Comments on FiveThirtyEight: Politics Done Right: Minnesota Challenged Ballot Primer

酒店經紀人, 菲梵酒店經紀, 酒店經紀, 禮服酒店上班, 酒店小姐兼職, 便服酒店經紀, ...

2009-10-02T05:15:17.423-04:00

酒店經紀人,
菲梵酒店經紀,
酒店經紀,
禮服酒店上班,
酒店小姐兼職,
便服酒店經紀,
酒店打工經紀,
制服酒店工作,
專業酒店經紀,
合法酒店經紀,
酒店暑假打工,
酒店寒假打工,
酒店經紀人,
菲梵酒店經紀,
酒店經紀,
禮服酒店上班,
酒店經紀人,
菲梵酒店經紀,
酒店經紀,
禮服酒店上班,
酒店小姐兼職,
便服酒店工作,
酒店打工經紀,
制服酒店經紀,
專業酒店經紀,
合法酒店經紀,
酒店暑假打工,
酒店寒假打工,
酒店經紀人,
菲梵酒店經紀,
酒店經紀,
禮服酒店上班,
酒店小姐兼職,
便服酒店工作,
酒店打工經紀,
制服酒店經紀,
酒店經紀,
菲
梵,

外遇外遇外遇外遇外遇外遇外遇外遇外遇外遇外遇外遇外遇外遇外遇外遇外遇外遇 ...

2009-04-16T04:16:00.000-04:00

^^ very nice 徵信, 徵信社, 徵信, 徵信社, 徵信, 徵信社,徵信, 徵信社, 徵...

2009-03-12T22:20:00.001-04:00

^^ nice blog!! ^@^徵信, 徵信網, 徵信社, 徵信社, 徵信社, 徵...

2009-03-12T22:20:00.000-04:00

情趣用品,情趣用品,情趣用品,情趣用品,情趣用品,情趣用品,情趣用品,情趣用品,情趣用品,情趣用品,...

2009-02-08T21:13:00.000-05:00

Thanks to those who responded to my comment above ...

2008-11-27T20:50:00.000-05:00

Thanks to those who responded to my comment above about Barkley's non-representation at the recount.

The gist of some of the responses (as I interpret them) is that, of course, a candidate with no possibility of winning would not have a right to be present.

As a student of comparative elections, what strikes me is how undemocratic that is. The 'party' in any election dispute should be the voters, not (some of the) candidates.

In this race, a majority of voters will have voted for someone other than the eventually declared winner, and a substantial minority will have voted for a candidate other than the two whose vote totals are currently too close to call.

The best way to represent "the voters" is not to have only agents of the top two present. This point would hardly seem odd to almost anyone from outside the US.

(Thanks to Alan, a frequent commenter at my place, for having often pushed so forcefully the point I am echoing here.)

Tejas--Signing a ballot invalidates it because the...

2008-11-26T14:19:00.000-05:00

Tejas--
Signing a ballot invalidates it because the vote is no longer secret: other people can identify which ballot is yours. If they can identify your ballot, they can punish or reward you for voting a certain way.
Making identifying marks on a ballot illegal is meant to discourage both pressure on a voter and bribery of a voter for a certain vote.

Note that no one really cares if you can identify the ballot yourself, only if others can.

@ Dave C.:For type 3, I agree that the 13% I assig...

2008-11-26T00:19:00.000-05:00

@ Dave C.:

For type 3, I agree that the 13% I assigned for Franken are more than "slightly larger" than the 7% I put for Coleman. This can be adjusted, for sure. Which breakdown would you suggest? (note that Coleman's and Franken's totals should add up to 100% so there isn't that much margin to play with).
Also, note that the type 3 challenges only resulted to a handful of ballots, so this is not the most crucial parameter.

But it's definitely a good idea to evaluate the error bars: I will play with that tomorrow. In the meantime, I'm open to any adjustment of the parameters: for the challenge type breakdown, what about
Coleman: 81%, 5%, 8%, 6%
Franken: 74%, 10%, 12%, 4%
(it's not easy to make the type 3 challenges only slightly different from one another while keeping the type 1 challenges not too similar)

What about the chances that each challenge type is accepted? (I used so far 2%, 80%, 10%, 10% respectively).

That said, using the current numbers, the latest update is (with now nc=1,853 Coleman challenges and nf=1,773 Franken challenges):

Each candidates gains that number of ballots from each challenge type:
.. Coleman ... Franken
1-.. 1268 .... 1489
2-.... 74 ..... 142
3-.... 13 ..... 23
4-.... 64 .... 100
--------------------
Total 1419... 1754

Net gain for Franken from recount: 335 ballots!!
With the official Coleman lead at 231, this suggests that Franken may in fact be ahead by over a hundred votes if the challenges are resolved like that.
[details in previous comment, 2.40pm, above]

Could someone please explain why signing a ballot ...

2008-11-25T22:58:00.000-05:00

Could someone please explain why signing a ballot invalidates it? I don't quite understand the reasoning for why it should disqualify a person from voting if they leave a distinguishing mark like a signature. Thanks.

Oh duh, never mind shock about the global 3400---t...

2008-11-25T22:52:00.000-05:00

Oh duh, never mind shock about the global 3400---that probably reflects the challenges, to a degree that depends on how many are "Types 1 & 4". But that still leaves several hundreds left over, and maybe that's what the "missing ballots" assertion is talking about?

In reply to MNLatteLiberal:What I originally meant...

2008-11-25T22:45:00.000-05:00

In reply to MNLatteLiberal:

What I originally meant was that in the STrib's county table, Washington showed more total votes lost (-122 now) than total challenges (110). Of course the dropoff could come from machine pickups of non-legal votes, but (a) when you factor in that not all challenges deduct votes, you get about 40 such (with NO new votes) which strikes me as high, and (b) I subsequently read about assertions of actual "missing votes" in that county.

Now when I go to the by-precinct data, I should compare the sum of columns G,H,I (recounted totals for Coleman, Franken, and Other) with column M for "total vote(r)s in precinct", right? Over the Washington rows 2852--2941, I get a net minus of about 80. Now to my shock when I do that for all precincts, I see a difference over 3400! Am I doing something wrong, or misinterpreting what these columns mean? Replying quickly---KWR

"Mrs B said "...Bob X, whom I have no reason to do...

2008-11-25T21:30:00.000-05:00

"Mrs B said "...Bob X, whom I have no reason to doubt, says there is no legal right for Barkley to have his vote properly recorded.
IS THAT REALLY TRUE?
You mean in a properly run election in the US only SOME of the candidates get to have their votes counted properly?"
Well, technically, NONE of the candidates has a right to get the number figured out precisely. The law is interested in one question and one question only: which candidate won. There is no legal interest in knowing the order or margins by which the trailing candidates failed to win. It is all reduced to the binary, "1" (you get the office) or "0" (you don't).

@KWRegan,I was a Franken observer during the Washi...

2008-11-25T21:11:00.000-05:00

@KWRegan,
I was a Franken observer during the Washington County recount, and as such I would greatly appreciate any detail/insight you might have in regards to your statement above.

On my watch we "discovered" 27 ballots missing in two of Woodbury's precincts. Day before that in Stillwater, afair, 8 ballots were missing.

Do you have more on that story?
Please share.

wv - unces: _ittoheads.
unces see a conspiracy behind Franken closing the gap.

@KWRegan-I agree that the latest Franken numbers, ...

2008-11-25T20:46:00.000-05:00

@KWRegan-

I agree that the latest Franken numbers, which seem honest, are the best base to extrapolate from, because only a small extrapolation is needed. But be sure to set aside the 'Eagle' machines before doing the extrapolation.

It turns out that I don't need to revise my estimates based on the hand-count review, because as I posted earlier they are in "unreasonably good agreement" with the data released by Franken.

Ad CH Truth: Agreed. Regarding the absentees, whe...

2008-11-25T20:37:00.000-05:00

Ad CH Truth: Agreed. Regarding the absentees, when Franken believed on Day 2 that he was down by 700-odd, of course he had to pull every straw. Having raised that issue, he can't go back even if he wanted to.

Ad Dave C. and Pat & Flo, and agreeing with Michael BW: Franken's spokesman was specific that the estimate of 84-behind "is based on how election judges in the counties have ruled so far on challenged ballots." I.e. it is the gap resulting if Manskey's dictum that his board will win all challenges comes true. The upshot is exactly that yours and all projections, even Nate's, should be revised to reflect this number. Indeed, helping this site may even have been a secondary motive for being so specific, and I see no reason for any spin here (indeed, has the Coleman camp contradicted it?---I see no sign).

The Franken camp may believe that a fair portion of what Nate classifies as F-2 (LX) and F-3 (XX) challenges may go their way: if they total 10-15% of Franken's 1500 challenges, and average 20% success rate, that's 30-45 more votes. Maybe a net of 20 over Coleman's similar cases. That could make this project to be a cliffhanger. Since they have this in reserve, and didn't show their whole hand, again I see no reason to doubt the "84". Hence IMHO we should all plug it in, and re-estimate the %s of types of challenges accordingly. I think doing so will peg Nate's previous estimate of 87% of Coleman's challenges being in categories 1 & 4 (vote deducted from Franken) as being too high, which I felt at the time.

The other thing that revises the models is this business of "hundreds of `missing' votes". I guess we can get an exact count from the SoS data (comma-separated values file importable into Excel etc.), but even in some whole-county data one can infer this is going on, such as I noted in Washington county.

Finally, how homogeneous is the 44% sample we have from Minneapolis? I assume that recounting precincts one-at-a-time will make it reflect times of day equally, but how about geography, mindful that incidents like this (saw on theUptake.org) may have caused some stressful balloting situations whose effects haven't been seen yet.

WV "sublogi": a frequent commenter who doesn't have eir own blog :-).

Here's a first-person account of the challenges th...

2008-11-25T20:36:00.000-05:00

Here's a first-person account of the challenges that seems to corroborate Nate's intuition:

http://blog.aclu.org/2008/11/25/challenges-counting-duplicates-and-other-minnesota-recount-complaints/

Jack-be-Numbskullbegs the questions but...WV - 'sc...

2008-11-25T19:41:00.000-05:00

Jack-be-Numbskull

begs the questions but...

WV - 'scror in u'

Here is the video link of Franken conceding the se...

2008-11-25T19:24:00.001-05:00

Here is the video link of Franken conceding the seat:

http://www.youtube.com/watch?v=mPw1f5kZxus

Read it and weep.

@Pat & FloMy only question would be the percentage...

2008-11-25T19:24:00.000-05:00

@Pat & Flo

My only question would be the percentages of each type that you assigned. You put numbers in place of Nate's words, and I think you may be overdoing it.

Example: For a Type 2 vote Nate says "Probably about 5-10% of all challenges. May be slightly more common from the Franken campaign, if Democrats tend to be responsible for a higher fraction of sketchy ballots.". You take this word "slightly more" to be double or a 100% increase. I would read slightly more to be a few percent difference, or at most say a 20% difference. Therefore if Coleman has 5% Type 2 challenges then Franken would have 6%.

These unknowns will case your prediction of 272 votes to have very large error bars. I would suggest you vary those percentages to get an idea of your error bars. I would guess they would be quite high.

-Dave C.

thanks for the lie jack be liar.franken has not co...

2008-11-25T18:54:00.000-05:00

thanks for the lie jack be liar.

franken has not conceded anything

Franken has just conceded the race. See the post ...

2008-11-25T18:38:00.000-05:00

Franken has just conceded the race. See the post on the Trib online.

It seems odd that he would resign before the challenge. Maybe he knows that the challenges are bullshit. Oh well, even libs don't want him in office.

New numbers have been posted: now were at 1,600 Co...

2008-11-25T17:27:00.000-05:00

New numbers have been posted: now were at 1,600 Coleman challenges vs. 1,557 from Franken.

Using Nate's estimate of the number of challenges by type (let's say for Coleman 82%, 5%, 7%, 6% and for Franken 73%, 10%, 13%, 4% respectively) and assuming that type 1 challenges are accepted only 2% of the time, type 2 are 80%, and types 3 and 4 are accepted 10% or the time, we now have
Gain for Coleman: 1,245 votes
Gain for Franken: 1,517 votes
Net gain for Franken from recount: 272 votes.
Since Coleman's official lead is now 217 votes, that means that Franken is actually leading by 272-217=55 votes if the challenges end up being resolved according to these estimates.

(see detailed explanation in the comment above, at 2:40 pm)

If anyone has suggestions concerning the probabilities of each type, it might be interesting to discuss it. (Do you think types 3 and 4 have more than 10% chance to be accepted? less? any other modification?)

I don't understand Nate to be advocating for eithe...

2008-11-25T17:00:00.000-05:00

I don't understand Nate to be advocating for either side, but merely analyzing the counting process and trying to get some clarity out of the fog of frivolous challenges.

I suggest that the number of type 2, 3, and 4 challenges is very small. Take the day 1 results, and extrapolate them to 100% of the vote. That approximates the results of the actual count. Then add the number of Coleman's type 1 challenges to Franken's vote total and the number of Franken's type 1 challenges to Coleman's vote total. I predict that to be the final vote number.

Anybody have access to the number of type 1 challenges from each camp?

by the way - what you call 'plurality', we call 'f...

2008-11-25T16:29:00.000-05:00

by the way - what you call 'plurality', we call 'first past the post' - but you probably knew that.

Thanks for indulging these off topic comments. It is hard to find something new to say about Franken/Coleman figures because we just DON'T KNOW.

We are all finding it hard to wait for the answer - imagine how Franken and Coleman are feeling.

thanks Katharriet.I would imagine that with a very...

2008-11-25T16:26:00.000-05:00

thanks Katharriet.

I would imagine that with a very strong 2 party system, those 2 strong parties wouldn't have much interest in opening up politics to other groupings, which is what would happen if any form of PR was widely introduced. I know vested interests are what is holding back progress here.