Here's is your definitive cut-to-the-chase scoreboard for the manual recount.
UPDATE Time: Tuesday, Dec. 14, 4:40pm
[includes Chelan, Clark, Cowlitz, Franklin, Grant, Klickitat, Okanagan, Whitman]
NOTE: I'm compressing and combining the two tables I had earlier. The new item, "Momentum Change in Lead", indicates the change in the lead considering the new ballots that have been added.
| Counties Reported |
32
|
|
| % of State Vote Reported |
31.8%
|
|
| Apparent Gain in Lead (+Rossi / -Gregoire) |
+64 | |
| Imputed Lead (assuming no new votes in unreported counties) | +106 | |
| Momentum Change in Lead | no change | |
| Net change in 2-party votes | +432 | +0.05% |
| Tabulation error | 0.05% | |
| Total Lead in Reported Counties ( +Rossi / -Gregoire) |
+128,493 | +14.9439% |
| Lead in Machine Recount from Unreported Counties | -128,387 | -6.8091% |
| Rossi lead in machine count in unreported Ex-King | +26,322 | |
| Extrapolated Rossi lead in unreported Ex-King | +26,334 | |
| Required lead in King for Gregoire to reverse | -154,827 | |
| Gregoire lead in machine count in King | -154,709 | |
| Required increase in Gregoire's lead in King manual recount for reversal |
-118 | |
| New two-party votes required in King (assuming same breakdown as in machine recount) |
655 | |
| Imputed tabulation error in King County to produce required new two-party votes | 0.08% |
| County |
Net Lead in Recount
(+Rossi / -Gregoire) |
Tabulation Error
|
| TOTAL |
+64
|
0.05%
|
| Adams |
-5
|
0.10%
|
| Asotin |
+5
|
0.18%
|
| Benton |
+2
|
0.01%
|
| Chelan |
+1
|
0.00%
|
| Clallam |
0
|
0.02%
|
| Clark |
+2
|
0.03%
|
| Columbia |
+1
|
0.05%
|
| Cowlitz |
+1
|
0.01%
|
| Douglas |
0
|
0.02%
|
| Ferry |
0
|
-
|
| Franklin |
+5
|
0.15%
|
| Garfield |
0
|
-
|
| Grant |
+7
|
0.03%
|
| Grays Harbor |
-2
|
0.06%
|
| Island |
-4
|
0.03%
|
| Jefferson |
-2
|
0.07%
|
| King | ||
| Kitsap |
+10
|
0.13%
|
| Kittitas |
0
|
-
|
| Klickitat |
0
|
-
|
| Lewis |
+7
|
0.04%
|
| Lincoln |
0
|
-
|
| Mason |
+3
|
0.09%
|
| Okanogan |
+4
|
0.11%
|
| Pacific |
0
|
0.01%
|
| Pend Oreille |
+2
|
0.03%
|
| Pierce | ||
| San Juan |
0
|
0.01%
|
| Skagit | ||
| Skamania |
+2
|
0.08%
|
| Snohomish | ||
| Spokane | ||
| Stevens |
+2
|
0.01%
|
| Thurston | ||
| Wahkiakum |
-1
|
0.04%
|
| Walla Walla |
+11
|
0.08%
|
| Whatcom | ||
| Whitman |
-2
|
0.05%
|
| Yakima |
+15
|
0.07%
|
How would you handicap the re-cap at this stage?
BTW, I really love this site.
Posted by: Good Captain on December 10, 2004 12:51 PMhttp://www.votewashington.org/recount/recount2_sw.tpl?c=0
Some fellow in Centralia by the name of Larry Hewitt runs this site. He runs elections websites for several county auditors, mostly for the smaller counties. His website usually posts results for these counties before they are posted on the SOS official website.
votewashington.org has Asotin County reporting +5 for Gregoire, +10 for Rossi, and +1 for Bennett. That makes Rossi a net +23 in the manual recount so far, or ahead by +65 in the overall votes.
e.g. Walla Walla changed +2 Gregoire, +13 Rossi, -3 Bennett, which is a total of 18 changed votes, divided by 23,269 votes in the machine recount.
For the statewide number, just add up all the county changed votes and divide by total number of ballots in the machine recount statewide.
Posted by: Stefan Sharkansky on December 10, 2004 12:57 PMThe democrats planted them so when King County comes in at +103 and Gregoire wins by 2 votes, they can say, "everyone was off and we were just off a little more..."
- Mike
Posted by: Mike on December 10, 2004 01:52 PMKitsap: Machine Gregoire +15 Rossi -19 Net Gregoire +34. Manual Gregoire +72 Rossi +82 Net Rossi +10. Overall Gregoire +87 Rossi +63 Net Gregoire +24.
Walla Walla: Machine Gregoire +59 Rossi +37 Net Gregoire +22. Manual Gregoire +2 Rossi +13 Net Rossi +11. Overall Gregoire +61 Rossi +50 Net Gregoire +11.
So much for accusations of Republican voter fraud. Besides, the auditor in Kitsap County, Karen Flynn, is a Democrat. Not sure about Walla Walla County.
What is interesting about Adams County is that they counted 5,200 ballots the first time, 5,207 ballots in the machine recount, and 5,204 ballots in the manual recount. (BALLOTS, NOT VOTES) Granted, 7 ballots is not a lot, but it is large in proportion to county size (total variance in ballots counted of 0.14%).
What worries me about King County is not so much whether or not new votes will be found on ballots that were already counted, but whether new ballots will be counted that weren't counted the first two times. I am not talking about the Dems lawsuit, but about having 336 new ballots show up in the machine recount without explanation (898,574 vs. 898,238).
If 336 more new ballots showed up in the manual recount from this same unexplained variance phenomenon, this would be expected to add about 60 net votes for Gregoire, given her 18 point margin in King County.
Posted by: Richard Pope on December 10, 2004 02:27 PMAgain, anyone have any real information? Are there other possible explanations? I think I know which of my three explanations is most likely, which has been the crux of the opposition to the hand count all along.
Posted by: srogers on December 10, 2004 02:31 PMAnd does the loss of five Dino votes mean, therefore, that during the hand recount, five of the votes that were "found" for Rossi (or were counted for him from the 7 additional ballots) were, on second thought, rejected (as in "we now determine that the voter didn't intend to vote for him")?
Posted by: srogers on December 10, 2004 02:43 PMThis is now +21 Rossi in the manual count for the counties that the SOS has on its website, or +26 Rossi if Asotin County (+5 Gregoire, +10 Rossi, net +5 Rossi) is considered -- reported on votewashington.org, but not on the SOS for some reason.
Thanks to Jefferson County for the correction and the additional Rossi vote.
Posted by: Richard Pope on December 10, 2004 03:04 PMThanks
Posted by: Ralph on December 10, 2004 03:15 PMAsotin County still is not on the SOS website, even though their manual recount was completed several hours ago. They are posted on votewashington.org +5 Gregoire, +10 Rossi, or a net of +5 Rossi. This would make the overall manual count of +33 Rossi.
I called the Asotin County Auditor's office, and they said they had sent the report to the SOS office several hours ago. I called the SOS office, and they said Asotin County had not done so. The lady in Asotin County said she would look into it, and hopefully correct the problem.
Posted by: Richard Pope on December 10, 2004 03:56 PMEach county has x number of ballots that went to their canvassing board in sealed envelopes for review. Franklin County had 20 according to the Republican head site monitor. I bet Franklin reports the mass count, but won't include the 20's disposition until later.
If true, be ready for county numbers to change a bit again before the bitter end.
Am I a pundit yet?
Posted by: Matt on December 10, 2004 10:29 PM"I was under the absolute impression not only I voted, but followed the instructions correctly," Phillips said. "If it can happen to the King County Council chairman, it can happen to anyone else."
Told you guys that Republicans would lose this one, now what, to the courts, whine about it, or do what you told the Democrats to do, suck it up and admit defeat.
Your call
Posted by: Sha on December 13, 2004 02:17 PMLike I said: "garbage in, garbage out."
Posted by: David Goldstein on December 13, 2004 02:50 PMshould criminal procedings be initiated everywhere? Even where Rossi gains votes?
Posted by: tom on December 13, 2004 03:51 PMActual truth is both sides will claim whatever they want to get their guy elected. Honestly saying that would make things much more palatable for the citizenry.
Posted by: tom on December 13, 2004 04:22 PMA 0.08% tabulation error would only apply if 712 new two party (D or R) votes were required out of King County. But since 535 new two party votes are estimated out of the 561 newly countable ballots, this means only 177 more votes need to be divined from the existing ballots. And the tabulation error to divine 177 more votes from the 898,574 (or so) existing ballots is only 0.02%.
The situation, needless to say, does not look very good. An additional tabulation error of 0.02% from King County in a manual recount would be fairly easy to achieve.
Posted by: Richard Pope on December 13, 2004 06:33 PMOf course this doesn't take into account the error rate in any of the uncounted counties (including King). It really is going to be a toss-up. Again.
Posted by: Nathan Azinger on December 13, 2004 07:56 PMGregoire is gaining votes in the hand recount at a faster rate than Rossi. In 24 counties reporting so far, she has turned 231,875 machine-counted votes into 232,019 hand-counted votes, a gain of 144 votes. Rossi has turned 317,910 into 318,100, a gain of 190. If the additional votes were distributed as the machine counted ones were, Rossi would have gained 193 and Gregoire only 141. Rossi is gaining about one vote less per hundred among the additional votes than among the machine counted votes. This is why I say Gregoire is gaining votes in the hand recount at a faster rate than Rossi.
Extending this rate of vote-gaining over the entire machine counted vote suggests Rossi picks up a total of 820 votes in the hand recount and Gregoire picks up 852. If that's the end of it, Rossi wins by 10 votes, 42 vote lead from machine recount less Gregoire's imputed 32 vote margin in additional votes from the hand recount.
However, there are the additional 561 ballots in King county. If these are distributed as the machine counted votes were, Gregoire gets 316 and Rossi gets 219. So this 97-vote margin gives Gregoire a win by 87 votes.
Here's some of what's wrong with my analysis.
Different voting technologies are likely to produce different amounts of additional votes. Are the different technologies in use represented in the same proportions in the 20% of the ballots so far reported as they are in the state as a whole? Probably not.
It would be great if we had a list of the counties showing the technology used for polling place votes (SecState website shows this piece) and the one used for absentee votes. Then, we would also need to know the number of votes for each of the two candidates in each county produced by each technology. Is this information available somewhere?
Richard Pope's thoughts on the additional King county ballots suggest that the machine-counted ballots may not be a useful predictor of the votes on the additional ballots.
I'm interested to learn what else is faulty in my analysis.
You're a great host, Stefan.
Posted by: John Milem on December 13, 2004 09:27 PMWith the King County 531 extras, every + overall helps.
Posted by: Steve on December 14, 2004 03:51 PMBut it sure would make life easier if anyone had any hard data...
Posted by: Timothy on December 14, 2004 04:47 PMFind Just.Enough.Votes late, from your Friendly Polling Places. At least there hasn't been any reported use of "dead" names, or animals. Yet.
-jbas.
Posted by: jbas on December 14, 2004 04:54 PM
Chris Vance just said on KVI that the 'found' ballots are no longer 561... it's 577.
Stephan, just move the 'Votes they need to find' down to, say, 100 so they'll stop finding more.
The 561 'erroneously excluded' ballots:
http://www.metrokc.gov/elections/news/2004_12_13.htm
The 1500 ballots in the Supreme Court case:
http://seattlepi.nwsource.com/local/voters14ww.html
If so, is it safe to assume that Rossi will win by about 250 like he did on the first count?
Thanks
Posted by: scott on December 14, 2004 06:25 PMAll previous explanations boil down to: democrats lie/cheat and republicans do not. We had some brave souls admit it's not about right/wrong, but about political influence--the most honest thing said in this whole process.
Would everyone agree? Let the recount, as provided by law, happen and act upon the result as you see fit (presumably Rossi and gang will sue if Rossi is behind; Gregoire will finally figure out three strikes is too many and fold)...
That is, an unregistered person could fills out a ballot. Then, after the election, comes in and gets their card filled out.
So when was the Councilman's voting card signed? How about the 561 others? Because if the elections department had followed standard procedures competently, all of these would have been entered into the computer, checked, vetted etc all before the ballots were mailed out! (You mean their database doesn't check to see if there's a signature on file first?!?)
So these weren't standard somehow. Handing the card in late (like, say, _today_) would be non-standard.
Posted by: Al on December 14, 2004 06:28 PMAgain, the caution that I have only done a "machine count" and "machine recount" to determine these numbers. If anyone wants to do a "manual recount", be my guest :)
Interesting, the 573 "no signature on file" list does not start until the last name of "Chartes-Maclean". This same voter is number 234 on the larger list of 1554 allegedly mismatched signatures. That means that 233 of the 981 actually mismatched signature voters were before "Chartes-Maclean" alphabetically, or some 23.75% of the larger sample.
I don't know whether 23.75% of the population in general, or more properly, the King County voter population, falls before "Chartes-Maclean" alphabetically. My (expire in) May 2005 Seattle Metro white pages has 1418 pages of residential listings. The name "Charters" (yes, a real person's name -- this portion does not have the business listings) falls on page 205, which happens to be only 14.46% of the way through the telephone book.
So it does look like disparate treatment of voters by thoroughly incompetent King County elections personnel, based upon where someone's name was in the alphabet. My theory on this as a possible explanation:
1. The absentee ballots were initially divided out alphabetically for signature verification, with each employee getting a certain portion of the alphabet to verify against the optical imaging system.
2. Someone in the initial range of A to Ch was a lot more thorough or strict at rejecting names, which explains why the 981 real mismatched signatures come 23.75% from this category, while only about 14.46% of the King County population falls in this category.
3. Other employee(s) were assigned to verify absentee ballots whose signatures were not on the optical imaging system by looking at the original registration records. This was an extremely tedious job, so when someone got to the "Ch" range, they simply gave all and decided to arbitrarily reject all of the remaining ballots.
There is another explanation as to why the no-signature record ballots start with "Chartes-Maclean", which is more omninous, if not more likely. Please note that number 234 out of a 1554 person list happens to fall 15.06% of the way down the list -- almost exactly the same as the 14.46% suggested by looking at the white pages of the telephone directory.
So it is possible that whoever Dean Logan commissioned to look for the "no signature on file" ballots started at the bottom of the list, and worked his or her way towards the top. For whatever reason, the worker or workers stopped with "Chartes-Maclean" and did not go any further forward in the alphabet.
One way this could happen, without intentional malfeasance, would be if Dean Logan had the list of 1554 voters printed out to check -- let's say this came out to 29 names on a page. The first eight pages of the list, comprising the first 232 mismatched absentee voters, get lost or misplaced somehow. Only the remaining 46 pages of the list (which would be 54 pages in total) are handed out to election workers to check. Therefore, the "no signature on file" list that is compiled by this process does not start until "Chartes-Maclean".
If this is what happened, then they found 573 "no signature on file" ballots out of the last 1321 or so names on the 1554 person list, or 43.48% of that portion of the list.
If they go back to the 1554 person list, and look at the 233 names before "Chartes-Maclean", they should be expected to find another 233 x 43.48%, or roughly 101 additional "no signature on file" ballots.
Posted by: Richard Pope on December 14, 2004 07:54 PMRossi gains another 6, now +70 or +112 overall
Whatcom County +16 and +22
Posted by: Steve on December 15, 2004 01:42 PMGregoire -16
Rossi -5
or a net +11 for Rossi, now up +81 and +123 overall
Posted by: Steve on December 15, 2004 02:46 PM