Mathematically speaking, how big is the likelihood of the Brexit turning out 50/50 +- 10,000 votest?
0%, because Cameron and Merkel have rigged it.
>>8158171
50%
>>8158184
Not really, because there's still 9% undecided. Not to say they're just not wanting it to be close - The closer the election, the less suspicious people are to be.
>>8158184
Yes but only if the polls (which are useless due to tiny sample set, polling a couple of thousand people is not representative of tens of millions) showed a consistent lead for Leave outside a margin of error. The polls instead have shown that neither side has a clear lead. However, the sudden swing to Remain after Cox has been neutralised back within the margin of error. This momentum may continue.
Good data to look at is markets, particularly GBP/USD - see attached. The huge rally a couple of days ago shows huge confidence in GBP, and a strong GBP indicates likely Remain. Bloomberg this morning said that Leave is 37% likely, and this may be the most reliable figure due the fact that international businesses trading hundreds of millions of GBP per day have significantly more at stake than the general public, at least in the short term.
>>8158171
Assuming everyone is voting randomly? Very, very close to 1. Here, look at the distribution function.
The analytical answer is erf(2 sqrt(5/3)) which is about 0.999739
>>8158200
UK polls are especially shit-tier if you look at their last election.
>>8158231
forgot to mention, this is assuming 30 million voters
>>8158232
They really are. How many individuals have been polled in the last week? 20,000 - 50,000, across multiple polls, at an extreme estimate? Assuming this polls everyone, and turnout is 60%, that represents less than 0.1% of the country. You really need 5% or so to get anything useful beyond margin of error, 2% for a rough idea.
>>8158248
>posting that on sci
polls margin of errors have nothing to do with percentages of the total population lad, only the absolute sample size matters
or maybe you're trolling
>>8158231
Yeah, but I would assume only a small minority is voting randomly. Which makes it a little more complicated I suppose.
>>8158200
>after Cox has been neutralised
As expected of /sci/.