My Favorite Maps from Redraw the States

Last week, I debuted a tool called Redraw the States, which allowed users to rearrange U.S. counties into different states and see what would happen to the Electoral College vote. In the post debuting the tool, I offered a few examples of maps which would have flipped the result of the election but, either, made imperceptible changes to the map, or involved moving only three counties from one state to a neighboring state.

Three counties move: Lake IL -> WI, Lucas (Toledo) OH -> MI, and Camden NJ -> PA. Clinton wins 279 to 259. Blue means Clinton wins the state. Red means Trump. Deeper colors mean a bigger win.

In the week since, I’ve gotten a lot of feedback, including a nice article in the Washington Post’s Wonkblog and another nice one in Daily Kos Elections. I’m grateful to those outlets for writing these up.

But I’m also grateful to all the people who have written to me via Twitter and via email, sharing some of the maps they’ve made. In the spirit of the original article, demonstrating the capricious nature of the Electoral College and the states’ methods of allocating electors, I’d like to share some of my favorites.

Extreme Maps

The first class of maps I’d like to share are the “extremes.” How much could Clinton win by? How much could Trump win by?

First, let’s start with “Clinton shuts out Trump,” that is, Clinton gets 538 votes and Trump gets 0. (In this imagining, we’ll assume Maine and Nebraska use the winner-take-all system for allocating electors, even though they do not.) My favorite of these comes from Brad Slavens, who sends the following map:

Funny looking states, but Clinton wins 538 to 0.

Brad set himself some ground rules: States should be contiguous and retain their capital cities. As much as possible, states should also retain their flagship school (usually the University of State) and, secondarily, their land grant school (usually “Name of State” State University), and their largest city. He did pretty well! Some highlights: Only KS, MN, MO, MT, and SD are missing their largest cities (he had to move them to turn other states blue), only ID, MN, NV, and TN are missing their flagship school, and only six land grant or “other notable schools” aren’t kept in their original state.

I love the sheer whimsy of this map: California now stretches all the way to Tennessee, which would represent a ludicrous political jurisdiction, but makes for a nice academic discussion.

But what I like most about this map is the recognition that what we generally think of as “New York” or “Nebraska” often covers only a small area of the actual state. Or, to say another way if you are arguing for popularly electing the president: people vote, not trees.

Trump 409 to Clinton 109. Some states have simply been removed!

On the other end of the spectrum, it is impossible for Trump to win all the Electoral votes as he lost the popular vote, so how close to 0 can we get Hillary? The best example of this I’ve received comes from Patrick Derocher who tweeted me with the heading “Make Deseret Great Again,” which is wholly appropriate. He isolated the extreme blue counties of King (Seattle), Multnomah (Portland), Cook (Chicago), as well as Hawaii, Blue New England, and “The Bay to LA”.

The remaining states he manages to turn red, though it does require getting rid of a few. For instance, New Mexico is now simply part of Texas, and Utah absorbs Arizona, Nevada, Colorado, and much of California (“New Deseret”). Still this is a great illustration of the opposite point of Brad’s map: we can arrange the states so that Trump doesn’t just win, but really wins. To me, this is simply more evidence that the winner-take-all-based Electoral College is a terrible idea.

Historical Maps

In another vein, Nate Cohn of the New York Times’ Upshot had some fun rearranging the states as if some historical compromises had never happened.

ME rejoins MA, WV never left VA, 36°30' to the Pacific, a different outcome to the Toledo War, and Alabama gets the Florida counties it asked for in its first century of its existence. In this map, the final vote is 269–269.

There’s a lot of historical accident to the way the states are drawn. Maine seceded from Massachusetts as part of the Missouri Compromise of 1820, West Virginia seceded from Virginia to avoid seceding from the Union at the start of the Civil War, the Upper Peninsula of Michigan was given to Michigan to resolve the Toledo War. And it is only because California’s Gold Rush had already solidified much of its borders that it wasn’t subject to the Compromise of 1850, codifying once and for all that 36°30' would be the boundary between slave and free states.

This map is perhaps my favorite that I have seen, simply because it points out how arbitrary state borders truly are, subjects of compromises over bad surveying (the Toledo War example) and, most strikingly, slavery. There are many ways that American slavery still today affects the franchise of African-American voters, but that so many state boundaries were drawn to tamp down tensions leading up to the Civil War is one way often forgotten.

Nate’s original tweets can be found here, here, here, here, and here.

Small swings for Trump

Stafford NH -> ME, New Castle (Wilmington) DE -> MD, Fairfax VA -> MD, Clark (Las Vegas) NV -> AZ. Trump 329 to Clinton 309.

Many people pointed out that you could flip states from Clinton to Trump by moving a single county to a neighboring state as well. In this map, we can flip Nevada by jettisoning Las Vegas, and NOVA’s population, specifically Fairfax County, is so large and so blue compared to the rest of the state that putting it into Maryland makes Virginia go to Trump. Most crazy, though, is that Clinton’s lead in New Hampshire was so small that moving her 5,000 vote advantage in Stafford County into Maine is sufficient to give its four electors to Trump.

In my original post, I was concentrating on the capriciousness of the Electoral College by showing that the result could easily be flipped if only the small changes in the map were made, but this map shows that small changes can reinforce the result as well.


LA to Texas! LA borders the Pacific, TX borders the Gulf, those are connected by the Panama Canal and the Atlantic! Water Contiguity! (Also, Trump 269 to Clinton 269; or, counting ME-2, Trump 270 Clinton 268.)

There were many entirely absurd maps that people made as well. My favorite, which I first saw from Alon Levy, moves Los Angeles County to Texas. Justified by “water continguity” and the Panama Canal, this map shows that Clinton had such a commanding lead in LA (at the time I pulled the data 1.25m votes, though now 1.7m), that if it were part of any red state, including Texas, it would flip that state blue.

This map also led to my favorite conversation, with one user asking why Alon hates LA County so much, and his replying that, “On the contrary. I love it, & want it subject to Texas development rules to build more housing.”


As I said in my last post, Trump won fair and square according to the rules. But the winner-take-all nature of states’ elector assignments is enormously distortionary, leading to results that appear arbitrary and capricious. A few counties moving to neighboring states can swing the election to Clinton, or push up Trump’s total. And the historical compromises around state borders, almost all related to slavery, are still felt today.

This sort of caprice alone is enough evidence that the Electoral College should be abolished and that the President should either be chosen (parliament-style) by the House, or simply be popularly elected.

Farmer by upbringing. Mathematician by training. Data scientist by profession. Yeller at clouds.

Farmer by upbringing. Mathematician by training. Data scientist by profession. Yeller at clouds.