Washington 2020: The Calm Before the Drizzle

[Alcon]

Yeah, second-wave absentees are much more Republican, it seems.  Spokane County just fell from Obama +3 to barely Obama +0.  It will likely flip to McCain.

On the other hand, King County is still underrepresented in returns.  It's 29% of the state's population and about 44% of remaining ballots.

Edit: Although Pierce County is not at all reporting remaining ballots.  Still, I doubt the Democrats are going to expand their margins much.

[Alcon]

Late ballot effect looks pretty big.  Wahkiakum swung something like 2.5-3%, and there are still military ballots out.  Obama will probably finish up 14% or 15%, is my guess.  Gregoire will tighten a few points.  That includes the high pending rate in King County.

The Presidential race in Okanogan County, write-ins only:

Ron Paul - 32 (not inc. "John Paul")
Hillary Clinton - 19
Bill Cosby - 2
Bob Barr (idiot) - 1
Jesus Christ ("because no one above is worthy") - 1
Mitt Romney - 1

The best part: 7 of Clinton's 19 write-in votes were misspelled.

In the congressional race, Ralph Nader (2) leads, tailed by Joe the Plumber (1).

Someone wrote in "Elvis" for every office on the ballot other than President.  And in the Attorney General race, someone actually wrote in "Undecided."  A Mr. Dick Wanker put up a strong showing in the 7th District, with three votes.

[Alcon]

OK, so the going rate seems to be that, in the later absentees, Obama loses about 4.5% and McCain gains the same.

Under this paradigm, I set all projected county turnouts to 2004 level * 10%.  A few counties may break this, but it should be about the going rate.

Extrapolated, assuming I did my math correctly, I can now extrapolate the following results.  In order for this race to tighten, you'd need either a comparatively low King County turnout or exponentially better GOP results in later balloting.  In order to make this a 15% win for Obama, McCain would have to swing late ballots like 15% or the aforementioned low Democratic turnout vs. 2004.

President
Obama 2,518,385 (57.97%)
McCain 1,755,114 (40.40%)
Other 70,688 (1.63%)

Margin: Obama +17.57%

I didn't check, but I assume Gregoire is seeing about the same loss.  As such, I can project the following mathematical results

Governor
Gregoire 2,273,159 (53.53%)
Rossi 1,973,025 (46.47%)

Margin: Gregoire +7.06%

[Alcon]

State Treasurer
McIntire 2,024,645 (51.38%)
Martin 1,915,540 (48.62%)

Margin: McIntire +2.76%

Commissioner of Public Lands
Goldmark 1,994,180 (50.63%)
Sutherland 1,944,850 (49.37%)

Margin: Goldmark +1.26%

[Alcon]

Can you do that with Burner/Reichert or is it too complex to break it up?

Well, Reichert leads and my methodology assumes that later ballots are going to be more Republican than current ballots...

But King County's dump today was actually slightly more Democratic, and then another more Republican one, I think.  That also happened in San Juan -- the dump today was 73% Obama when the original was 71% Obama.  I think Reichert briefly gained a bigger lead and then it shrunk?

In any case, my model says:

Burner 232,571 (46.51%)
Reichert 267,494 (53.49%)

But I'm way uncomfortable with using it this for this.  Either way, I do think Reichert will win.  King County's WA-8 portion is barely under-represented vs. Pierce County's (about 49% in vs. about 48%).  Late ballots being more Democratic seems a lot less likely than them being more Republican, even if they're barely more Republican.

[Alcon]

Want some consolation?

King County just dumped about 75,000 new ballots.  The breakdown was essentially identical to the pre-existing Democratic margins, over 70% Obama.

Which oddly made my model's margin tighten, which doesn't make sense, since he beat expectations.  wtf?  arghh.

[Alcon]

Commissioner of Public Lands
Goldmark 1,893,427 (50.21%)
Sutherland 1,877,463 (49.79%)

[Alcon]

I don't like this model.  It's exponentializing more than I expected.  I'm going to create a new model tonight, which extrapolates uniformly based on new ballot returns.  This will bias toward the Democrats if late absentees are even more Republican than second-wave absentees, as I expect they will be.  But as it stands, my model is too biased toward the GOP.  I expect that the truth will be about 2/3 of the way between the current model and the new one.

I'll figure this out!

[Alcon]

OK, so here are the new models.  I also fixed a stupid Excel error that was causing it to inflate turnout drastically.

Model A (dumb projection) assumes that all future ballots will be cast in exact proportion to their current ratio.  This measure does not take into account the late ballot effect at all, and thus it has a Democratic bias.

Model A (Obama +19.79%)
Obama 1,857,676 (59.07%)
McCain 1,235,248 (39.28%)
Other 52,068 (1.66%)

Model B (old model) assumes that all future ballots will be about 4.5% more Republican and 4.5% less Democratic than current ballots.  This should be the most Republican-friendly model, and should become moreso as time goes on.

Model B (Obama +16.33%)
Obama 1,803,305 (57.33%)
McCain 1,289,619 (41.01%)
Other 52,068 (1.66%)

Model C (new model) takes the earliest number I have for each county, and then subtracts it from the current totals.  It takes this new number, which is the break-down for second-wave ballots, and extrapolates based on it.  I think this is the most neutral, but two things tell me it may have a slight Dem bias.  First, early King County second-wave ballots look more like first-wave ballots to me, but I can't tell.  Secondly, my suspicion is that last-wave ballots are even more Republican because they're more likely to be military.  On the other hand, if there is no new ballot information, Model C defaults to Model B's Republican-leaning methodology.  So, Model C's slight Democratic lean is probably neutralized at the moment.

Model C (Obama +17.76%)
Obama 1,825,548 (58.05%)
McCain 1,266,946 (40.28%)
Other 52,498 (1.67%)

My current margin projection would be about Obama +16.5-18.0%.  Unless of course my modeling has a fatal flaw, which is always possible.  We shall see.

Gov, Lands Comish, and Trez coming up.  I don't have older counts for these races, so we'll have to wait for new data to come in for Model C to work.

[Alcon]

President
Model A (Obama +19.79%)
Obama 1,857,676 (59.07%)
McCain 1,235,248 (39.28%)

Model B (Obama +16.33%)
Obama 1,803,305 (57.33%)
McCain 1,289,619 (41.01%)

Model C (Obama +17.76%)
Obama 1,825,548 (58.05%)
McCain 1,266,946 (40.28%)

Likely range: Obama +16.5-18.5

Governor
Model A (Gregoire +9.82%)
Gregoire 1,726,993 (54.91%)
Rossi 1,417,999 (45.09%)

Model B (Gregoire +6.32%)
Gregoire 1,671,834 (53.16%)
Rossi 1,473,158 (46.84%)

Model C
Insufficient data

Likely range: Gregoire +6.5-8.5

State Treasurer
Model A (McIntire +5.75%)
McIntire 1,662,859 (52.87%)
Martin 1,482,134 (47.13%)

Model B (McIntire +1.88%)
McIntire 1,602,096 (50.94%)
Martin 1,542,896 (49.06%)

Model C
Insufficient data

Likely range: McIntire +2.0-4.0

Commissioner of Public Lands
Model A (Goldmark +3.77%)
Goldmark 1,631,765 (51.88%)
Sutherland 1,513,228 (48.12%)

Model B (Sutherland +0.06%)
Goldmark 1,571,510 (49.97%)
Sutherland 1,573,483 (50.03%)

Model C
Insufficient data

Likely range: Goldmark +0.0-2.0

[Alcon]

Unless King County stops delivering dumps that are more Dem-favorable than their initial results, which it just did for the third time in a row, Sutherland is toast.

Martin is now probably dead in the water either way.

Burner may hold on yet (she tied the latest batch), but it's unlikely.

[Alcon]

I also have a feeling that the Pierce County results are going to be more Democratic than the last few batches. Not as Democratic as the initial batch though.

I have a feeling you may be right.

Every single second-wave release so far today (Adams, Grant, King, Skagit, Whatcom, and Whitman, Yakima) has been better for Obama, and even moreso the down-ticket Democrats, than the initial results.  No idea why.  McCain may even fall under 40%, at this rate (maybe.)

But, seriously, if these sorts of numbers hold up in the Puget Sound, the Washington GOP is officially no longer a viable political party.  Gregoire is now projecting as beating Kerry's margin by nearly half a point.  That has to sting for the GOP.

[Alcon]

(Blog copy)

Model A: "Dumb" model (no late ballot effect)
Model B: Old model (Republican adjustment method)
Model C: New model (New ballot extrapolation method)

Other than prez, Model C still could use some more information, but it's solid enough.  It's the best one.

President
Model A: Obama +19.14 -0.65
Model B: Obama +16.37 +0.04
Model C: Obama +17.83 +0.07

Governor
Model A: Gregoire +9.07 -0.75
Model B: Gregoire +6.31 -0.01
Model C: Gregoire +7.67

State Treasurer
Model A: McIntire +4.66 -1.09
Model B: McIntire +1.96 +0.08
Model C: McIntire +2.65

Commissioner of Public Lands
Model A: Goldmark +2.90 -0.86
Model B: Goldmark +0.22 +0.28
Model C: Goldmark +1.40

U.S. Congress - Dist. 8 (still not comfortable)
Model A: Reichert +1.48
Model B: Reichert +8.73 (lol, what can you do?)
Model C: Pending

All in all, we’re probably look at about Obama +17.5,  Gregoire +7.5, McIntire +2.5 and Goldmark +1.5.  I think, in pretty much every case, a slightly closer race (due to late absentees) is more likely than bigger Democratic margins.  I can’t promise anything, though.

Reichert looks pretty solid in WA-8 at first, but I want to give Model C a go when Pierce and King release new information tomorrow.  Why?  Because the late ballot effect seems much less pronounced here (especially King), and accordingly Reichert +8.7 is utterly ridiculous.

[Alcon]

Not sure what way the poll votes are going to cut, though.

In the February primary, 59% of absentee voters voted Democratic compared to only 45% of poll voters.

In the August primary, Gregoire got 48% of absentee voters and 45% of poll voters.

Of course, dramatically different turnout situation...in a General, poll voters are probably more Democratic, you think?

[Alcon]

Poll voters have always been more Republican for as long as I can recall... were they in the 2004 general?

I don't have the 2004 data file anymore, which sucks, because it cost $20.  I'll have to trust you on them being more GOP.  It sounds plausible.

If they're still more Republican, and that last batch was absentees+them, very good news for Democrats in Pierce County races.

Would also help explain why Marilyn Rasmussen is down.

[Alcon]

Yeah, the latest King County batch was def. on the Republican side.  McCain actually got 31.2% of it (!!!!!!!!111~~).  Unless King goes back to more more-Democratic-than-before batches (wonder what that was) Burner is done for.  Although she probably is either way.

The CPL call was a mite premature, imho.

[Alcon]

Burner = doomed

Model A: Reichert +3.1
Model B: Reichert +9.2
Model C: Reichert +5.3

Probably won't even be especially close.  Rodney Tom?  Can't believe the WA Dems nominated her again.

[Alcon]

Early results show that McCain did not break 33% in the 48th.  Words fail.

That's an 18-point swing from 2004 in a suburban district.  Tentatively, the 5th swung 15 points; the 45th swung 16.

This is an area where Bush generally declined between '00 and '04, and that trended Democratic during the 1990s.  This is also a very polarized area.  That's just all-around pretty incredible.

[Alcon]

Newest King County bump was good for Burner.  Not good enough.  We've had five Kingco waves, and all of the remaining ballots would have to be 40% better for Burner than the best of those.  Reichert wins.

Can't believe the AP called CPL way before this race, but whatever

[Alcon]

Early PierceCo numbers (this will become more GOP).  Includes only Pierce County portions (which is like 20 votes in Pacific), does not include consolidations.  Kerry in parentheses.

Municipality
Auburn: 55-44 Obama (49-51)
Bonney Lake: 51-48 Obama (46-52)
Buckley: 48-48 Obama (46-53)
Carbonado: 61-36 McCain (34-63)
DuPont: 50-49 Obama (41-57)
Eatonville: 49-49 McCain (43-56)
Edgewood: 51-47 Obama (47-51)
Fife: 61-37 Obama (57-42)
Fircrest: 61-38 Obama (54-45)
Gig Harbor: 53-46 Obama (46-53)
Lakewood: 56-42 Obama (51-48)
Milton: 55-44 Obama (51-48)
Orting: 50-48 Obama (45-53)
Pacific: 52-48 Obama (37-63)
Puyallup: 54-44 Obama (48-51)
Roy: 57-39 McCain (38-62)
Ruston: 66-32 Obama (63-33)
South Prairie: 52-48 McCain (46-51)
Steilacoom: 52-46 Obama (49-49)
Sumner: 55-43 Obama (52-47)
Tacoma: 68-30 Obama (62-37)
University Place: 59-40 Obama (52-47)
Wilkeson: 49-48 Obama (52-46)

LD
2nd: 50-48 McCain (42-56)
25th: 53-45 Obama (47-52)
26th: 52-47 Obama (46-53)
27th: 69-30 Obama (63-36)
28th: 56-42 Obama (50-49)
29th: 64-34 Obama (59-40)
31st: 51-47 Obama (46-53)

Predictably, McCain just lost it in the middle-class suburbs here too.

[Alcon]

Huh:

Fort Lewis Army Base (28-465 & 28-469)
Barack Obama 525 (50.68%)
John McCain 497 (47.97%)
Bob Barr 5 (0.48%)
Chuck Baldwin 4 (0.39%)
Ralph Nader 2 (0.19%)
Gloria La Riva 2 (0.19%)
James Harris 1 (0.10%)

[Alcon]

What about the McChord precinct?

And can you do that for Gregoire/Rossi?

Sure, I'll be happy to when I get home.

McChord was like 61% McCain, and I'm pretty sure Bush broke 70% in a walk.  I'll check.

Are those the voters for the entire base? I feel like there should be alot more...

Transitory population.  Many vote at home, many don't vote.  Besides, we're only half counted or so.

[Alcon]

Pierce County ballot release tonight was really Republican, like Obama +5 at best, bringing the second wave ballots down to about Obama +8.  If that pattern continues, the county is going to tighten to single digits.

King County dump was slightly better for the GOP, but probably not enough to bring Obama under 70%.  Sutherland is looking a little closer to being hosed, principally because of this.

Meeker, to answer your question: McChord (28-467) is 62-38 McCain now, and was 72-27 Bush in 2004.  In 2004, Fort Lewis (now 51-48 Obama) was 59-40 Bush.  Smaller swing than I remembered, but still pretty big.

Governor
Auburn: 51-49 Rossi
Bonney Lake: 56-44 Rossi
Buckley: 52-48 Rossi
Carbonado: 66-34 Rossi
DuPont: 54-46 Rossi
Eatonville: 55-45 Rossi
Edgewood: 52-48 Rossi
Fife: 55-45 Gregoire
Fircrest: 58-42 Gregoire
Gig Harbor: 52-48 Rossi
Lakewood: 54-46 Gregoire
Milton: 51-49 Gregoire
Orting: 54-46 Rossi
Pacific: 57-43 Rossi
Puyallup: 51-49 Gregoire
Roy: 61-39 Rossi
Ruston: 60-40 Gregoire
South Prairie: 53-47 Gregoire
Steilacoom: 50-50 Gregoire
Sumner: 51-49 Gregoire
Tacoma: 64-36 Gregoire
University Place: 56-44 Gregoire
Wilkeson: 50-50 Rossi

LDs
2nd: 57-43 Rossi
25th: 51-49 Rossi
26th: 53-47 Rossi
27th: 65-35 Gregoire
28th: 54-46 Gregoire
29th: 62-38 Gregoire
31st: 54-46 Rossi

[Alcon]

Preliminary (although not really that far from done) Snoho results:

Arlington: 54-45 Obama (45-53)
Bothell: 62-37 Obama (55-44)
Brier: 64-35 Obama (57-42)
Darrington: 53-45 Obama (45-53)
Edmonds: 66-33 Obama (60-39)
Everett: 64-34 Obama (58-41)
Gold Bar: 53-44 Obama (49-49)
Granite Falls: 56-41 Obama (51-46)
Index: 83-18 Obama (76-23)
Lake Stevens: 57-42 Obama (49-50)
Lynnwood: 66-32 Obama (61-38)
Marysville: 55-43 Obama (48-51)
Mill Creek: 57-42 Obama (48-51)
Monroe: 55-43 Obama (46-52)
Mountlake Terrace: 70-28 Obama (64-35)
Mukilteo: 63-36 Obama (54-45)
Snohomish: 59-39 Obama (52-46)
Stanwood: 49-49 McCain (45-54)
Sultan: 52-46 Obama (44-54)
Woodway: 52-48 Obama (45-54)

[Alcon]

I think there's a chance that Seattle beat San Francisco.  A good chance, in fact.

Congressional district 7 returns:

Obama 120,114 (84.60%)
McCain 20,131 (14.18%)
Other 1,734 (1.22%)

Margin: Obama +70.4%

Now, at the time, Obama's margin was 44.3%.  It's likely to decline to about 42.0% if Obama holds up well in late-late KingCo returns.  I'll assume a decline of 2.3% or so.

Now, in 2004, WA-7 was Kerry +60.1.  Meanwhile, Seattle was Kerry +62.6.

That puts Seattle at about Obama +70.6%.  San Francisco, returns show, is currently Obama +70.2%.  Plenty of things could notch Seattle down a bit, but these early returns indicate Seattle is favored.