Republican Acela Corridor Tuesday results thread (all polls close at 8pm ET)

[Alcon]

CT exit poll: Trump 57%, Kasich 26%, Cruz 15%

MD exit poll: Trump 53%, Kasich 25%, Cruz 21%

PA exit poll: Trump 58%, Cruz 19%, Kasich 17%

No DE, RI exit poll

[Alcon]

Looks like Cruz is staying over 10% in RI-1. Just under in RI-2.

Say he gets 9.9% in RI-2 would that be brought up to 10% or would it say at 9%

It's a threshold, so I really doubt they round up.

[Alcon]

You must be logged in to read this quote.

Actually, Trump managed to lose ground. He needed all the delegates tonight. Did well everywhere but Rhode Island.

My original tabs put him about 50 short of the nomination.

I haven't looked at your methodology, but really?  It seems like it would make more sense to estimate where Trump is at in remaining states, and keep increasing his performance in each state linearly until he hits 1237.  That obviously wouldn't have involved winning every delegate in Rhode Island (unless I misremember their allocation methodology), because that would require such an absurdly strong performance that he would have clinched via big victories in Indiana and California first.  No?

[Alcon]

You must be logged in to read this quote.

He needed to be at 1k after April and the NE.  Getting zero delegates in Colorado and losing Wisconsin bigly hurt him badly for the nomination.

Coming into WI and CO put him right at 1235. Right now I have him at 1185, falling short of the nomination.

I had him winning every delegate tonight and 75 in New York. So my prediction ran about spot on.

You just told me nothing about your methodology and didn't answer the concern I raised

[Alcon]

You must be logged in to read this quote.

I ran off the delegate totals and estimated that Trump would be regionally stronger in the NE, with additional Kasich support and Cruz support. You're right, he wasn't expected to do as well in RI, but that was the only path I saw to him getting to 1237.

I was not far off. I had him under in NY and so the numbers come out correct. He still needed to be at 1k to be on course for the nomination.

But, again, unless I'm missing something, that makes no sense.  You're saying he fell short of what he "needed," but you had him winning all delegates in Rhode Island.  A world in which Donald Trump wins Rhode Island by enough to win all of its delegates, is a world where Donald Trump obviously wins California and Indiana by enough that it doesn't matter.  You're incorrectly modeling what Trump "needs."

[Alcon]

edit based on your last post: dude that is not what "pledged delegates" are.  Delegates do not become "pledged" once they vote.  "Pledged delegates" are delegates who are elected or appointed with the specific directions to support a given candidate at the convention (at least on the first ballot).  Mr. Morden's point is that there are a bunch of unpledged delegates that will be voting on the first ballot, so Trump doesn't need to enter with 1237 pledged delegates to be on-track to win, unless you expect he'll get no votes from unpledged delegates.

You must be logged in to read this quote.

Again, no I'm not. Trump needs 1237. To get there, there are several paths, but all of them have him at 1k delegates at this point in the race.

This is why he threw his trumpertantrum over Colorado's delegates. He did the math and realized that even with a sweep tonight - which he was denied, he would fall short.

You argued that Trump "needed all the delegates tonight" to hit 1237.  You argued this based on your model.  I'm asking why your model assumes Trump's most likely (let alone necessary!) route to 1237 involves a sweep in Rhode Island.  A Trump delegate sweep in Rhode Island would have required an epic overperformance -- obviously a much more unlikely outcome than Trump overperforming enough in other upcoming states enough to hit 1237.  If Trump is keeping Cruz and Kasich below threshold in Rhode Island, he's probably "on track" to win the remaining states by way more than he needs to.  If that's the case, your model doesn't represent what Trump needs to be "on track," which makes it effectively useless except as an obscure hypothetical.

Am I missing something here?  Mr. Morden?  Anyone?

[Alcon]

The delegates have to come from somewhere. I didn't feel comfortable predicting a sweep of NY, and I felt more comfortable in predicting a sweep tonight. I come out just about right, and more right than 538.

I'm just going to keep repeating this until you understand it.  You're claiming that your model indicates what Trump "needs" in each state to be on track to win 1,237 pledged delegates -- or, in fact, to fall 50 short.  You were claiming that Trump "lost ground" tonight on what he needs to lose by 50 pledged delegates, because he failed to stay on-track by not sweeping Rhode Island.

In other words, your model assumed Trump needed to keep Cruz and Kasich under 10% in both Rhode Island congressional seats to stay "on-track."  That would have been a huge overperformance, obviously.  If Trump overperformed in every subsequent state to that extent, he'd easily win more than 1,237 pledged delegates.  In other words, a delegate sweep of Rhode Island isn't "on-track" to win 1,237 pledged delegates (let alone fall 50 short).  It's way overboard.

That's not a model of what Trump "needs."  It's not a model of what keeps him "on-track."  It's a theoretical scenario you've created where Trump falls short of 1,237 pledged delegates while winning a freakish super-victory in Rhode Island for no reason.  That seems pointless to me as a "model."  It's also objectively not a sound rebuttal of the post you originally replied to, which claimed that Trump was on-track.

So far it's called Wyoming (Cruz sweep), Colorado (Cruz sweep), Wisconsin within a point, and got the delegate totals correct for NY and the corridor tonight.

You're right, that perhaps Indiana will be a win for Trump, but I'm predicting a Cruz victory here.

What were your predictions tonight?

Things that have nothing to do with the validity of my logical criticism of your model:

1. Whether you predicted delegates accurately in past races.

2. Whether I predicted delegates accurately in past races.

3. Anything about either me or you.

Things that do have to do with the validity of my logical criticism of your model:

1. The logical validity of my logical criticism of your model.

[Alcon]

I'm sorry, the math was done long before Wisconsin, and before Wyoming and Colorado. Trump has been in trouble for a long time, since he lost Ohio to Kasich. Said it then, say it now.

We are just about done the primary season. Romney had it wrapped up by Wisconsin, and his run was very late. Trump has not managed to do that. By losing in Wisconsin, he assured that he would not mathematically be able to clinch pledged delegates before California. Cruz now has no incentive to drop before California.

None of that has anything to do with how little sense it made to model a sweep in Rhode Island.

You're assuming that each state is linked to each other. I am not. I am treating each state as it's own mini-election. Trump's strongest states are in this corridor, ergo I assign delegates to his strong states before assigning them to his weak states.

Even if Rhode Island is one of Trump's best states, that doesn't mean it's easier for him to sweep the Rhode Island delegates than pick up the equivalent number of delegates in, say, Indiana or California.

Here's an extreme version of that illogic.  If you assume that every delegate is easier to obtain in strong Trump states than in weaker Trump states, Trump will never be "on-track" if he's losing any delegates in any of his strongest states.  That's obviously not the case.

If your model doesn't mean to establish whether Trump is "on-track," and is simply a hypothetical scenario involving a ridiculously bad prediction of Rhode Island, OK, fine.  But if that's the case, it makes no sense as a rebuttal to the claim that a contested convention just became less likely tonight.

It's not nonsense. I'm sorry you can't understand that Trump has been behind since he lost Ohio. Those delegates need to be made up. It's like Alice and Wonderland. He has to run twice as hard to catch up.

My argument is that you are using the wrong way of establishing Trump's "baseline" of need.  It says absolutely nothing about whether I think Trump is above or below his baseline.

That it predicted the delegate counts correctly is a solid indicator that my model is working well.

I'm not sure you have a "model" so much as a set of predictions, but in any case...

It's quite possible that the rest of your "model" is logical.  Even if it's not logical (or a model at all), it's quite possible that the rest of your "model" involves reasonable predictions.  But your use of your model as a baseline for what's "on-track" was not logical, and your model's prediction for Trump's baseline in Rhode Island was not reasonable.

[Alcon]

Yes, it does because he lost delegates that a sweep would have gotten him. Every delegate counts. I said the same after New York and folks laughed at me then.

...Do you really not understand how this totally fails to address my criticism?  I'm happy to explain it in great, excruciating detail, but seriously dude, can you re-read my post and think it over for a few minutes first?  (My entire point is about prioritizing which delegates Trump "needs" to be "on-track" and which he doesn't.)

Also, as the gentleman above notes, Delaware is winner-take-all.  I assume you maybe meant Rhode Island, in which case "but in the end, my model predicted [Trump] would be 50 short" is a total non-sequitur.  I feel like you're talking more than you're thinking here.

[Alcon]

the idea that trump underperformed in Rhode Island because he didn't get the 67% he needed in both districts to get a sweep is silly; especially since making a landslide prediction like that based on absolutely no evidence for it (no polls had him that high, he were in the high 50s sure but he'd have needed literally every undecided voter to get to the 2/3rds mark and that almost happened, surprisingly).  Its almost as if you didn't actually look to see what the delegate rules where in every state before you began randomly giving numbers to people...

It's actually even sillier.  In RI you get delegates for hitting 10% both statewide and congressionally.  Cruz is currently winning 3 delegates because he didn't fall under 10% in RI-1 or statewide.  Ben's Rhode Island baseline (Trump's "on-track" number there) would have required Trump to keep both Cruz and Kasich under 10% statewide and in both congressional districts.  Obviously, a world where Trump manages that is very likely a world where he's on-track to win 1,237 pledged delegates in a walk.  Think of it this way: Trump would have had to take almost two-thirds of Kasich's Rhode Island vote share.  It's ridiculous that Ben would require that for Trump to be "on-track."

Ben, are you seriously postulating that Trump gets little to none of the "unpledged" PA delegates, even among those who've publicly endorsed Trump or at least to support their district's winner? if so, Alcon's rght that your model is whack.

To be totally clear, I do think that assumption is wrong, but I also have other problems with the model.  Basically, that he's determining what delegates Trump needs to be "on-track" to clinch based on something besides a probabilistic evaluation of the 1,237 easiest pledged delegates for Trump to get (or 1,237 minus 50 or whatever).  That means his "model" doesn't indicate whether Trump is "on-track" at all...I'm not sure what he thinks it does, honestly.

[Alcon]

It's actually even sillier.  In RI you get delegates for hitting 10% both statewide and congressionally.  Cruz is currently winning 3 delegates because he didn't fall under 10% in RI-1 or statewide.  Ben's Rhode Island baseline (Trump's "on-track" number there) would have required Trump to keep both Cruz and Kasich under 10% statewide and in both congressional districts.  Obviously, a world where Trump manages that is very likely a world where he's on-track to win 1,237 pledged delegates.

When I was watching MSNBC last night they mentioned something about getting all the delegates automatically if you got over two-thirds of the votes; perhaps they got that wrong, I dunno.  Either way its ridiculous, the "Trump has to limit the other candidates below 10% of the vote everywhere!!" thing is just more so.

It's confusing, but I think the deal is that if all three candidates broke 10% in a congressional district, all three of them get delegates.  However, if one candidate breaks 67%, they're guaranteed two of the three delegates -- but they only get all of them if no other candidate breaks 10%.

For instance, RI-1 looks like it was about 61% Trump, 26% Kasich, 11% Cruz.  That means all three candidates get one delegate, since all three broke 10% and no candidate hit 67%.

RI-2 looks like it was about 66% Trump, 23% Kasich, 9.8% Cruz.  Since Cruz is probably under 10%, he doesn't get any delegates.  That means Trump is likely to get 2 delegates and Kasich 1 (proportional).  If Cruz goes over 10%, the split there will be 1-1-1.  However, if Cruz goes over 10% but Trump also goes over 67%, Cruz does not receive a delegate, because hitting 67% guarantees Trump 2 of the 3 delegates.  If that happens, the delegate split will remain Trump 2, Kasich 1 (since Kasich beats out Cruz for the third delegate).

This all means that Trump would have had to keep Cruz and Kasich under 10% of the vote (statewide and in each CD) to be "on-track" in Ben's math.  That would require Trump to have won (statewide and in each C) with an absolute minimum of just over 80%.*  After all, keeping Cruz and Kasich under 10% requires Trump to be that high.  

Even if Rhode Island gave a delegate sweep at 67%, Ben's "on-track" number would be too high there, but since it doesn't, his number is completely ridiculous.

(this all assumes I'm not messing some detail up -- this is complex stuff)

_{* - I'm not sure whether this math includes the other ballot options -- Uncommitted, Marco Rubio, and Write-in -- but it doesn't matter much.  That just knocks Trump's minimum down to like 78%.}