tennessee, West Virginia, Wyoming, Maine’s 2nd Congressional District, Nebraska’s 1st Congressional District, Nebraska’s 3rd Congressional District
Not too many surprises here. Still, there are a few states and congressional districts to keep an eye on. Most notably, Maine’s 2nd Congressional District has been categorized as a toss-up by many pundits and election prognosticators, and the Demographic model agrees with that categorization. If Bernie voters turn out en masse for Clinton in Maine’s 2nd Congressional District, it’s probably the most likely out of the “Safe Trump” states and districts to vote for Clinton. Still, our Facebook “Likes” model is confident enough in Trump’s chances there to push Maine’s 2nd Congressional District all the way to “Safe Trump”.
Three other states/districts to keep an eye on are Nebraska, Nebraska’s 1st Congressional District, and Kansas. These states/districts all vehemently opposed Trump during the primaries – Kansas elected Cruz with 60% of the vote – and are mostly filled with “traditional conservatives” who might not align themselves with Trump’s brand of Republicanism. Traditional conservatives staying home or voting third party probably won’t be enough to push any of these states/districts into Clinton’s column, but the Facebook “Likes” model is bullish on Clinton’s chances in these three states/districts.
Clinton Is The Favorite, But Trump Has An Easy Path To Victory
The main reason Clinton is the favorite to win the election is simply because she has more viable paths to victory. If she wins Texas, which she wins in 47.3% of simulations, she is almost guaranteed to win. Even in simulations where she doesn’t win Texas, she can cobble together different paths to victory by simply winning one of Florida, Arizona, or Pennsylvania while holding on to all of the states she’s favored to win.
Still, for a candidate losing by a similar margin in the popular vote to Clinton as Mitt Romney lost to Obama four years ago, Trump is in a considerably stronger position to win the Presidency than Romney. This is because Trump has an extraordinarily simple path to reaching 270 electoral votes. Indeed, all Trump has to do is win every state where he is favored to win and also win toss-up Texas in order to reach 273 electoral votes, a remarkably simple path to the Presidency for a candidate only pulling in roughly 40% of the popular vote.
Nonetheless, Clinton is probably going to win, even if her path to victory isn’t as easy as she would like. Just remember that a 79% chance of victory is far from a sure thing.
akelly96