What Went Down In The New York Primaries

Something else worth watching: the exit poll that had Clinton up by only 4 percentage points statewide had her beating Sanders by 10 percentage points in New York City. So far, however, Clinton has about 85,000 votes from the five boroughs as compared with about 52,000 for Sanders, which equates to a 62-38 advantage for Clinton in the city overall.

We can start to pull out a couple of interesting tidbits from the exit polls in New York, starting with the Democrats and how they view the candidates and the race. While Sanders generally gets the media play of being the more inspirational candidate, Clinton is leading on that question in the New York exit poll, a reversal of fortune for her since Wisconsin, where Sanders won on the question of heartstrings. Voters in New York also see Clinton as inevitable — seven in 10 believe she’ll be the nominee, as opposed to three in 10 who think it will be Sanders. That said, voters still seem to have some discomfort with Clinton on a personal level — six in 10 say she’s honest, leaving over a third who might have some questions about her email servers or her Wall Street speeches. Sanders, meanwhile, is seen by eight in 10 voters as being honest and trustworthy.

Kumar, it’s not that they are being ignored. The problem is that even though Asian-Americans make up 13 percent of New York City’s population (and far less statewide), they make up even less of the voting population statewide. It’s just beyond the ability of the exit poll to get that much precision on a relatively small group.

Jody, I am conflicted. I know I just said that I think the Wisconsin system of same-day registration is great. (We also have open primaries, so no need to commit to a party in advance.) But I have reservations about the possibility of achieving real democracy within parties. Parties occupy a weird place in politics. I mostly agree with Jonathan Bernstein here that parties are distinct organizations that should be able to choose their own nominees – provided they are open to new members. The Supreme Court has upheld the rights of parties to control their nominations as part of their members’ Constitutional freedom of association. Looking at the post I made awhile ago about momentum and sequential voting — it’s not clear to me that without major changes, the nomination process can ever really be open and democratic. But while I believe in parties as entities, I also think that people should be able to decide fairly late in the game that they want to participate. Ideally, I’d love to see a system with more face-to-face contact — caucuses and state conventions – but these require a lot of time and energy, and our culture isn’t set up to encourage or reward civic engagement. I worry about shutting people out of primaries like this one, which has taken on a lot of significance, or contests where the primary stands in for the general. At the same time, a lot of advocates for a more democratic nomination system seem to be envisioning another big national election that looks a lot like the general — and I seriously doubt that’s a good idea.

https://twitter.com/PatrickRuffini/status/722593954231934976

We enjoyed the Wall Street Journal’s feature this week on Trump’s best and worst counties, so it’s worth watching whether he can beat his previous high of 69.7 percent of the vote in Buchanan County, Virginia. Richmond County (Staten Island) looks like it will probably be up to the job. Trump has 79.9 percent of the vote there so far with 53 precincts reporting.

As I’m sure Harry will remind us several times before the night is out, New York is really two states: New York City and its suburbs, and everywhere else. That’s true economically as in so many other ways. New York City itself is doing reasonably well economically. The unemployment rate, at 5.7 percent in March (not seasonally adjusted), is higher than for the U.S. as a whole, but it has been falling steadily. Job growth has been strong and wages have generally been rising (though more slowly in recent months). Wall Street, Sanders’s favorite punching bag, has struggled a bit over the past year, but so far that hasn’t proved much of a drag on the broader economy. New York’s northern suburbs – the counties of Westchester, Orange and Rockland – are doing even better. Their unemployment rate was just 4.2 percent in March, lower than for the U.S. as a whole, and they have experienced strong wage growth. (Nassau and Suffolk Counties, on Long Island, also have a low unemployment rate and solid job growth but haven’t seen the same recent wage gains.) But the story is different in upstate and western New York. Outside of a few particularly hard-up counties, those areas aren’t necessarily experiencing high unemployment, but job and wage growth have been painfully weak. The Buffalo area, for example, has added just 3,500 jobs in the past year, and employment in the Albany area has actually declined. Those numbers only begin to tell the story of the long-term struggles of upstate cities and towns, which are grappling with declining industrial bases, high rates of joblessness (not all of it captured in the official unemployment rate) and shrinking populations.

You want to talk about divides in the Democratic primary? Look at the age and racial breakdowns in the exit polls. Sanders is winning voters ages 18 to 24 by a margin of 85 percent to 15 percent, while Clinton is winning voters ages 65 and over by a margin of 70 percent to 30 percent. Among racial groups, Sanders carries whites by 9 percentage points, while Clinton takes Latinos by 18 percentage points and blacks by 43 percentage points. All-in-all it makes for a closer race on the Democratic side (in the exit polls at least) than pre-election polls suggested.

I’m not sure if it’s because the races have gone on for a while, or because of the populist tinge of the races, but one of the themes that seems to have come out of the last couple of weeks is discussion about which groups count as “the people.” Populist candidates — and others — tend to talk about the people as if they can speak with one voice, if only the nasty, selfish elites would get out of the way. But the people — all voters, or all citizens — comprise lots of groups who may not, as it turns out, agree with each other. Tucker Carlson tweeted earlier that “once again Sanders has swept… people under 30.” People under 30 don’t vote at the highest rates, but they are still voters! Sanders himself has run into some trouble by talking about how Clinton has mostly won primaries in the South. His point was that the South is conservative, but a number of people have pointed out that that’s not really true of Democratic primary voters there. There’s an interesting tension here between an election with a lot of rhetoric about “the people” and a renewed interest in democracy within parties, and the seemingly unavoidable tendency to divide the electorate up into groups and suggest that some voices have more, or different, significance than others.

Want proof that the Sanders team really wants a Cinderella win in New York? They outspent the Clinton team by $3 million in ad buys, cashing in $6.8 million total in the state. Sanders also made New York’s hottest clubs … its parks. He threw marquee rallies in areas where “the youth” congregate, Manhattan’s Washington Square Park and Brooklyn’s Prospect Park, and they had everything: about 28,000 people each, revolution, guys wearing graphic T’s with a frumpy 74-year-old’s head plastered all over them, Vampire Weekend, birds, Wall Street reform talk and in all likelihood, some psychotropics.

In the past, we’ve found that a mix of three-quarters exit polls and one-quarter our ”polls-only” model is an effective mix for predicting the eventual results. Given where the exit polls stand… That would project a finish of Trump 57, Kasich 25 and Cruz 17 on the GOP side, and Clinton 53, Sanders 47 for Democrats. The latter race is close enough that it isn’t safe to assume Clinton will win.

Trump, Clinton and Sanders all have strong New York connections, and their Twitter followings somewhat reflect that: A greater share of their followers are in New York than Cruz’s. Kasich gets more of his following in New York than Trump does, but Trump also has more than 25 times as many followers — and 40 percent of Trump’s followers are monitoring his latest insults and polling updates from outside the U.S., compared to just 12 percent of Kasich’s followers who are outside the U.S. No surprise: Sanders and Clinton have a higher share of their followers in the reliably blue state than do any of the three Republicans.

Who’s following the candidates on Twitter?

	SHARE OF FOLLOWERS WHO ARE
CANDIDATE	IN NY	MALE
Trump	5.0%	66%
Cruz	4.3	65
Kasich	6.2	64
Clinton	6.3	62
Sanders	7.7	59

These are estimates based on Twitter users’ profile text from TweepsMap, a Toronto Twitter analysis company. The company also estimates the gender of users — based on text in profiles like “husband,” “mother” and “daughter” — and finds that people who both use gendered language in their profiles and who follow one of the remaining candidates are heavily male. Trump’s following is the most male-dominated; Sanders’s is the least.

https://twitter.com/ryanstruyk/status/722591427692208128

The preliminary exit polls have Clinton’s lead at just 4 percentage points. If that becomes the final result, it would be closer than all the pre-election polls. It wouldn’t necessarily help Sanders in the delegate count, but it would allow him to continue to claim some type of “momentum.”

ABC News has projected that Trump has won the state of New York. We don’t know how many delegates he has won yet.

Earlier today, Nate wrote about Sanders performing better relative to expectations as the campaign has gone on — consistent with national polls showing him gaining on Clinton, the front-runner, in the Democratic race. One problem with the national polls, especially once voting starts, is that they include people who can still vote, and ones who can’t because the states they vote in have already held their primary or caucus. So we asked the online polling and media company Morning Consult to break down its latest national poll into respondents in states that have and have not voted.

Candidates’ support by registered voters in their party

CANDIDATE	STATE HAS VOTED	STATE HAS NOT VOTED
Clinton	46%	46%
Sanders	43	44
Trump	46	47
Cruz	26	25
Kasich	13	12

It turns out that, at least based on presidential preference, Democrats and Republicans in New York and the states that vote after today are almost identical to those in the states that have already voted, according to Morning Consult. So at least for now, national polls could be a pretty good indicator of how people who have yet to vote are likely to vote. Three caveats:

• The numbers for previous states don’t really line up with the actual vote shares for previous states. On the Republican side, that’s partly because many prior votes were cast when more candidates were in the race. On the Democratic side, it’s partly because many of Sanders’s strongest states were caucus states, where there aren’t many votes to go around. And on both sides, the discrepancy between votes and the poll is partly due to people changing their mind after their state’s primary day.

• National vote shares don’t really matter; what matters is delegates, and how the support expressed in polls translates into votes and later delegates has a lot to due with the arcana of states’ delegate rules. National polls can’t tell us anything about those.

• Sanders needs to do better than trail Clinton narrowly in the remaining states; he needs to win by a lot to have a shot at catching up to her big lead in delegates.

Hockey update: Although Rick Nash scored the first shorthanded goal of his playoff career to put the Rangers up 1-0, Sidney Bleeping Crosby deflected the puck in for a power play goal to tie the score the final minute of the second period. It’s tied 1-1 heading to the second intermission.

And here’s how New York’s Democrats voted in their last competitive primary, in 2008:

We’re about 10 minutes from the polls closing in New York, so for context, here’s how the state’s Republicans voted in 2012 and 2008:

Our default view on primary nights is that Republican results have higher stakes than Democratic ones. There are several reasons for that. First, Trump’s status as the Republican frontrunner is more tenuous than Clinton’s as the Democratic one. Second, Democratic delegate allocation is extremely proportional, while Republicans sometimes adopt winner-take-all or winner-take-more rules that can put more on the line. Third, a Trump nomination would pose more of an existential threat to the GOP than Sanders would to Democrats. Tonight is potentially an exception, however. We’re expecting Trump to do very well, and the difference between his worst-realistic-case and best-realistic-case isn’t all that wide. If he gets something like 60 percent of the vote statewide, he might win something like 90 of the 95 delegates at stake tonight. If if he disappoints and only gets 48 percent of the vote or so — anything lower than that would be pretty shocking — he’d still probably get 70 to 75 delegates, hardly a terrible evening. Every delegate counts, so that isn’t a meaningless difference. But even if Trump gets all 95 delegates, he’ll still have his work cut out for him in Indiana, California and other states. And even if he gets “only” 70 or 75, he’ll have an opportunity to make up ground later on. By contrast, a blowout win for Clinton tonight could make Sanders’s delegate math all but impossible, while a Sanders win would represent a huge upset and a real cause for concern for the Clinton campaign.

The CNN discussion of momentum and narrative is making me a little cranky. First of all, while there’s some evidence of how both of these things work, they aren’t the same thing. Narratives are stories about why something happened, and they do sometimes play a role after elections in constructing an account of why the outcome occurred and what it means for policy. In the context of a primary, this can mean stories about whether candidates over- or under-performed expectations, which voters they’ve won over, and which they have yet to court. Momentum can start with the idea of a candidate outperforming expectations, but political scientists define the concept as something related to the sequential nature of voting. Namely, voters are learning something from their counterparts in earlier states. As Barbara Norrander describes in her classic work, Super Tuesday, some scholars define this in terms of learning which candidates are viable (can win the nomination) and electable (can win the general election). This is the definition used in Why Iowa? Others don’t give voters that much credit, and suggest that momentum is just about gaining familiarity. The CNN crew has been saying that there’s been no momentum in this race. But we don’t have enough evidence to say that for sure. Because Sanders is a somewhat unexpected and unconventional candidate, signals about viability and electability have the potential to be very important. At the same time, that kind of signal won’t cancel out the preferences of people who simply prefer Clinton. And while it seems like Sanders has been around forever at this point, it’s not beyond imagination that some voters only became aware of him and his message when he did well in a few key early races.

Tonight marks a whiplash-inducing shift in the GOP primary calendar. And for Trump, transitioning from a mishmash of red states and caucuses to a final phase dominated by blue state primaries is not unlike Rafael Nadal entering clay court season in tennis. Up to this point, just 28 percent of GOP delegates at stake have been decided by primaries in blue states (484 of 1,703). But starting with New York’s primary tonight and ending with California on June 7, a massive 76 percent of delegates will be awarded in blue state primaries (586 of the remaining 769). That’s Trump’s wheelhouse. Trump has won eight of 10 blue state primaries so far, losing only Ohio (Kasich’s home state) and Wisconsin, which is more of a purple state anyway. Not counting Ohio, Trump has won 58 percent of the delegates in these contests, 13 percentage points higher than his national share. States like Maryland, New Jersey and California are why Trump still has a decent shot at 1,237. Cruz has done best with deeply conservative voters in red states while Kasich has often divided anti-Trump voters in blue states, allowing Trump to carry the most heavily Democratic states and districts, which often provide the most delegate bang for the buck. So while a winnowing field threatens to derail Trump, his best states could be coming to his rescue just in time.

We should expect a ridiculously slow count of results tonight. I was just looking back at FiveThirtyEight’s liveblog from the 2010 statewide primary. Carl Paladino wasn’t declared the winner of the Republican gubernatorial primary until after 11 p.m., despite winning by over 20 percentage points. The other important thing to note is that the votes from outside the New York City metropolitan area came in first. If that holds tonight, that probably means that the initial statewide percentages for both Clinton and Trump should rise in later tallies.

New Yorkers’ votes count today in two competitive primaries. In November, though, their votes probably will matter a lot less than those of voters in other states. Just how much less depends on who wins the nominations and what the polls show about the likelihood of a close race, but based on the last two decades of presidential elections, the Democrat should coast to a win of New York’s 29 electoral votes. In 2008, Columbia University statistician Andrew Gelman and his colleagues estimated the probability that one vote would swing a presidential race, based on the state in which it was cast. No single vote had much of a chance, but some had a much bigger chance than others. Each New York voter had one of the lowest probabilities of deciding an election: one in 1.9 billion. New Mexicans had a much better chance, at one in 6 million. And residents of the nation’s capital had just a one in 490 billion probability of deciding who would reside in its White House. I’ll make the same disclaimer my boss did in 2008: This doesn’t mean New Yorkers shouldn’t vote in November. (Not to mention all the other races besides the presidential one.)

Julia just praised Wisconsin’s same-day registration for being more voter friendly. New York’s primary has a lot of people we know thinking about open/closed primaries, registration deadlines, and so forth. I’m curious what the others on the blog (and the commenters!) think. If it were up to you, how would you design the system? Who could vote, when would they have to register, and how else would you change the way we vote in primaries? I’ll be sure to forward your ideas to the New York City Board of Elections.

I’m just going to say it: voter registration is mostly a crap idea. It’s true that registration was originally a reform aimed at a party system that was dominated by a few corrupt party bosses. But it’s also true that this was an anti-immigrant move. Furthermore, there’s some evidence that institutional barriers to voting affect turnout, especially for people at lower income and education levels. Voter registration laws build on the worst aspects of 19th-century politics, maintaining rules that are highly varied across different states, while simultaneously making it more difficult to vote. We certainly have our problems here in Wisconsin, but same-day registration appears to allow for high turnout in general elections (lots of challenges to causal inference there, so I say this with some caution). In this year’s primary, same-day registration created some of its own chaos, but people who were energized by the race late in the campaign were generally able to participate.

As my colleague Leah Libresco explained earlier, New York’s voter registration laws are some of the strictest in the country. Only voters registered with the Democratic and Republican parties can vote in their respective primaries. The registration deadline for new voters was March 25. And the deadline for switching parties was Oct. 9, more than six months ago. Perhaps as a result, there’s been very little growth in voter registration year over year:

Little growth in voter registration in New York

	ACTIVE REGISTERED VOTERS
PARTY	APR. 1, 2015	APR. 1, 2016	CHANGE YEAR-OVER-YEAR
Democrats	5,262,004	5,268,431	+6,427
Republicans	2,563,924	2,554,996	-8,928
Conservative	148,484	149,564	+1,080
Green	22,928	23,394	+466
Working Families	43,170	42,947	-223
Independence Party	433,659	430,172	-3,487
Women’s Equality Party	12	1,261	+1,249
Reform Party	8	373	+365
other parties	4,925	5,186	+261
no party	2,229,938	2,252,424	+22,486
Total	10,709,052	10,728,748	+19,696

The number of active, registered Democrats has grown by only about 6,400 from a year ago, paltry in a state with 10.7 million registered voters. Republican registration has actually shrunk. And the number of voters who are registered but aren’t affiliated with a party has grown, even though those voters don’t get a vote on primary day. To be clear, this doesn’t mean there are only 6,400 new Democrats. That’s a net figure; it includes some unknown number of newly-registered Democrats, minus some number of Democrats who switched parties or who were dropped off of active voter rolls. Still, a lot of voters went to the polls today without realizing that they’d failed to switch their voter registration on time.

Problems at the polling places aren’t the only issues with voting administration in New York City this year. In fact, one of the the big stories over the past couple of days has been the thousands of Democratic voters who fell off the registration rolls since November 2015. WNYC, the city’s public radio station, broke the story yesterday that over 60,000 active registered Democrats had dropped from the rolls in Brooklyn; by today, that number had been revised up to 126,000 according to data from the city’s Board of Elections. Lots of people are crying foul. As Ben Casselman just wrote, the city comptroller, Scott Stringer, said that he would be conducting an audit of the Board of Elections, and one of its goals is to see if there had been any improprieties involved in the removal of the voters. Mayor Bill de Blasio released a statement supporting Stringer’s audit and enumerating some of the claims of irregularities coming in from around the borough: “It has been reported to us from voters and voting rights monitors that the voting lists in Brooklyn contain numerous errors, including the purging of entire buildings and blocks of voters from the voting lists,” the mayor’s statement said. “I am calling on the Board of Election to reverse that purge and update the lists again using Central, not Brooklyn borough, Board of Election staff.” According to WNYC, the 126,000 number includes “12,000 people who moved out of the borough, 44,000 people who were moved from active to inactive voter status and 70,000 voters removed from the inactive voter list.” The executive director of the elections board, Michael J. Ryan, spoke to The Times and said that the number did not surprise him, given that a report had surfaced a while back showing voters who did not belong on the list. “Now we take people off the list who don’t belong there, and there’s some blowback, criticism that we’ve done it inaccurately,” Ryan said.

New York City polls opened at 6 a.m. on Tuesday – or at least they were meant to. Social media quickly filled up with complaints of closed polling places, missing registrations and other issues. This afternoon New York City Comptroller Scott Stringer said that amid “widespread reports” of voting irregularities, his office would “undertake an audit of the operations and management of the Board of Elections.” I saw some of those issues first-hand. I arrived at my polling place in Brooklyn’s Fort Greene neighborhood just before 8 a.m. to find mass confusion. The doors were open but workers were still setting up and no voting was taking place. Several dozen people waited in line, many of them on cell phones calling the New York City Board of Elections — which has a long history of incompetence in administering and counting the votes — or the media. “I came here at 6, came back at 7, they weren’t open and we’re still waiting,” said Mary Simpson-Driver, 66, who said she had lived in the neighborhood for some 40 years. This was her first time voting at this location, Simpson-Driver said, brandishing a change-of-site notification she had received in the mail; her old polling place had been in a nearby school, where voting had always gone smoothly. “They’re trying to steal our vote,” added her husband, Eddie Driver, 72. Poll workers explained that there was no coordinator on site, and that voting couldn’t begin without one. (They declined to comment further.) Voting finally began around 8:15 a.m. – more than two hours late — and by about 8:45 the line had mostly cleared out. Most of the people who were there when I arrived stayed around to vote, but I don’t know how many people gave up earlier in the morning. It’s hard to know how, if at all, the irregularities could affect tonight’s results. Most of the anger about the voting issues, at least on Twitter, seems to be coming from Sanders supporters, many of whom seem to suspect dirty tricks on behalf of the Clinton campaign or its Democratic Party backers. But most of the people waiting in line at my polling place were older African-American voters, a demographic group that has strongly favored Clinton. (In 2008, my district voted roughly 70-30 for Obama over Clinton, although the neighborhood has become whiter since then.) But whether or not the problems favor one candidate over another, they are a threat to the democratic process. As she left the polling place on Tuesday after finally casting her vote, Regina Robinson complained that in the general election in November, lines will be longer and the weather will be colder. She worried that next time people will leave without voting or won’t show up at all. She urged everyone she saw to call the Board of Elections to complain. “You’re not going to cheat us out of our vote,” Robinson said. “This place always votes.”

We originally published our Facebook Primary project on Feb. 11, a week before the South Carolina primary. A lot has happened in two months! Donald Trump took control of the Republican race. Bernie Sanders — while still a long-shot for the nomination — has proven himself a formidable opponent to Hillary Clinton. Ben Carson and Marco Rubio have dropped out. We now have Facebook candidate data through April 17, our last update since February 28. How has the Facebook primary changed? Facebook “likes” don’t shift all that quickly, it turns out. For example, the three Republican dropouts still make up 30.5 percent of likes in our dataset, only down from 34.3 percent on February 11. Nevertheless, it’s clear that some of the remaining candidates have had a better couple of months than others:

Bernie has gained the most likes nationwide

	SHARE OF CANDIDATE LIKES
CANDIDATE	FEB 11	APRIL 17	CHANGE
Bernie Sanders	22.7%	24.6%	+1.9
Donald Trump	22.5	23.8	+1.3
Hillary Clinton	7.8	8.7	+0.9
John Kasich	0.9	1.4	+0.5
Ted Cruz	11.7	11.1	-0.7
Carson/Rubio/Bush	34.3	30.5	-3.9

Bernie Sanders added the most to his national share of likes, and Cruz’s share of likes actually declined. But national likes, like national polls, aren’t quite right for this stage of the race. New York votes today in a critical primary matchup — Sanders has made gains in the state’s likes since February 11, but more notably Trump has been flat:

Trump’s likes have been stagnant in New York State

	SHARE OF NEW YORK LIKES
CANDIDATE	FEB 11	APRIL 17	CHANGE
Bernie Sanders	32.0%	33.7%	+1.6
Hillary Clinton	14.1	15.1	+0.9
John Kasich	0.4	1.0	+0.6
Donald Trump	26.4	26.8	+0.4
Ted Cruz	6.9	6.1	-0.7
Carson/Rubio/Bush	20.2	17.4	-2.8

Our primary model heavily favors Trump in today’s primary, regardless. But if it turns out that he has a vote ceiling, then maybe he has a Facebook ceiling as well. See all the updated numbers here.

I live within a few blocks of Madison Square Garden — close enough that I can see it from my apartment — and few things annoy me more than being stuck at home when I could be at a New York Rangers playoff game. Unfortunately, the state of New York decided to schedule presidential primaries for April 19, failing to anticipate that they’d conflict with Game 3 of the Rangers’ opening-round series against the Pittsburgh Penguins. The game is tied 0-0 after a scrappy first period. Chris Kreider had appeared to score a goal for the Rangers, but the goal was disallowed by superdelegates after a replay revealed the Rangers were offside on the play.

Do you have a handle on what in the world is happening with this whole delegate mess this year? No? You’re in luck. I’m going to be writing a weekly Monday column that updates you with all you need to know about the shifting landscape of the Great Delegate Hunt of 2016 — we’re calling it Conventional Wisdom. (Sign up for it in newsletter form.) If you don’t get the pun, think about it for 5-10 seconds longer. (Hint: there’s a convention in Cleveland at the end of this 1,237 mess.) Only six more days until the next dispatch!

The other day, a reporter emailed me to ask: “Can Clinton deliver a knockout punch to Sanders in New York? What’s it take?” To me, the question sounded not too far from: “If the Rockets cut the Warriors’ lead from 38 points to 28 points with four minutes left in the 4th quarter, how many more threes would Golden State need to hit before one qualifies as ‘the dagger’?” The answer, of course, is that there is no such thing as a “knockout blow” in the Democratic race, because the psychology of Democratic voters and most media outlets is quite detached from the reality of Clinton’s advantage in the delegate math. The Sanders campaign will likely press on until June, gradually transitioning from a competitive effort to a symbolic and cathartic one without a clear dividing line. It’s newsworthy that Sanders appears to be gaining ground with voters even as he becomes less mathematically viable each primary. But there’s a chicken-and-egg question here: Do media outlets keep hyping up every twist and turn in the race because Democratic voters want it to end, or is Sanders benefiting from self-serving media coverage that misleadingly frames each contest as “critical?” Either way, this decoupling and dissonance is to be expected. It’s not much different from what we witnessed in 2008, when Clinton’s campaign persisted for months after Obama built an insurmountable delegate lead. Clinton’s campaign and each relatively inconsequential primary continued to receive wall-to-wall coverage, and her lack of viability didn’t deplete her vote shares. Sure, New York is a compelling story insofar as Clinton and Sanders can tout political and personal roots in the state. But tomorrow, the show moves on. Just don’t spoil the ending for the next set of voters.

A lot of New Yorkers couldn’t vote today. New York requires voters to register a new party affiliation at least 193 days before the election in order to participate in the primaries. That’s by far the earliest cutoff date in the country. But New York isn’t the only state to set an earlier deadline to change parties than to register for the first time.

A few weeks ago, we published a sort of best-case scenario for Sanders in which he wound up with exactly 2,026 pledged delegates, the number he’d need to clinch an elected delegate majority over Clinton. (Leave aside the thorny issue of superdelegates for now.) The path would require almost everything to go right for Sanders — including narrow wins in states such as Pennsylvania and New Jersey, and double-digit wins in California, Indiana and other states. Sanders has had a good couple of weeks, however. He fell only two delegates shy of our path-to-2,026 projection in Wisconsin. He also fell four delegates shy in Wyoming, where his results were disappointing. However, Sanders has gained a few extra delegates at state conventions and from previous states revising their delegate counts as their results became official. Because of these changes, Sanders has kept exactly on pace with the path to 2,026 so far. Tonight’s task is much harder, however. Our path had Sanders winning New York by a couple of percentage points and netting 128 out of 247 delegates there. Here’s what the rest of his path would look like on the unlikely-but-not-quite impossible chance that he does so:

Sanders’s path to 2,026 after a narrow New York win

STATE OR TERRITORY	NO. ELECTED DELEGATES	SANDERS’S PATH-TO-2026 PROJECTION	POPULAR VOTE MARGIN NEEDED TO REACH TARGET
California	475	274	Sanders	+15
Pennsylvania	189	101	Sanders	+7
New Jersey	126	67	Sanders	+6
Maryland	95	43	Clinton	+9
Indiana	83	48	Sanders	+16
Oregon	61	45	Sanders	+48
Puerto Rico	60	33	Sanders	+10
Connecticut	55	31	Sanders	+13
Kentucky	55	33	Sanders	+20
New Mexico	34	18	Sanders	+6
W. Virginia	29	19	Sanders	+31
Rhode Island	24	16	Sanders	+33
Delaware	21	10	Clinton	+5
Montana	21	16	Sanders	+52
S. Dakota	20	14	Sanders	+40
D.C.	20	9	Clinton	+10
N. Dakota	18	14	Sanders	+56
Guam	7	4	Sanders	+14
V. Islands	7	4	Sanders	+14

By contrast, suppose that Sanders has a bad night in New York, losing to Hillary Clinton by roughly 20 percentage points and claiming only 99 pledged delegates. By falling 29 delegates short of his target in New York, Sanders would need to make them up somewhere else. I designed an algorithm to redistribute those delegates to future states and here’s what it came up with:

Sanders’s path to 2,026 after a big New York loss

STATE OR TERRITORY	NO. ELECTED DELEGATES	SANDERS’S PATH-TO-2026 PROJECTION	POPULAR VOTE MARGIN NEEDED TO REACH TARGET
California	475	284	Sanders	+20
Pennsylvania	189	105	Sanders	+11
New Jersey	126	69	Sanders	+10
Maryland	95	45	Clinton	+5
Indiana	83	50	Sanders	+20
Oregon	61	47	Sanders	+54
Puerto Rico	60	34	Sanders	+13
Connecticut	55	32	Sanders	+16
Kentucky	55	34	Sanders	+24
New Mexico	34	19	Sanders	+12
W. Virginia	29	20	Sanders	+38
Rhode Island	24	17	Sanders	+42
Delaware	21	10	Clinton	+5
Montana	21	17	Sanders	+62
S. Dakota	20	14	Sanders	+40
D.C.	20	9	Clinton	+10
N. Dakota	18	14	Sanders	+56
Guam	7	4	Sanders	+14
V. Islands	7	4	Sanders	+14

Sanders would have to increase his margins by 4 or 5 percentage points across the board — winning California by 20 percentage points rather than 15, for instance, Pennsylvania by 11 points instead of 7, and Kentucky by 24 points instead of 20. Most likely, Sanders won’t win New York tonight, but won’t lose by 20 points either. Still, every delegate Sanders falls short of our original target of 128 will make the rest of his path harder, even if he only narrowly loses New York to Clinton.

Some people like to complain that the media is too New York-centric. If you feel that way, this live blog may not be the one for you. Today, New Yorkers from Buffalo to Montauk cast ballots in the Democratic and Republican primaries. Polls will close at 9 p.m. EDT, though we’ll have some preliminary exit poll data before that. So what are we expecting? Donald Trump will probably run up the score in his home state, after a string of losses. He’s at 52 percent in our polling average, while Ted Cruz and John Kasich are stuck around the 20 percent mark. If he gets more than 50 percent, Trump will secure all 14 of the statewide delegates. The big question is which congressional districts will give him a majority and thus all three of their delegates. Our estimate is that he’ll miss 50 percent in about half of the 27 congressional districts and win somewhere between 83 and 85 delegates statewide. Such a win would be impressive, but it is already baked into our calculations, which point to Trump falling short of 1,237 pledged delegates at the end of the primary season. Hillary Clinton, too, is hoping that New York provides her first win in almost a month. Chances are it will, as she leads Bernie Sanders by 14 percentage points in our polling average. Clinton has long been expected to emerge victorious in New York, as it has a diverse and relatively wealthy electorate and holds a closed primary. Still, Sanders needs to win in New York in order to have a realistic shot at closing Clinton’s large 200+ lead in elected delegates. If Clinton does win by 14 percentage points, she would expand her elected delegate lead by 35 delegates and turn a long shot into a moon shot for Sanders. A win by Sanders, on the other hand, may signal that something about this race has fundamentally changed and that he has a legitimate shot to catch Clinton. So that’s what on the line tonight. After the polls close, it will take a while for most of the congressional district results to be counted, but we’ll be staying up late analyzing them.

Tweet us a question or comment

Leave a comment, and send us questions @FiveThirtyEight.