As I wrote on Monday, King County's recent discovery of the 93 previously unaccounted-for absentee ballots proves that the "Mail Ballot Report" that was presented to the canvassing board was a fraud. Today's Seattle Times also reports on this:
an internal report ... said election workers knew in November that they had lost track of 86 ballots — now known to be 93 — even while managers were publicly reporting they had accounted for every single vote. It's unknown if any managers were aware of the discrepancy.In fact, I suspect the latest internal report is also covering up the true magnitude of the discrepancy, which I now estimate to be 881 more ballots counted than valid ballots returned. It's hard to imagine that election workers didn't know about this discrepancy back in November. The county needs to release the full set of Absentee Ballot Audit Trail files immediately so we can find out what's really going on. Posted by Stefan Sharkansky at April 06, 2005 10:04 AM | Email ThisThe discrepancy wasn't reported to the King County Canvassing Board, which is responsible for certifying election results. The official Mail Ballot Summary Report to the board showed no discrepancy.
The internal report, written by assistant elections superintendent Garth Fell, said workers in November found an unexplained shortage of 48 ballots in one batch of absentee ballots, 27 in another and four in a third before election results were certified. Smaller discrepancies were found in other batches.
The bigger problem -- which affected the outcome of the election -- was the inclusion of hundreds of absentee ballots in the vote tabulation which didn't belong.
Logan's gang knew of the recorded discrepancies showing several batches of ballots were short, and I also believe they knew that they were counting the votes on hundreds of ballots in excess of the number they had accepted as valid.
They claim the list of voters credited with participating in the election isn't an indication of a problem, but that list was created from the database containing the record of each voter credited with voting during the canvassing of the absentee ballots.
The list contained too few names to account for all the absentee ballots, because there weren't that many voters whose absentee ballots were accepted during the canvassing process.
Aside from the computerized voter registration records containing those entries -- made during the canvassing process -- crediting absentee voters with voting, is there any other record of the number and identity of voters whose absentee ballots were accepted?
Isn't there a way to query that computerized record to determine the number of ballots accepted as valid?
That number of ballots accepted as valid isn't on the "Mail Ballot Report" for a reason -- its presence on the report would have revealed that they counted the votes on too many ballots, not too few.
The Mail Ballot Report was written in a way that made it appear the number was being reported, but note that the report combines "accepted as valid" and "counted" into one line.
It makes it look as though the number of valid absentee ballots equalled the number of counted ballots -- which we know isn't true.
How can we find that number now?
Posted by: Micajah on April 6, 2005 10:55 AMSo they counted some twice to make up for the 86 or 93 that they didn't count. Did they count 86 votes for Gregoire twice and misplace 86 Rossi Votes? If so that is a difference of 172 that is 43 more then 129. How much was Rossi ahead after the second recount? Wasn't it close to that number of 43? Interesting, very interesting.
Posted by: Kirk on April 6, 2005 11:11 AMIf we get the actual public record request of the Excel file that generated this report, I think that the number in Item 2 -- Total number of ballots returned -- will be generated by the addition of all of numbers in Item 3 -- Total number accepted as valid and counted -- and Item 4 -- Total number of ballots rejected.
The ballot numbers in Item 4 for rejected ballots would come from counting the rejected ballot envelopes in each category. King County Elections puts rejected ballots -- such as late postmark, no signature, mismatched signature, etc. -- into separate storage trays or boxes. So those would simply be counted and input that way.
The ballot numbers in Item 3 for counted ballots actually come from the tabulation statistics in the November 17, 2004 initial certification. That way, those two numbers match exactly.
In reality, King County Elections should have either (a) counted the total number of voters credited as casting mail ballots, including unregistered FWAB's and confidential addresses, or (b) counted up the total number of outer ballot envelopes that were accepted.
Of course, the number of rejected inner envelopes (i.e. multiple ballots, wrong election ballots, etc.) would have to be subtracted from this figire to come up with Item 3 numbers.
As for Item 2, King County Elections maintained an actual count of the total number of mail ballots returned, and published this on its website every single day until about one or two days before the November 17, 2004 certification.
However, this actual running total number only got only to 566,291 as of November 15, 2004 -- a date on which 396 absentee ballots were received.
http://www.metrokc.gov/elections/2004nov/ballotsreceived.htm
The November 17, 2004 "Mail Ballot Report", however, alleges a total number of 568,333 absentee ballots returned -- an increase of 2,042 in just two days, when the trend was for only about 400 ballots a day to be received.
Like I say, the "Total number of ballots returned" in Item 2 of the November 17, 2004 "Mail Ballot Report" is not an actual count, but simply an addition formula of all the categories in Items 3 and 4.
The November 17, 2004 "Mail Ballot Report" is also proven to be total B.S. and a fraud by the extra absentee ballots counted in the November 24, 2004 machine recount -- an increase of 198 -- and the December 23, 2004 manual recount -- an increase of 28. If the Item 2 ballots returned number was genuine on November 17, 2004, then the total number of ballots counted could not have increased by 198 and then another 28 in the later counts.
(By the way, the 566 "Larry Phillips" ballots would not have thrown the report out of balance. They would have been taken from the 1,561 ballots listed as "Not voter's signature" in Item 4.c and added to the total number counted in Item 3.)
Since the November 17, 2004 "Mail Ballot Report" is a total fabrication, I think this is the reason why Dean Logan did not prepare updated reports reflecting the November 24, 2004 machine recount and December 23, 2004 manual recount. Some other numbers -- most likely the total number of ballots received (!) -- would have needing changing to making it perfectly balanced, making teh fraud obvious.
How would counting the ballot envelopes give us an accurate accounting?
If they included among the "accepted" ballots and envelopes several hundred "rejected" ballots and envelopes, then counting the envelopes in each category wouldn't reveal how many illegitimate ballots were included in the vote tabulation.
Recounting the envelopes would simply tell us how many ended up in each category -- whether they belonged in the category where they are now stored or not. Wouldn't it?
Posted by: Micajah on April 6, 2005 11:27 AMI don't know if there's a way to figure out why Rossi and Bennett "lost" votes in King County during the manual recount. In their Dec. 22 report, King County reported that Rossi's total was 12 votes less than what he had from the first recount, and that Bennett's total was 10 votes less. ("undervotes" were reported as having been reduced by 45.)
The inclusion of the 566 ballots counted on Dec. 23 made those negative numbers for Rossi and Bennett disappear from the final report -- and perhaps from the memories of most people.
Oddly, since the 566 ballots would have had no effect on "undervotes," the number of undervotes changed to minus 10 in King County's final report on Dec. 23. It had been minus 45 the day before, so how did it change to only minus 10?
It seems that plenty of questions arise from a close examination of the numbers reported by King County, but how to answer those questions is an enigma.
Granting that the rejected ballots were supposed to be stored separate from the envelopes of accepted ballots, my suspicion is that the 800-900 extra absentee ballots in the vote count were "rejected" ballots that weren't pulled out of the batches -- but were instead sent along to the vote tabulation.
If my suspicion is correct, then counting the envelopes in the categories in which they are now stored won't disclose anything.
You have to know how many envelopes belong in each category, then compare those numbers to the numbers you find when you count the envelopes.
My question is: How can we tell how many belong in each category?
Posted by: Micajah on April 6, 2005 11:48 AMThat certainly is a possible explanation of why there are 800 to 900 extra absentee ballots above the number of voters credited with casting them -- rejected envelopes were instead opened and counted.
I do believe that all of the rejected ballot categories in Item 4 of the "Mail Ballot Report" represent actual counts of rejected ballot envelopes in each category that have been stored separately.
That said, it still doesn't explain why the numbers don't add up, and why they are not being otherwise reported consistently.
For example, let's say that 881 ballots postmarked after 11/02/2004 were properly not credited to the voter, but were mistakenly opened and counted anyway. That would explain the discrepancy in voter crediting and ballots counted.
However, this would make Item 4(a) too small by 881 and Item 3 too big by 881. The overall numbers would still add up to be the same.
This doesn't explain why the total number of ballots alegedly received as of 11/17/2004 is higher by 2,042 than the total number actually received as of 11/15/2004 -- when ballots were coming in at less than 400 per day.
Nor does it explain the additional 198 absentee ballots that materialized out of nowhere to be counted in the 11/24/2004 machine recount or the 28 additional ones counted in the 12/23/2004 manual recount -- much less the 93 additional ballots discovered recently that weren't counted.
Posted by: Richard Pope on April 6, 2005 12:22 PMThat's what we need here.
And for those wondering about the _loss_ of voters and the loss of 'undervotes' think MARK ADDED TO PAPER.
The choices:
- ballot got lost
- ballot changed category
The easiest way to change category would be to turn it into an overvote by marking a second vote (for anyone, Bennett's fine). We already knew about the reduction of undervotes by marking - that's what the canvasing board was doing. A vote for 'Christine Rossi' would be an undervote in the machine (re)count.
I'm very curious about how many of the overvotes could be shown to be made with two different pens.
Posted by: Al on April 6, 2005 01:11 PMDid you read Fund's book? He described how overvotes are easily made on punchcard ballots. Take a stack of and push a nail through the "Gore" (or "Gregoire") hole. Any ballots already punched for Gore wouldn't be changed; any ballots marked for Bush or Buchanan would show as an overvote and be tossed. Now THAT's stealing votes!
Using a marker on individual ballots is also pretty easy, but not as easy.
Posted by: Shannon K on April 6, 2005 01:20 PMThey were comparing their ballot count to the previous recount -- so if 35 of the newly counted 566 ballots were "undervotes" then there would be only 10 fewer "undervotes" compared to the previous recount.
I had thought of the 566 newly added ballots (previously rejected as "no signature on file") as a new group of ballots which couldn't reduce the discrepancy between the two recounts.
They, of course, reduce discrepancies any which way they can. So, if there really were 35 new "undervotes" among those 566 new ballots, they simply used those new "undervotes" to cancel out most of the discrepancy.
Posted by: Micajah on April 6, 2005 05:02 PMhttp://www.vatican.va/holy_father/john_paul_ii/apost_constitutions/documents/hf_jp-ii_apc_22021996_universi-dominici-gregis_en.html
"68. After all the ballots of the Cardinal electors have been placed in the receptacle, the first Scrutineer shakes it several times in order to mix them, and immediately afterwards the last Scrutineer proceeds to count them, picking them out of the urn in full view and placing them in another empty receptacle previously prepared for this purpose. If the number of ballots does not correspond to the number of electors, the ballots must all be burned and a second vote taken at once; if however their number does correspond to the number of electors, the opening of the ballots then takes place in the following manner."
So if the number of ballots doesn't match the number of electors, they are discarded entirely (by burning, presumably in the usual manner of black smoke) and not even counted. It doesn't matter whether the discrepancy in numbers would have affected the result of the particular round of balloting.
Dean Logan is presently on vacation. Hopefully, he is not visiting Rome. Otherwise, we will never be able to elect a new pope ...
Posted by: Richard Pope on April 6, 2005 05:27 PMI was just double-checking to see if the "undervotes" discrepancy of minus 45 really was reduced to minus 10 by additional undervotes among the 566 ballots counted on Dec. 23. (It was, according to my comparison of the daily reports posted on the web for Dec. 22 and 23.)
And, I found an old note I had made on Dec. 22 when I copied and saved that daily report (and forgotten -- must be my extreme youth).
On Dec. 22, before those 566 new ballots were added to the count, Gregoire had gained 47 votes compared to the machine recount.
Rossi and Bennett had lost 12 and 10 respectively.
Here's the interesting part: The vote totals for Bennett, Gregoire and Rossi in King County on Dec. 22 equalled 875,924; while the machine recount had totaled 875,899 for those three candidates.
That's a difference of 25 more votes counted in the manual recount compared to the machine recount. And, the only candidate who gained votes in the manual recount of all but the added 566 ballots was Gregoire, who gained 47. She seems to have received the votes on all the extra ballots counted in the manual recount.
You've noted earlier that there were 28 more absentee ballots counted during the manual recount compared to the machine recount.
28 and 25, hmmmm. Coincidence? Just the way things worked out as the numbers bounced around?
Prior to the addition of the 566 Larry Phillips ballots, Gregoire picked up exactly 59 net new votes out of King County: Gregoire + 47, Rossi - 12.
And there were exactly 59 new ballots added to the machine recount before the 566 Larry Phillips ballots: 28 absentees and 31 poll/provisionals?
Coincidence or not? Go figure ...
If we sent Dean Logan to Rome, we could even end up with an American woman as the next pope ...
Posted by: Richard Pope on April 6, 2005 07:12 PM