President Donald Trump faces a large deficit in the polls right now. If the election were held today, he’d very likely lose. But if Trump were to pull off a surprising win, how would he do it? A look at the electoral map reveals just how narrow Trump’s possible paths are at this point. What Trump probably needs right now isn’t some magical map, but rather a fundamental shift across the board to move the environment in his direction. Build your own road to 270 electoral votes with CNN’s interactive map You can see this first by giving him the states he won four years ago by 1.2 points or greater. This includes a number of states where Trump is slightly trailing in the polling averages right now, like Iowa and Georgia, and also gives Trump states such as Arizona, Florida and North Carolina, where he won by less than 5 points four years ago. Trump is down by 3 to 4 points in the polls in all of those states at this point. Were former vice president Joe Biden to win any of these states, Trump’s chances begin to shrink very, very quickly. But even if Trump won all of these states, he’d still be at only 260 electoral votes. Trump needs to somehow find more votes. One potential path for Trump to win at least one of the Rust Belt (i.e. Great Lake) battleground states he won by less than a point in 2016. That is, win either Michigan (16 electoral votes), Pennsylvania (20 electoral votes) or Wisconsin (10 electoral votes). A victory in any of these states gets him to at least 270 electoral votes. Here’s the problem: he’s losing by a lot in all of these states right now. As I outlined on Sunday, Trump’s down by 7 to 8 points in all of them to Biden. An alternative path might be to try and win some of the electoral votes that went to Hillary Clinton in 2016. The two most likely suspects are Nevada and New Hampshire. Clinton won those by about 2 points or less last time around, and Trump has spent money on advertisements in both of those states. If Trump won all the states where he won four years ago by 1.2 points or greater, Nevada and New Hampshire, he’d reach 270 electoral votes. Again, the issue here is that the polls in neither Nevada nor New Hampshire are particularly close at this moment. Trump’s not doing any better in either state than he is in the Great Lake battleground states. Biden’s up by 7 points in Nevada and by double-digits in New Hampshire. Indeed, in none of the states that Clinton won in 2016 is Biden’s lead less than 7 points. Trump’s electoral map difficulties come down to the fact that it’s pretty much impossible for a candidate to win when they’re trailing by 10 points nationally. As I have previously noted, the swing state polls look a lot like you’d expect it to with the national polling where it is. For Trump to have a realistic shot of winning, his national deficit would have to be sliced by at least half in the final result. Most likely, Trump probably can afford to lose by 3 or 4 points nationally in order to replicate the popular vote/electoral college split that he did in 2016. In other words, he probably needs to do 6 to 7 points better nationally than he is currently doing. (Not surprisingly, if Trump does 6 to 7 points better in Michigan, Pennsylvania and Wisconsin than he currently is, then those states are about even.) Now, Trump doesn’t need to gain back all 6 to 7 points before the election. If he can merely close Biden’s advantage by a few points, Trump could hope to get a polling miss that benefits him. Remember, Trump won in 2016, despite being down by about 4 to 5 points in the final polls of Michigan, Pennsylvania and Wisconsin. Trump, though, still needs to be doing better than he is right now for a polling error to be anything more than a true long shot to give him a second term. Otherwise, Trump’s paths will be possible, but they won’t really be realistic.