PA-18 Special Election - Lamb by a nose

[Tintrlvr]

Saccone's got to be the worst candidate I've ever seen.

If anyone is looking for Martha Coakley analogies, look no further. We'll see what the result is, but Saccone is surely much more similar to Coakley's generally haplessness than Roy Moore was.

[Tintrlvr]

Any updates on the ground?

Some people are voting. Other people are not voting.

THE ANECDVOTING IS REAL

Fixed.

[Tintrlvr]

I don't think the race is as over as some on this board are predicting. NYT needle still shows it very very close.

we have to see what's in Westmoreland before we now

I think we will know before Westmoreland reports, but we need more results to have a good sense. A lot is coming in fast right now.

[Tintrlvr]

Lamb now has two districts in Greene!

What was the second one? I don't see it on NYT. They're projecting all remaining Greene precincts for Saccone so would be interesting.

[Tintrlvr]

One area that is being overlooked but will be significant is the suburbs in Washington County. A lot is reporting from Washington County thus far, but only rural areas and the smaller cities. There's almost nothing from Peters, in particular, which has a large portion of the population of Washington County and is a wealthy and quite Republican Pittsburgh suburb, very different from the rest of Washington County. We don't have a great sense yet of how much it will swing further to Lamb relative to 2016.

[Tintrlvr]

We're just going to have to assume what Westmoreland is reporting is a representative sample of the county as a whole. Which is hard to be sure about, but it is true that Westmoreland County is relatively uniform as far as these things go. Saccone's current 56-44 (R+12) lead in Westmoreland with 53% reporting is probably not enough to overcome Lamb's performance elsewhere, but it's close (his benchmark was R+13).

[Tintrlvr]

Re: county benchmarks generally...

Allegheny: benchmark D+14, currently D+16 (close to done reporting so this can't change much)
Westmoreland: benchmark R+13, currently R+12 (about halfway, could change somewhat)
Washington: benchmark R+6, currently R+0 (only one-third reporting, and some of the largest R areas aren't reporting yet, so take with some salt)
Greene: benchmark R+19, currently R+16 (basically done reporting, no real room to change)

So, at a guess, Lamb wins by around 2 points, depending a bit on what happens in Washington County more than Westmoreland, but hard to be too certain without knowing more about what's outstanding in Westmoreland.

[Tintrlvr]

The new dump of Westmoreland left it at R+12, so Saccone is still slightly behind benchmarks there as well. With 77% reporting, it's basically caught up to Allegheny. The only room for Saccone remaining is a major overperformance in the suburban parts of Washington County, which haven't reported yet, but that seems highly unlikely given the rest of the results.

I'm ready to call this for Lamb.

[Tintrlvr]

CNN just reported that Allegheny absentee count is estimated to be complete around midnight.

Implying that there are absentee votes not counted yet in at least Allegheny. Anyone know about other counties?

Good that they are counting tonight. Would be a disaster if the absentees weren't counted till tomorrow or even later and changed the result either way.

[Tintrlvr]

ALLEGHENY ALL IN.  LAMB BY 755.

They're saying 100% but unclear if this includes absentees. Greene is counting absentees tomorrow and says they are 100% reporting.

[Tintrlvr]

Lost in all of this: monmouth pooped the bed.

their high turnout model (which this race is) is off by labout 5 points.

Being off by five points in a congressional district is honestly a great result for a pollster. Congressional district polling is notoriously off-base.

[Tintrlvr]

755 votes. I don't think Rick can make up raw vote and absentee should boost Lamb around midnight  

Not so sure. Most of the remaining precincts are in Peters, where there are six precincts remaining. It's been strong for Saccone, with his average net votes per precinct around 120 or so. It will be very close before absentees.

The other two Washington county precincts are one in North Strabane, where Saccone should net around 50 votes or so, and one in Chartiers, which should be around a wash.

[Tintrlvr]

Wow, I was very wrong about Peters Township. Thought it would net Saccone much more. Crow is being eaten.

Western Peters seems to be less Republican than eastern Peters. Western Peters was what was late while eastern Peters was already in. Western Peters was only around 55% Saccone while eastern Peters was more like 62%.

Can't say Lamb ahead pre-absentees because the two remaining precincts are in Westmoreland, so we don't know where they are, and some precincts have been netting Saccone more than 95 votes on their own. However, at this point I think Lamb should win overall by a few hundred votes once absentees are counted regardless of what the pre-absentee results are.

[Tintrlvr]

Alleghaney Absentee

1930 Lamb
1178 Snowcone

Is that all of the absentees? I thought there were supposed to be ~4200 in Allegheny.

[Tintrlvr]

Alleghaney Absentee

1930 Lamb
1178 Snowcone

Is that all of the absentees? I thought there were supposed to be ~4200 in Allegheny.

That many were sent out I believe. These were returned.

Got it.

Does seem like Lamb wins at this point. Can't see Saccone managing 848 votes net from maybe 3000 absentee ballots, max.

[Tintrlvr]

Saccone needs to pull 850 votes out of the remaining 3200. I think he'll net about 100 absentee votes in Greene but probably not enough in the other two to make up for it. I'm still really nervous about calling this though.

He can't net 100 in Greene if there are only 200 ballots. That would be a huge overperformance compared to on-the-day results. If absentees are the same as on-the-day (and they were significantly better than on-the-day for Lamb in Allegheny), Saccone nets 30-35 votes from Greene.

[Tintrlvr]

Saccone needs over 70% of outstanding votes to win, according to CNN.

If he couldn't do that with Westmoreland and it was his best county.  No way he does it with Greene and Washington

Greene was his best county, actually, but he came nowhere close to 70% there. And obviously did not in Washington, either, which is the vast majority of the remaining absentees.

[Tintrlvr]

Interesting thread...because the DCCC is so openly bragging after the fact and outlining what it did:

https://twitter.com/dccc/status/973890765981011968

https://twitter.com/dccc/status/973892233928822784

https://twitter.com/dccc/status/973892236126703616

https://twitter.com/dccc/status/973892241864478721

Honestly probably a much better investment than spending a zillion dollars on TV ads, which should only be a focus for candidates already completely flush with cash for organization anyway (think Obama 2008 GE).

[Tintrlvr]

Basically, there'll be 200 absentee ballots from Greene that break 50-50, 200 provisional ballots counted that break 60-40 Lamb, and 200 military ballots counted that break 60-40 Saccone. The end result? Not a damn thing changes, neither in outcome nor in margin.

Even if 100% of those votes went to Saccone, he would still lose.

Didn't Doug Jones win the military ballots in Alabama? I would not assume they favor Saccone.

[Tintrlvr]

Interesting. The patterns are all very general, but it seems like Casey did best in super-Republican, ancestrally Republican areas, Wolf did best in ancestrally Democratic areas that have swung hard to the Republicans and Lamb did best in suburban areas that have swung towards the Democrats.

[Tintrlvr]

Are the absentee and military ballots going to be reported at the precinct level or the County level?

The absentees are all counted and were reported at the precinct level. Don't know about military ballots. There are also provisional ballots still to be counted. Not sure on PA law on the issue, but usually voters have a few days or weeks to verify their eligibility to vote to get their provisional ballot counted, so those will be counted over the next couple of weeks or so. Military ballots likewise have some time because they only have to be postmarked by election day, so those will be counted over the next couple of weeks also.

I doubt there are more provisionals+military ballots than Lamb's current 627-vote lead. Even if there are, provisionals will favor Lamb (while military may or may not favor Saccone; they tend to be more variable), so there's no chance of votes still outstanding changing the result. The only way the result could change at this point is if a serious tabulation error emerges. But there don't seem to be any precincts that might be prospects for such an error.

NYT now has Westmoreland precinct data.

I hate that you can't zoom in on their maps.

[Tintrlvr]

Another terrific map by Miles.

Wow, Rothfus is done. I didn't realize that Mt. Lebanon and the surrounding area was nearly all in the new PA-17; Lamb should be a lock to win.

Both Romney and Trump won that District overall. Lamb is favored but not a lock.

Yes and no. He wins if he gets exactly the same percentages as Clinton in the parts of the district that aren't in the old PA-18 and gets exactly the same percentages he got in the special election in the parts of the district that are in the old PA-18. That's as close to a guarantee as you can get.