The new Emerson College poll for New York provides some good news for Donald Trump — for now, anyway. With more than five weeks to go before the Empire State’s primary, Trump trounces Ted Cruz for the state’s 95 delegates by more than 50 points, 64/12, with John Kasich only getting 1% despite his big win in Ohio. That puts Trump well over the 50% threshold for winner-take-all allocation, while leaving Cruz and Kasich well below the 20% threshold for any allocation in the event of a plurality. Trump, the home-state candidate, also gets the most favorable marks in the GOP field, +48 against Cruz’ +8 and Kasich’s +20. The small sample of 298 likely Republican primary voters leaves the margin of error at 5.6%, but with gaps like these, the MoE is an afterthought.

However, when it comes to the general election, the news looks worse — much worse, in fact, for the argument that Trump will expand the GOP map:

Looking ahead to the general election, the two Democrats do equally well in a head-to-head matchup with Trump. Clinton (55% to 36%) and Sanders (53% to 36%) lead the reality show star by 19 and 17 points, respectively, while Cruz loses to Clinton by 31 points (61% to 30%).

The MoE on the full survey is a much more reasonable 3.5% on a sample of 768 likely general-election voters, so these numbers are a little more solid. The only bright spot for Trump in the general-election numbers is that Cruz does worse at 61/30, but that’s a cold comfort in a general election, where Electoral College votes are winner-take-all.

Furthermore, these numbers are almost exactly the same as two successive Siena polls got in the past two months, surveys with even larger sample sizes and a traditional live-call model. In February, the race was 57/32 Clinton, and two weeks ago Siena had it at 57/34. That’s close to the margin by which Barack Obama beat Mitt Romney in 2012 (62/36). All three polls show no upside for Republicans by having Trump on the ticket, not even with New York being his home state and the base of his business operations. Clinton may be underperforming Obama, but not by much.

Of course, the general election is still seven-plus months away, but that would only matter if New Yorkers hadn’t gotten to know Donald Trump well (or Hillary Clinton, for that matter) before now. Trump’s ubiquitous in New York, though, and voters there know him better than voters in most other states, with the possible exception of Florida — where Trump actually looks more competitive, but where Republicans have been much more competitive in presidential elections anyway. Romney only lost Florida by a percentage point, getting to 49%, while the RCP average of head-to-head polling there has Trump down 2.2 points … and Trump going no higher than 47% in any of them.

Given this lack of change over Romney’s 2012 performance in New York, where’s the evidence that Trump expands the map? If at this point Trump can’t get to 40% in his home state in head-to-head polling, what makes anyone think that Trump could beat Hillary in California or Massachusetts, or even in Pennsylvania or Michigan? The theory that Trump expands the Republican general-election map and provides a new path to 270 has no evidence supporting it. Right now, it looks like pure fantasy.

Update: I’ve received quite a bit of feedback about this post on Twitter. The reason why this matters is that Trump and his campaign argue that they expand the map for the GOP, and so do not need to worry about traditional swing-state strategies or demographic expansion. The core of that argument was that Trump could carry New York for the GOP for the first time since Ronald Reagan did it in 1984. These three polls show no evidence that this is the case — which means that the GOP would need to win back the swing states it has lost in the last two presidential elections by expanding their reach into key demographics. That’s the case made in my upcoming book Going Red. Trump has shown no particular indication of following that path either, instead offering the same kind of 30,000-foot national-messaging campaign that helped Obama win two elections.