SI
SI
discoversearch

We've detected that you're using an ad content blocking browser plug-in or feature. Ads provide a critical source of revenue to the continued operation of Silicon Investor.  We ask that you disable ad blocking while on Silicon Investor in the best interests of our community.  If you are not using an ad blocker but are still receiving this message, make sure your browser's tracking protection is set to the 'standard' level.
Strategies & Market Trends : VOLTAIRE'S PORCH-MODERATED
 Public ReplyPrvt ReplyMark as Last ReadFilePrevious 10Next 10PreviousNext  
To: abstract who wrote (23781)12/10/2000 10:42:08 AM
From: Jim Willie CB  Read Replies (3) of 65232
 
Paul, I heard an expert on the punchcard machine counter
he said it has an error rate of 15 cards per thousand
he compared to optical scanner rate with pencil marks of 5 cards/1000

(I quoted 2% incorrectly, bent my mind to recall)

so if cards are perfectly punched, then 0.15% accuracy
figure 6 million votes in FLorida
and margin of 600 votes would be 0.01% margin victory
VERY TROUBLING THAT MACHINE ACCURACY IS 15 TIMES VICTORY MARGIN

if we were to begin with premise of 600 TRUE margin
then each machine recount would vary about that margin
figure some random variation from normal bellshape distribution
consider 5 standard deviations offers 99.9999% tolerance
give or take a "9"
(I never was a 6-Sigma advocate)

now figure the range requires plus/minus, so divide by 2
0.15% accuracy translates to 8000 votes
I get machines count with stand dev of about 8000 x 1/5 x 1/2 = 800 votes

remember the first count showed Bush 1600 ahead
and the second count showed Bush 800 ahead
absentees and other amended results took it to 600 vote margin
let's give benefit of doubt to second 600 victory margin

it is probably closer to 1000 (given 1600 first count)

now each machine recount would have avg change of 600 votes
difference between first & second was 800 votes
Gore might actually be ahead on 10% or more of the recounts

the normal law provides for 2/3 of samples to lie within one standard deviation from the true mean
so 1/6 of samples would come in less than mean minus one stdev

enough speculating
funny how the media doesnt discuss how successive machine recounts would differ slightly
/ jim
Report TOU ViolationShare This Post
 Public ReplyPrvt ReplyMark as Last ReadFilePrevious 10Next 10PreviousNext  

HomeMailHotSubjectMarksPeopleMarksKeepersSettings
Terms Of UseContact UsCopyright/IP PolicyPrivacy PolicyAbout UsFAQAdvertise on SI
© 2025 Knight Sac Media.  Data provided by Twelve Data, Alpha Vantage, and CityFALCON News