Wednesday, November 4, 2009
Election How Many Ballots Are Left to Count?
Posted by Eli Sanders on Wed, Nov 4, 2009 at 5:16 PM
According to King County Elections:
Director Sherril Huff estimates 129,000 mail ballots are on hand and remain to be counted. This number does not include ballots still to be delivered by the United States Postal Service.
Today 53,412 mail-in ballots were counted, bringing the total number of ballots counted thus far to 307,673.
So, calling all Slog math wizards and political psychics: In the McGinn-Mallahan race, in which the candidates are now separated by just 462 votes, what's the most likely outcome? (Factors to consider: Are people who mail their ballots at the last minute more likely to vote for McGinn or Mallahan? What percentage of all outstanding ballots are likely to come from Seattle voters as opposed to greater King County voters? How far is 129,000 from the actual number of ballots remaining to be counted once those still-to-be-delivered ballots arrive via the U.S. Postal Service? What does GOD want?)
Get thee to your calculators and Ouija boards!
Share via
2
Are those votes from particular geographic areas, or are they later arrivals? I dunno. If you don't know that, I don't think your calculator is going to help you. The fact that the second batch narrowed the gap might suggest that the next batch will narrow it further, but probably not.
3
Depends. Will the ballots be driven in a Ford Efficiencies that's moving forward? If so, McGinn is toast.
4
the estimate is based on traditional turnout - and probably doesn't include two LA ballots winging their way to Seattle for McGinn.
Main prob is Joe might pull a Dino on us and we'll have to do the recount.
Main prob is Joe might pull a Dino on us and we'll have to do the recount.
5
@4: "might"? You forget both Dino and Malladroit were mentored by Slade Gorton.
A recount is assured, even if he wins.
A recount is assured, even if he wins.
8
We voted and mailed in our ballots on the very first day -- and I just checked and neither of ours has been tallied yet.
9
Hmmm, its 129,000 mail ballots that remain uncounted.
Has anyone else tracked their ballot's "status" via the county's website: https://info.kingcounty.gov/elections/Ma… ?
My vote doesn't show as being received even though I dropped it off in person at the county bldg on 4th avenue. I know the ballots are still being counted but my primary ballot, also delivered in person, never showed up as being received.
Any idea how reliable King County's ballot tracking mechanism is?
Has anyone else tracked their ballot's "status" via the county's website: https://info.kingcounty.gov/elections/Ma… ?
My vote doesn't show as being received even though I dropped it off in person at the county bldg on 4th avenue. I know the ballots are still being counted but my primary ballot, also delivered in person, never showed up as being received.
Any idea how reliable King County's ballot tracking mechanism is?
11
I put mine in the Lake City neighborhood drop box on Monday and it's not yet showing received. I wish they would say how many of these ballots are still outstanding.
12
My ballot just started showing as received by King County (Track Point #2) tonight, and I put it in a drop box before Halloween. It takes some time.
13
My vote doesn't show as being received and I dropped it off in person at the county bldg on 4th avenue around 7pm last night.
14
Mine and my wife's were dropped in the downtown box last Wednesday, and both of our statuses are at step 5: "signature verified- waiting to be opened". We both voted for McGinn.
15
The bigots are trying to 'pray in a victory'.
What if 'they' lose...did God ignore their prayers?
Assholes.
www.protectmarriagewa.com
What if 'they' lose...did God ignore their prayers?
Assholes.
www.protectmarriagewa.com
16
I voted on Day 2 after receiving my ballot, and while my signature was verified, my ballot hasn't been processed yet. (I voted McGinn.)
First in, last counted?
First in, last counted?
17
Maybe they order these by zip codes, per chance? I have a higher number zip code than most zips in Seattle.
21
"The bigots are trying to 'pray in a victory'. What if 'they' lose...did God ignore their prayers?"
Joke's on them, there is no god. But clinging to their myths gives them comfort as the pray for big government to have more say or people's legal and emotional relationships.
Joke's on them, there is no god. But clinging to their myths gives them comfort as the pray for big government to have more say or people's legal and emotional relationships.
22
The numbers you give Eli are county wide.
Here's what I get for SEATTLE ballots...
Joe won this round of 9,675 mayoral ballots counted (not including write-in) 51%-49%. As near as I can tell...
KC Elections had 136,424 ballots in hand as of 8pm last night.
They have counted 109,874 as of 4:30pm today.
If the rough proportions of ballot deliveries are the same as in the primary, we should see about 67,000 ballots received today added to the Seattle totals at the 8pm tally tonight.
That's roughly 93,550 Seattle ballots left to count.
51% of that is 47,710
49% of that is 45,840
Difference is 1,870 in favor of Mallahan.
If the percentages seen today hold, Mallahan wins. McGinn gained ground in later counts in the primary, so it is not a foregone conclusion the percentages will hold.
Here's what I get for SEATTLE ballots...
Joe won this round of 9,675 mayoral ballots counted (not including write-in) 51%-49%. As near as I can tell...
KC Elections had 136,424 ballots in hand as of 8pm last night.
They have counted 109,874 as of 4:30pm today.
If the rough proportions of ballot deliveries are the same as in the primary, we should see about 67,000 ballots received today added to the Seattle totals at the 8pm tally tonight.
That's roughly 93,550 Seattle ballots left to count.
51% of that is 47,710
49% of that is 45,840
Difference is 1,870 in favor of Mallahan.
If the percentages seen today hold, Mallahan wins. McGinn gained ground in later counts in the primary, so it is not a foregone conclusion the percentages will hold.
23
Look up your vote online and make sure the signature was validated.
If the Powers That Be - the ones that want us to pay $15,000 per voter for the cost overruns for the Billionaires Tunnel - force a recount, make sure they count yours.
Seattle is not in thrall to the "stakeholders" who don't LIVE here or VOTE here - it belongs to US.
If the Powers That Be - the ones that want us to pay $15,000 per voter for the cost overruns for the Billionaires Tunnel - force a recount, make sure they count yours.
Seattle is not in thrall to the "stakeholders" who don't LIVE here or VOTE here - it belongs to US.
24
@16 - you and me both, brother.
Worst case scenario, we call out the turtles and shut this city DOWN.
Worst case scenario, we call out the turtles and shut this city DOWN.
25
Both my wife's ballot and my ballot are at step three despite the fact that both ballots were mailed well over a week ago.
Count two more votes for McGinn.
Also, refused to vote for Suzie Hutchison loving Sheriff Sue Rahr, so I wrote in David Goldstein and my wife, Dan Savage. Does KC Elections print a list of all the write-ins that were received?
Count two more votes for McGinn.
Also, refused to vote for Suzie Hutchison loving Sheriff Sue Rahr, so I wrote in David Goldstein and my wife, Dan Savage. Does KC Elections print a list of all the write-ins that were received?
26
Doesn't it seem odd that even in this little comments thread there are so many votes that were mailed early but haven't yet been tablulated. How can that be? WTF is our newly elected head of Elections, Sherril Huff bringing to the table?
Maybe someone from The Stranger can give a call down there tomorrow and report back???
Maybe someone from The Stranger can give a call down there tomorrow and report back???
27
Either that or get someone to explain how to design database query results that don't create more questions than they answer.
Regardless, when all is said and done they have NO IDEA how many people ended up mailing in their ballots - cause they haven't even RECEIVED them yet.
Regardless, when all is said and done they have NO IDEA how many people ended up mailing in their ballots - cause they haven't even RECEIVED them yet.
28
I don't think they start counting ballots until the actual election day. Hence, the amount of people still showing their vote as uncounted.
30
Me & my s/o's two McGinn votes landed in the Union Station drop box yesterday evening but are not yet showing as received by the county. Not a big surprise.
I expect McGinn has this one. For some reason I intuitively view last-minute slacker voters without landlines like myself as the core of his constituency.
I expect McGinn has this one. For some reason I intuitively view last-minute slacker voters without landlines like myself as the core of his constituency.
31
wow, yeah. i mailed mine in over a week ago and it's been received and validated, but not tallied yet. one more for McGinn.
sincerely,
diggum
sincerely,
diggum
32
People, PEOPLE:
your vote was received, and validated, and "prepared for counting", but that's as far as you're ever going to find out. They CAN'T tell you when your ballot is counted, because they don't know. That's the whole point of stripping the ballots out of the outer envelopes -- the only one with your name on it.
your vote was received, and validated, and "prepared for counting", but that's as far as you're ever going to find out. They CAN'T tell you when your ballot is counted, because they don't know. That's the whole point of stripping the ballots out of the outer envelopes -- the only one with your name on it.
33
That makes absolute sense Fnarf and in fact did cross my mind. I think the confusion comes from the language on the KC Elections site:"Your returned ballot packet will soon be opened and your ballot will be prepared for counting." Clearly leads one to believe that the votes are still pending count.
35
"In the McGinn-Mallahan race, in which the candidates are now separated by just 462 votes, what's the most likely outcome? (Factors to consider: Are people who mail their ballots at the last minute more likely to vote for McGinn or Mallahan? What percentage of all outstanding ballots are likely to come from Seattle voters as opposed to greater King County voters? How far is 129,000 from the actual number of ballots remaining to be counted once those still-to-be-delivered ballots arrive via the U.S. Postal Service? What does GOD want?)
Get thee to your calculators and Ouija boards!"
Here's the some answers a la dan savage: Some Freaky Deaky jello sex; put a condom on it and it won't matter what it's made out of or smells like; 673; is this some sort of trick question? see answer 1.
oh, and calculators are the devils tool.
Get thee to your calculators and Ouija boards!"
Here's the some answers a la dan savage: Some Freaky Deaky jello sex; put a condom on it and it won't matter what it's made out of or smells like; 673; is this some sort of trick question? see answer 1.
oh, and calculators are the devils tool.
36
I too waited until the absolute last minute to drop my ballot in the mail, so that's another McGinn vote that won't be counted until the very end.
37
Oh, a real math problem on SLOG. Step aside, man, I'm a scientist...
Assume n=129,000 ballots are drawn distributed with p=0.500571 in McGinn's favor. The mean outcome is that 64,597 (np) of those votes go to McGinn, with a standard deviation of 180 votes (sqrt n p (1-p)). To loose his current 462 vote lead, McGinn would have to get less than 63,769 votes ((129,000-462)/2). That is 828 votes less than the mean (64,597 - 63,769), a 4.6 sigma event (828/180). The chance of a 4.6 sigma deviation to one side is less than 1 in 10,000. Under these assumptions, it's basically impossible for Malhallan to win.
The exact numer of ballots remaining is not very important to this conclusion. As long as it isn't off by an order of magnitude, the chance of a Malhallan win is tiny. But the assumption that the vote distribution of the remaining ballots is the same as those of the ballots already counted is very important. If late-breaking ballots have a different distribution, so that the value of p is different, the game can get very different even for very small changes to p. If you know something about late voters, let us know. If they tend to skew hippie, Malhallan is definitely lost. If they tend to skew corporate, Malhallan still has a decent chance.
Assume n=129,000 ballots are drawn distributed with p=0.500571 in McGinn's favor. The mean outcome is that 64,597 (np) of those votes go to McGinn, with a standard deviation of 180 votes (sqrt n p (1-p)). To loose his current 462 vote lead, McGinn would have to get less than 63,769 votes ((129,000-462)/2). That is 828 votes less than the mean (64,597 - 63,769), a 4.6 sigma event (828/180). The chance of a 4.6 sigma deviation to one side is less than 1 in 10,000. Under these assumptions, it's basically impossible for Malhallan to win.
The exact numer of ballots remaining is not very important to this conclusion. As long as it isn't off by an order of magnitude, the chance of a Malhallan win is tiny. But the assumption that the vote distribution of the remaining ballots is the same as those of the ballots already counted is very important. If late-breaking ballots have a different distribution, so that the value of p is different, the game can get very different even for very small changes to p. If you know something about late voters, let us know. If they tend to skew hippie, Malhallan is definitely lost. If they tend to skew corporate, Malhallan still has a decent chance.
39
As the primary count went on the tallies seemed to skew left. After everyone stopped paying attention, more votes showed up for McGinn and anyone but Hutchtison. Also the Bag tax results became a lot closer. The point is the liberals = lazy procrastinators theory is probably true in all mail elections, and McGinn will win.
Personally I always take my ballot to the drop box. Nobody trusts USPS with cash. Why in the hell would you mail your ballot?
Personally I always take my ballot to the drop box. Nobody trusts USPS with cash. Why in the hell would you mail your ballot?
40
The McGinn voters I know are my real-life friends, social network friends, people I know of who are one degree away from someone at The Stranger, of which there are many -- we're not that big a city.
Everyone outside that circle I know voted for Mallahan... most of which kinda surprised me.
(To be honest -- the Seattle mayoral race hasn't concerned me as much as some of the other races or initiatives. Not really biting nails here.)
Everyone outside that circle I know voted for Mallahan... most of which kinda surprised me.
(To be honest -- the Seattle mayoral race hasn't concerned me as much as some of the other races or initiatives. Not really biting nails here.)
42
Well, given the number of outstanding ballots, and the likely distribution of voters, I feel that I can confidently call this one in favor of Thomas Dewey
43
Even as we speak, tunnel boring contractors are digging into election central, to plant bags of "new ballots" for Mallahan...
44
Most Likely outcome?
Another 3 or 4 days of ballot counting. Tight gap. No matter who is on top, the other will demand a recount. Possibly a lawsuit or two to disqualify ballots. Final outcome around mid-December that will be hotly contested by the losing party. Winner won't be able to do anything because they don;t have a real voter mandate. Tunnel goes through as planned either way. Everybody wins and everybody loses.
Another 3 or 4 days of ballot counting. Tight gap. No matter who is on top, the other will demand a recount. Possibly a lawsuit or two to disqualify ballots. Final outcome around mid-December that will be hotly contested by the losing party. Winner won't be able to do anything because they don;t have a real voter mandate. Tunnel goes through as planned either way. Everybody wins and everybody loses.
45
There is only 1 data point at this time.
After 104960 Mayoral votes counted, McGinn leads by 462 votes, or .44%. That's all we know, and all we can know.
We have no way of knowing how the two different ballot drops were separated, we don't know anything about when they came in or from where.
So, if the "trend" continues, McGinn wins. The only trend is the number we see.
After 104960 Mayoral votes counted, McGinn leads by 462 votes, or .44%. That's all we know, and all we can know.
We have no way of knowing how the two different ballot drops were separated, we don't know anything about when they came in or from where.
So, if the "trend" continues, McGinn wins. The only trend is the number we see.
46
@37, I admire the calculations, but the assumptions are wrong. You are basing them on all of King County, not Seattle only. Ballots received today: 27560, total received in Seattle to date:164113, Turnout percentage to date: 43.52% from: http://your.kingcounty.gov/elections/abs…. Also, where are you getting your P value? It doesn't jive with the numbers that we're seeing.
@22, are you also basing your calculations on Seattle-only?
@22, are you also basing your calculations on Seattle-only?
47
I'm a Seattle voter living in Vancouver, and I forgot to even request my ballot until 4:15pm on Election Day. I did get it postmarked in time (up here our post offices are open until 8:00) so that's one more for McGinn, probably arriving in a couple of days.
48
@46: I simply trusted Eli's assertions that the relevant ballot counts were 307,674 done and 129,000 to go. If two candidates are seperated by 462 votes out of 307,674, then the votes favor the leading candidate by p = 154,068 / 307,674 = 0.500751.
It looks like those are not the right numbers for mayoral ballots, in which case the calculation needs revising.
It looks like those are not the right numbers for mayoral ballots, in which case the calculation needs revising.
49
The drivers in this election in Seattle were - Dow - Ref 71 and No on 1033.
Who are those voters in the mayors race? Time alone will tell.
Many people left it blank, did write ins, going to be close and the winner has NO big time mandate. Phone banks discovered voters very unhappy with choices in the mayors race.
Interesting. Burgess and Bagshaw will run the city - with Conlin keeping a wet tongue hanging out amid the lop sided grinning style.
Licata teaming with Clark and Rasmussen will be a big player.
Mc Ginn is really fat - surprise Dan is in bed with him. Old fat guy, kind of a breakthru. The Stranger in league with old fat guys ... hmmmmm.
Still floating above the fray cause Ref: 71 was APPROVED. Wow, "Were here, were Queer and we do a damn fine campaign in Washington State."
(The new mayor, with NO political capital, will be chewed and spit out by the old timers in Olympia who play very hard and dirty. Ed Murray will dictate many an agenda to the City.)
Who are those voters in the mayors race? Time alone will tell.
Many people left it blank, did write ins, going to be close and the winner has NO big time mandate. Phone banks discovered voters very unhappy with choices in the mayors race.
Interesting. Burgess and Bagshaw will run the city - with Conlin keeping a wet tongue hanging out amid the lop sided grinning style.
Licata teaming with Clark and Rasmussen will be a big player.
Mc Ginn is really fat - surprise Dan is in bed with him. Old fat guy, kind of a breakthru. The Stranger in league with old fat guys ... hmmmmm.
Still floating above the fray cause Ref: 71 was APPROVED. Wow, "Were here, were Queer and we do a damn fine campaign in Washington State."
(The new mayor, with NO political capital, will be chewed and spit out by the old timers in Olympia who play very hard and dirty. Ed Murray will dictate many an agenda to the City.)
50
Oh, I forgot. Mc Ginn is going to be kind, and cool, and talkitive and friendly and talk a lot to the Oly folks as Seattle tries to suck up every dime of shrinking state dollars that they can.
And the rest of the state is too dumb to know the game, so says Mc Ginn ... but it sounded good to his followers. Mighty Mike in Olympia doing a teach in ... with his hand out ..
He might find he can't get appointments even.
And do you think Olympia will cater to Seattle in an election year with the newly minted far right trying to get some seats in the outback? Kicking Seattle works better - works very well in fact - in Olympia in election years, and the long session has lots of chances.
January doth approach rapidly.
And the rest of the state is too dumb to know the game, so says Mc Ginn ... but it sounded good to his followers. Mighty Mike in Olympia doing a teach in ... with his hand out ..
He might find he can't get appointments even.
And do you think Olympia will cater to Seattle in an election year with the newly minted far right trying to get some seats in the outback? Kicking Seattle works better - works very well in fact - in Olympia in election years, and the long session has lots of chances.
January doth approach rapidly.
52
As you point out, math goes out the window when we don't know if there's a particular type of vote outstanding, or if that type of vote tilts one way or the other. If it's all Magnolia, for instance, Mallahan's in on a cakewalk; if it's late voters, and late voters are hippies, McGinn will surge. We really don't have a clue.
53
@37 I think Dino Rossi would like to talk to you.
The problem with this kind of analysis is that the assumption that the distribution between the groups will be the same is in no way valid. While we do not know how they are different (otherwise we would know who is going to win), we know they are. Swings of that degree happen all the time in elections (a lot more than 1 in 10,000) and in fact happened between Tuesday and Wednesdays count already.
The problem with this kind of analysis is that the assumption that the distribution between the groups will be the same is in no way valid. While we do not know how they are different (otherwise we would know who is going to win), we know they are. Swings of that degree happen all the time in elections (a lot more than 1 in 10,000) and in fact happened between Tuesday and Wednesdays count already.
54
#37 said: "Assume n=129,000 ballots are drawn distributed with p=0.500571 in McGinn's favor. The mean outcome is that 64,597 (np) of those votes go to McGinn, with a standard deviation of 180 votes (sqrt n p (1-p)). To loose his current 462 vote lead, McGinn would have to get less than 63,769 votes ((129,000-462)/2). That is 828 votes less than the mean (64,597 - 63,769), a 4.6 sigma event (828/180). The chance of a 4.6 sigma deviation to one side is less than 1 in 10,000. Under these assumptions, it's basically impossible for Malhallan to win."
I hope McGinn wins, but seriously that comment is totally wrong. Using that logic Wednesday morning would have made yesterday's 4:30 pm results "basically impossible." We don't know enough yet (about who voted when or what order votes are counted) to predict anything besides the fact that it is a close race - as yesterday's drop confirms. But we can still hope for the best.
I hope McGinn wins, but seriously that comment is totally wrong. Using that logic Wednesday morning would have made yesterday's 4:30 pm results "basically impossible." We don't know enough yet (about who voted when or what order votes are counted) to predict anything besides the fact that it is a close race - as yesterday's drop confirms. But we can still hope for the best.
56
The fact that the gap changed fairly significantly (in relative terms) just between Tuesday and Wednesday suggests any assumption that the count will follow past trends is fallacious.
57
That is why my FB status today is my thanks to Joe for him conceding.
Statistically, though, this presumes many - mostly incorrect - things.
It presumes:
a. population of voters in votes to date matches people who voted at the last minute (demonstrably false);
b. age/ethnic/wealth/sex groupings for the first group of voters matches the people who voted at the last minute (demonstrably false);
c. that we actually have comparable data from a similar election period - which doesn't even pass the sniff test;
d. that the trends in the batches match the theory - which they don't.
Look, I know Mike McGinn is our new mayor, but anyone who tells you they can prove it with stats is bullshitting you.
Statistically, though, this presumes many - mostly incorrect - things.
It presumes:
a. population of voters in votes to date matches people who voted at the last minute (demonstrably false);
b. age/ethnic/wealth/sex groupings for the first group of voters matches the people who voted at the last minute (demonstrably false);
c. that we actually have comparable data from a similar election period - which doesn't even pass the sniff test;
d. that the trends in the batches match the theory - which they don't.
Look, I know Mike McGinn is our new mayor, but anyone who tells you they can prove it with stats is bullshitting you.
58
@2 - thanks to Fnarf for remembering the geographic distribution comparison complication as well. We don't know enough about the groups to be able to say ANYTHING.
I know it's frustrating, but we'll just have to wait.
Caveat - it was obvious Susan had lost and her chance of regaining a win was beyond probability - which is why she conceded.
I know it's frustrating, but we'll just have to wait.
Caveat - it was obvious Susan had lost and her chance of regaining a win was beyond probability - which is why she conceded.
59
@53: Agree entirely! There's just too much goddamn noise between one batch and another. You can't pull any trend at all from the first two day's return. They're going to be different groups of voters and we don't know how they differ. There's a really strong argument for late voters going either way - the conventional wisdom of early voters are older and more conservative (Mallahands!) would give things to BikeSnobSEA, but that could be countered by a number of hypotheses, such as: conventional wisdom doesn't hold for an off-year municipal election because all voters are more motivated to turn out than the norm (norm being even-year, 'big' election voters) and so younger, liberal voters are as likely to return their ballots early or the particular last minute moves of either and both campaigns screw up CW: McGinn's "flip-flop" on the tunnel and the subsequent attack ads of Mallahan wins late voters for Mallahan, Mallahan's last-minute attacks backfire, McGinn's better ground-game turns out more voters...
etc., etc., etc.
The only thing the politically minded thing a person can do right now? Drink away the part of the day that you cannot sleep away. Until 4:30pm.
etc., etc., etc.
The only thing the politically minded thing a person can do right now? Drink away the part of the day that you cannot sleep away. Until 4:30pm.
60
Hey so I realize most of this comment string has nothing to do with the original math, but there's a pretty big error (I think). The original commenter correctly said Mcginn needs to get (129000-462)/2 votes or less to lose his lead. That number is 64269, NOT 63769 as the original author posted. The likelihood of that is not a 4.6 sigma deviaion, but a 1.8, and the chance of that is not less than 1 in 10000, but more like 3.5/100, or 3.5%
That said, the late returns are likely to have a stronger bias to McGinn, so p will increase, and Mallahan's chance of recovering will decrease.
That said, the late returns are likely to have a stronger bias to McGinn, so p will increase, and Mallahan's chance of recovering will decrease.
61
What is the link to look up your vote? I keep seeing this mentioned for "stages" but no one links it. Could Slog please link that in a proper post?
62
On most of our close races going back to Cantwell's first win against Skeltor the pattern is they toggle back and forth with each count and then at the very end they skew to the left, probably because of the fact that young , not republican people are more likely to be procrastinators. THe early vote counts are random and have no reason to be atypical. The only change in this pattern could be the lack of walk in voting.
63
https://info.kingcounty.gov/elections/Ma…
Also check out...
http://wei.secstate.wa.gov/OSOS/VoterVau…
Technically not related but have you seen this?
http://www.bikeexif.com/
Also check out...
http://wei.secstate.wa.gov/OSOS/VoterVau…
Technically not related but have you seen this?
http://www.bikeexif.com/
64
Yes, the most plausible theory is that the young are the procrastinators, and that favors McGinn in uncounted ballots. But even that theory is weak -- I'm nearly as ready to believe that there is no bias between early and late voters. I can see a lot of set-in-their-ways older people (Mallahan voters) who don't like all-mail elections, and who wait to mail their ballot until election day just to maintain the tradition. It's a toss up.
65
I second all the criticism about invoking normality for the election results for mayor. @37, the assumption only holds if the variate is _identically_ and independently distributed. There's not only a lack of evidence for this assumption, there's evidence against it. Just look at how much p differs for the ballots stratified by day counted. I'm not quite interested enough to actually calculate p stratified by day counted (and it's hard to find the results by day posted--the king county website only has the most recent results) but http://slog.thestranger.com/slog/archive… has some info. It doesn't take very much non-stationarity in p to totally overwhelm the effect of sampling error under normality in this case.
66
Hey everybody calling out my assuption of uniform sampling: I agree completely! In fact, I said as much in my original post: "If late-breaking ballots have a different distribution, so that the value of p is different, the game can get very different even for very small changes to p". I calculated what I could, with the given data. There was no data given about how the ballots skew with time.
(Actually, I didn't even manage that. @60 found an arithmetic error in my analysis and corrected it. Ah, the dangers of public calculation. Thanks, @60.)
(Actually, I didn't even manage that. @60 found an arithmetic error in my analysis and corrected it. Ah, the dangers of public calculation. Thanks, @60.)
67
The difference in vote totals is less than 4/10's of one percentage point of the remaining ballots. In statistical terminology there is a name for this type of race: Toss Up! There are far to many variables to figure something this close.
Most Commented on Slog
-
Slog Somehow Overlooked... by Dan Savage
-
VivaVulva Viagra! by Dan Savage -
Life and Death in a Recliner in South Carolina by Dan Savage
-
Health at Every Size by Dan Savage
-
"Let his days be few... Let his children be fatherless and his wife a widow." by Dan Savage
-
Feeling Sorry for Carrie Prejean by Dan Savage
-
Last Night's Colbert Report by Dan Savage
-
$2,600,000,000 by Dan Savage
-
All The Best Men Are Taken by Dan Savage
-
I Hate You, Garrison Keillor by Jen Graves
RSS
Facebook
Twitter
Comments (68) RSS