November 15, 2004
Monday's gubernatorial score card

I'll be updating the table below throughout the day on Monday as new reports come in. There are a few more columns than there were in Friday's table. See the bottom of the post for explanation.

Scheduled ballot returns from Whitman, Lincoln and Franklin counties did not materialize today.

Well here's an interesting surprise. King County, which claimed earlier to have 11,000 uncounted ballots, today reported 15,000 votes and claims there's another 4,000 ballots remaining. This narrows Rossi's projected lead, but doesn't predict a Gregoire victory. Of course, our model doesn't take into account the surprise discovery of thousands of previously non-existent ballots.

Time Rossi's
Current
Lead
Uncounted
Ballots
Remaining
Share of
Uncounted
in Rossi
Counties
Rossi's
Projected
Victory
Margin
"Latest"
Rossi's
Projected
Victory
Margin
"Overall"

Estimated
Lead in
Uncounted
"Latest"

Estimated
Lead in
Uncounted
"Overall"
Gregoire's
Hurdle
Counties
Reporting
Friday
end of day
1,920
41,623
60.4%
1,676
3,403
0.59%
3.56%
4.61%
 
Mon. 10:25am
1,935
41,423
60.2%
1,671
3,417
0.64%
3.58%
4.67%
Grays Harbor
11:02am
1,948
41,273
60.1%
1,658
3,419
0.70%
3.56%
4.72%
Clallam
12:17pm
1,988
38,873
57.6%
1,684
3,405
0.78%
3.65%
5.11%
Snohomish
3:53pm
1,974
36,573
55.0%
1,655
3,359
0.87%
3.79%
5.40%
Kitsap
4:23pm
1,978
36,393
54.7%
1,594
3,304
1.06%
3.64%
5.44%
Douglas
4:29pm
(657)
27,493
65.5%
448
1,365
4.02%
7.35%
*
King, Walla Walla
5:59pm
(158)
21,666
56.3%
514
1,430
3.10%
7.33%
*
Chelan,
Clark,
Kittitas,
Okanogan,
Pierce
EXPLANATION for the above table:
* A lead in Rossi's favor is shown in RED. A lead in Gregoire's favor is shown in GREEN.

* The two columns "Rossi's Projected Victory Margin" estimate Rossi's final lead under two different assumptions -- (1) "Latest", meaning that any new returns for a given county will break like the most recent returns from that county; (2) "Overall", meaning that any new returns for a given county will, on average, break like all of the previously counted votes from that county.

* Similarly, the two columns "Estimated Lead in Uncounted" estimate the percentage by which the leading candidate is ahead of the trailing candidate in the uncounted ballots using the "Latest" vs. "Overall" scenarios, as above. The number is in RED if Rossi is leading under the scenario; in GREEN, if Gregoire is leading under the scenario.

* "Gregoire's Hurdle" indicates the percentage lead she would be required to take in the uncounted ballots in order to win. Compare this number against the "Estimated Lead" figures to assess the feasibility of a Gregoire win. (For example, a 4% Hurdle for Gregoire combined with a 3% estimated lead for Rossi means that the lead in the uncounted ballots would have to suddently swing by 7 percentage points in order for her to win, which seems somewhat implausible )

A longer explanation and a table with projected leads by county (as of end of day Friday) are here

Posted by Stefan Sharkansky at November 15, 2004 12:01 AM | Email This

Comments
1. In 2000 Benton county was reporting 2,500
outstanding ballots, but the final batch of
votes was only about 50. I assumed that they
were reporting outstanding requested ballots
rather than ballots returned at hand.

Does anyone know whether Benton County's estimate
of remaining ballots is more accurate this year?

If it is, Dino's chances are excellent.

Posted by: Dave on November 15, 2004 08:17 AM
2. In the URL link for this comment I put a link to my blog entry on how to model the race in Excel. I have Rossi winning by 3400.

Posted by: Jeremy Rule on November 15, 2004 09:08 AM
3. The most excitement of the entire 2004 election story is right here in Washington.

Being one of those "immoral Democrats", I do have to admit that I voted for Dino because I felt he was the best suited for the job. In fact, the ONLY reason I went to the polls was to vote for Dino. I did not cast a vote for president because I could not in good conscience vote for any of them. I look for moderation in our political figures and Dino seemed to be the only one who offered that kind of balance.

I am not happy that on the federal level our government has moved from one extreme to another, but on the state level I am hopeful that we can move from the extreme left to the center.

Posted by: Political Pulpt on November 15, 2004 09:54 AM
4. Jeremy,

Those are terrific instructions.

I reached my Friday end-of-day "Overall" estimate of a 3,403 vote victory the same way that you reached your (rounded) 3,400 estimate.

This estimate assumes that all of the uncounted ballots will break the same way as the average of all previously counted ballots for their counties. But it's also reasonable to assume that the late counted ballots were cast by voters who are in some ways different from the voters who cast the earlier counted ballots.

Therefore, I also created the "Latest" estimate which assumes that the uncounted ballots will break more like the most recent returns. The final outcome will probably fall somewhere in between the two.

Posted by: Stefan Sharkansky on November 15, 2004 10:07 AM
5. Gray Harbor just reported 169 Dino 154 for CG

Posted by: Ralph on November 15, 2004 10:33 AM
6. According to the SOS website, Clallam just reported.

Posted by: Tim Ford on November 15, 2004 11:07 AM
7. For the record, Stefan, you rock!

Posted by: Nathan on November 15, 2004 11:14 AM
8. 1st two counties look for good for Dino. Numbers are right in line. The votes for the other candidate ticked a bit, but size of ballots is small.

Here's a trivia point for you watchers. Over 50,000 plus voters voted for Kerry or Bush and did not vote in the Governor's race. 50,000+.

Most were democrats, but some had to be republicans. It happens every four years people cast one ballot and left the booth. I suspect that even many of the absentee ballots had just one vote on them.

Posted by: Ralph on November 15, 2004 11:21 AM
9. Looks like Snohomish dumped some votes in.

Posted by: Nathan on November 15, 2004 12:33 PM
10. Snobomish reported at 12:17. Even for Dino. The libertian has picked another 1% here. Actually Dino and CG are done just a bit % wise, but Libo is rocking at 4%. This trend is not good for CG.

Posted by: ralph on November 15, 2004 12:35 PM
11. Just think what would have happened if a small % of the Libertarians had made a reality check in the poll booth. Rossi would be calling up the moving trucks and we wouldn't have this cliff-hanger to obsess over :)

ed.

Posted by: libertynews on November 15, 2004 12:53 PM
12. I have a question: Has the Republican party done anything in any of the Rossi-held counties to track down rejected provisional ballots? I haven't heard anything, and I was surprised that they hadn't done anything.

thanks

Posted by: nathan on November 15, 2004 03:13 PM
13. Nathan

Yes that is actually what spurred on the King County lawsuit. 19 of 21 counties that Rossi had good leads in had already released that information. Republicans and election officials were calling.

Posted by: Ralph on November 15, 2004 03:40 PM
14. The previous article about the King County votes connects two thoughts of mine.

1. If you look at the other or no votes in the batch of King County votes it totaled over 10%. The % of under votes and libertian votes in this race so far is about 5%. I believe that many of these provisional ballots will not have a vote for Governor. If it true that many of these student ballots it may be that will not do downballot.

2. The number of ballots out is no guarantee that a vote for Governor will be on it. I will try to look the Presidential vote changes and try to determine whether there is much of a dropoff.

Posted by: ralph on November 15, 2004 03:50 PM
15. Lead is down to 1,974. Not sure which county reported.

Posted by: Marc on November 15, 2004 03:57 PM
16. Kitsap

Everyone will be tired of me posting. 14 vote edge to Gregoire there.

Here is the big news???

1842 votes were cast in Governor's race on these ballots.

2106 votes were cast for President. 264 voters or over 10% did not vote in the Governor's race.
Now what would happen in these calculations if 10% of the remaining ballots did not have a vote for Governor on it. It would move Rossi's numbers higher and make it harder for CG.

Great news.

Posted by: Ralph on November 15, 2004 04:05 PM
17. I got the following from a Liberal Loon friend:

"Screw the numbers. I don't give a crap about numbers. The only numbers I care about are the final ones. I don't care about trends, spreadsheets, predictions, or any of that, just like I could care less about polls. I don't want to hear it. I get enough of it... already. Enough is enough!"

Valiantly defiant 'til the end. Poor guy - IT IS OVER (as I giggle).

Posted by: Josef on November 15, 2004 04:07 PM
18. Don't speak too soon. I just heard a rumor that King County has suddenly come up with 10,000 extra votes.

Posted by: Timothy on November 15, 2004 04:09 PM
19. Kitsap removed 2300 ballots from count so nearly 200 were invalidated. They only have 200 left to report.

A spoiled ballot, no vote on a ballot, or a vote for Bennett are really votes for Rossi.

I have to step out to get a child so my best wishes are with Rossi as I leave.

Posted by: Ralph on November 15, 2004 04:10 PM
20. Not wanting to plug myself to much on a conservative site, but just thought y'all would be interested to hear reports that King County found an additional 10,000 ballots: HorsesAss.org.

Just a friendly scoop from "Aluminum Hat Boy."

Posted by: David Goldstein on November 15, 2004 04:21 PM
21. HOLY S--T. WHERE DID THESE BALLOTS COME FROM? OUTER SPACE? THE GREGOIRE PRESS?? THE KGB ARCHIVES???
(Just having fun with ya posting about it. I cannot be brought to accept at this point anything except a Rossi victory.)

Posted by: Josef on November 15, 2004 04:23 PM
22. One last thing: The anal yahoos who whine about "AL hat boy" and the anal yahoos who whine about the Marlboro boy should both SHUT UP. As in Shut (many expletives) UP.

David Goldstein's different, that's all. He's also smart, you know.

The Marlboro boy of Fallujah is a hero and a patriot who should smoke, smoke, smoke if that's what keeps the boy alive.

Posted by: Josef on November 15, 2004 04:26 PM
23. King just dropped in 7,000 votes. Rossi is now down by 657 total votes.

Posted by: Marc on November 15, 2004 04:36 PM
24. Actually, as per my previous post, King dropped in 17,000 votes.

Posted by: David Goldstein on November 15, 2004 04:42 PM
25. King Co. just dumped nearly 15K ballots...they were only supposed to have 11K on hand. They say they have another 4K to go. My excel model now shows a projected Rossi win by 1900+ votes.

Can you say re-count????

Posted by: Kevin Shannon on November 15, 2004 04:45 PM
26. KIng county had 16,871 Presidential votes.

Only 14,989 Gubernatorial votes. Rossi is now down 657, but I believe that this will be a low point.

Walla Walla came in okay with only 2% dropoff.
2111 votes in Presidential and 2056 in Governor's race. King had 14.5% of ballots with either no vote for Governor or a vote for Bennett.

I still think Rossi has an edge.

Posted by: Ralph on November 15, 2004 04:50 PM
27. Well, if this truly is "a special sort of heaven for political bloggers," I guess things just got a bunch more "heavenly."

Still, though, these most recent votes broke better for Rossi than they have been recently, so things are still looking good. But, yeah, I can say "recount." We'd all better learn how to, unless other counties suddenly find a bunch more votes too.

Posted by: Timothy on November 15, 2004 04:54 PM
28. Yeah, looks that way. Hey, check out http://www.secstate.wa.gov/office/osos_news.aspx?i=TjrRx9TQ4Kv037C9GJJtuw%3d%3d and see how not a single recount has changed the final outcome.

Posted by: Josef on November 15, 2004 04:58 PM
29. I must say, I'm highly suspicious of this surprislingly high vote count from King Co.

How convenient for the Democrats that their best county suddently comes up with 40% more votes than originally thought.

And they still have 4K left...

Posted by: Kevin Shannon on November 15, 2004 05:00 PM
30. A thought...

"Everyone with a computer is a statistician, and everyone with a mouth is a pundit."

-- Gregoire campaign consultant Christian Sinderman
http://www.spokesmanreview.com/olympia/blog.asp

Posted by: Josef on November 15, 2004 05:03 PM
31. It's so funny that when ballots are counted which do not favor your hopes, and you were not expecting them, you term these ballots 'previously non-existent' instead of 'previously not reported.'

Of course it's not the truth that these votes previously did not exist. The truth is that so many ballots requires a lot of counting, and so the report was updated -- to your dismay.

Have a good night. I hope this helps you with your thinking.

Posted by: Heath Hunnicutt on November 15, 2004 05:03 PM
32. Well, why weren't adequate records kept of the ballots?

Posted by: Josef on November 15, 2004 05:08 PM
33. I don't know. Maybe you should get a job working the polls next election. Then you can see whether you think that a 4% tabulation error on 900,000 ballots after 8 business days have elapsed from the election still seems unreasonable to you.

Posted by: Heath Hunnicutt on November 15, 2004 05:16 PM
34. King County has been receiving ballots everyday up until the 12th. It appears to me that when they counted ballots from the 8th on they substracted ballots counted, but did add ballots received on the 9th, 10th, 11th and 12th. (I realize that the 11th was a postal holiday, but they must receive their ballots by PO box)

Bureacratic slipup. The uncounted ballot number supplied to the state is for informational purposes only. The site very clearly says that they cannot those numbers.

Posted by: Ralph on November 15, 2004 05:19 PM
35. That was one hell of an underestimate. Lets hope some Rossi friendly counties underestimated their remaining ballots too.

Posted by: Nathan on November 15, 2004 05:34 PM
36. Well, okay then I'm going to ask this:

How is it that it takes so long for ballots mailed 2 November to get to the county auditors?

I mean, I mail something to a friend in Olympia from Sedro-Woolley it takes 2, 3 days absolute max. I would tend to think the ballots with their special envelopes would be there in 2 days.

Josef

Posted by: Josef on November 15, 2004 05:34 PM
37. On the bright side, those King County votes didn't break nearly as badly as I thought they would. Working with the 11,000 ballot estimate I'd come to the conclusion that at this point Rossi would be down by over 1,000 votes. Cheer up, ya'll, 65% of the remaining ballots are in Rossi's counties.

Posted by: Nathan on November 15, 2004 05:38 PM
38. Clark County just posted. Bit of mixed bag for Rossi. This county, if accurate, had fewer ballots then expected. Pierce and Whitman are last hopes for tonight. 453 lead for CG. I still believe Rossi will win. Maybe 1,000 or maybe 500.

As my old buddy Dan Rather said "This race is tighter then a jar lid on a bottle of stewed tomatoes".

Posted by: Ralph on November 15, 2004 05:42 PM
39. Thanks, Nathan. Keep up the good work. The last gasp of Gregoire? We hope.

I just wish it was Goldstein (or Talmadge or to some extent, Sims) and not Gregoire. Then we'd have a REAL campaign on ideas and less of a divide. My .02 .

I hope all of you join me in praying this doesn't end up dividing the state and that both sides play this fair 'til the end.

Posted by: Josef on November 15, 2004 05:44 PM
40. Gregoire's lead is down to 324. I'm not sure who reported in.

Posted by: Nathan on November 15, 2004 05:49 PM
41. You know a lot more about statistics than I do, and I appreciate what you're doing. On the other hand, I've been watching election returns from places like Wayne County, Michigan, and Cook County, Ill. for more than forty years, and don't be surprised: if they need to find 5,000 more votes in King they'll find them, if they need to find l0,000, etc.

Posted by: George on November 15, 2004 05:51 PM
42. Heath:

Maybe YOU should get a job working at the polls next time, or keep your sarcasm to yourself.

I did work the polls this election, as a volunteer no less! I'll be waiting for your thanks, by the way!

But I DO know that every precinct made a physical count of ALL absentee and ALL provisional ballots THAT NIGHT!

So the questions are:
1. Why can't King County Election workers make a proper tabulation (sum) of a list of numbers?
2. Why would hourly King County Election employees want a recount? (Can anyone say OVERTIME PAY?)

Like I said, HEATH, I await your answers, and your thanks to me for taking a day off my high-paid IT job to VOLUNTEER on election day.

You're welcome!

Posted by: Larry on November 15, 2004 06:07 PM
43. Gregoire's lead is 158 now. I would love to see Rossi end the night with a lead.

Posted by: Nathan on November 15, 2004 06:18 PM
44. New counties reported since Stefan's last posting
Chelan, Clark, Okenagan, Kittias

I punched the new numbers (and revised uncounted estimates from washington state website) into Tim Goddard's model and it yields a projected lead of 189 to 1425 for Rossi.

Posted by: John B on November 15, 2004 06:20 PM
45. New counties reported since Stefan's last posting
Chelan, Clark, Okenagan, Kittias, Pierce.

Pierce county now reporting since my last post (previous post without Pierce just above)

PIERCE: 1984:Gregoire 2150:Rossi (good news for Rossi) (many more ballots and bigger rossi margin). Pierce has 106 more to count

Gregoire now leads by 158 in COUNTED ballots

I punched the new numbers (and revised uncounted estimates from washington state website) into Tim Goddard's model and it yields a projected final margin of 351 to 1457 for Rossi.

Posted by: John B on November 15, 2004 06:36 PM
46. Read Dave's comment (first comment all the way at top). He says Benton county drastically overstated its uncounted ballots in Year 2000. They said they had 2500 ballots in 2000 but only actually had 50 per Dave.

If Benton county turns out to again only have 50 ballots left this year rather than the 2,200 shown on washington state web site then the contest is MUCH closer!

Using Tim Goddard's model including 5:59PM Pierce County results (see above post) but INSTEAD using only 50 remaining uncounted Benton County ballots rather than 2200 yields a range of projected margins of Rossi winning by +625 to Gregoire winning by +396. Yikes!

Posted by: John B on November 15, 2004 06:52 PM
47. Hey John B - Where are you now? Need an update. Why is it that all of a sudden King County has thousands more votes?? However many they need for a G victory, seems they will get. Hopefully the republicans are on top of this!

Carol P

Posted by: Carol Petke on November 15, 2004 08:06 PM
48. I am still confident of Rossi's victory.

On the SOS's website that details the 21,066 ballots still to be counted look at % turnout.

The republican counties, for most part, are the counties with the lower % of turnout. That tells me that they do have more ballots to turn in.

I believe it will happen for Dino.

Posted by: Ralph on November 15, 2004 08:27 PM
49. Possible explanation for the "extra" 10,000 ballots reported today from King COunty.

After the Friday 11/05/2004 update report, King County said it had 119,000 ballots left to count.

I remember reading some report that King County counted 63,000 ballots on Monday 11/08/2004.

However, after the Monday 11/08/2004 update report, King County said it had 46,000 ballots left to count.

If 63,000 ballots were counted on 11/08/2004, this figure should have actually been 56,000 ballots.

I sure hope that Benton and Yakima Counties have at least as many uncounted ballots as they are reporting.

Posted by: Richard on November 15, 2004 08:33 PM
50. "Some report" isn't good enough intelligence. Heck, we launched a war to find tons of WMD on that and found enough to fill a pickup truck, not a squadron of C-5s.

That said, IF I WERE AFTON SWIFT AND THEREFORE ROSSI'S CAMPAIGN MANAGER (what would I give for bold): I'd rightfully demand an emergency injunction until the mystery ballots could be verified to keep them from the count. Not going to the courts on this one is stupid, if you ask me. If Gregoire would have done (and may do) the same thing, then there's nothing wrong with that.

Posted by: Josef on November 15, 2004 08:43 PM
51. Josef,

On what grounds would Rossi's people take this to court? On the grounds that King County underestimated the remaining number of ballots? That's hardly actionable.

Unless there is something a bit more solid to go on, charging vote fraud because the County's estimate was lower than the actual number of ballots would be like charging vote fraud because the exit polls were higher than the actual votes.

If we're going to (rightly) scoff at the liberal conspiracy theories, we can't start spouting our own. Now I'm not saying don't be skeptical here, but don't jump the gun either.

Posted by: Nathan on November 15, 2004 09:27 PM
52. Well, King County isn't Cook County, the director of elections in King County is actually a Republican, even though he works for Ron Sims, and the reason the count is so uncertain is primarily that the provisionals are not included in the original counts because they often are not from the precinct- or even the county- in which they were cast. They wouldn't show up in a count in the best of circumstances until they arrived in the right county.

And many of the provisionals are student votes, cast in Whatcom, Whitman, Kittitas, or wherever, and then sent to the home county- the bulk going to King, because that is the largest county.

Quirks in the counts showed up last week in Rossi counties- shortly after King reported today, Okanogan county (having previously reported no votes left to count) reported 8 Rossi, 7 Gregoire.

There is nothing sinister going on- and Rossi is still going to win.

Posted by: Jim King on November 15, 2004 10:18 PM
Post a comment
Name:


Email Address:


URL:


Comments:


Remember info?