Lies, Damned Lies and Meaningless Drivel

Talk about anything but keep it polite and reasonably clean.
Post Reply
User avatar
djewesbury
Graham’s 1970
Posts: 8165
Joined: 20:01 Mon 31 Dec 2012
Location: Gothenburg, Sweden
Contact:

Lies, Damned Lies and Meaningless Drivel

Post by djewesbury »

Some posts moved here from The Scottish Independence Referendum, 18th Sept 2014 by jdaw1.

jdaw1 wrote:
DRT wrote:That is quite a margin in a two horse race.
Innumerate rubbish.

Odds of 1/4 and 3/1 imply probabilities of 80% and 25% (the total exceeding 100% by the bookies’ profit). So divide by 1.05 to get implied probabilities of 76% and 24%.

Let’s assume that the Yes and No votes are normally distributed, with a standard deviation of ±5% (this σ having been guessed by me). So if the expected outcome were 53.6% : 46.4%, with that uncertainty, the probability of a Yes would indeed be about 76%. But 53.6% : 46.4% is quite a small margin.
Oh, it's suddenly gone very quiet.. Anybody out there…?
Daniel J.
Husband of a relentless former Soviet Chess Master.
delete.. delete.. *sigh*.. delete...
User avatar
DRT
Fonseca 1966
Posts: 15779
Joined: 23:51 Wed 20 Jun 2007
Location: Chesterfield, UK
Contact:

Re: The Scottish Independence Referendum, 18th Sept 2014

Post by DRT »

djewesbury wrote:
jdaw1 wrote:
DRT wrote:That is quite a margin in a two horse race.
Innumerate rubbish.

Odds of 1/4 and 3/1 imply probabilities of 80% and 25% (the total exceeding 100% by the bookies’ profit). So divide by 1.05 to get implied probabilities of 76% and 24%.

Let’s assume that the Yes and No votes are normally distributed, with a standard deviation of ±5% (this σ having been guessed by me). So if the expected outcome were 53.6% : 46.4%, with that uncertainty, the probability of a Yes would indeed be about 76%. But 53.6% : 46.4% is quite a small margin.
Oh, it's suddenly gone very quiet.. Anybody out there…?
please define standardly distributed.
"The first duty of Port is to be red"
Ernest H. Cockburn
User avatar
jdaw1
Cockburn 1851
Posts: 23613
Joined: 15:03 Thu 21 Jun 2007
Location: London
Contact:

Re: The Scottish Independence Referendum, 18th Sept 2014

Post by jdaw1 »

DRT wrote:please define standardly distributed.
Normally distributed. It’s not, of course, very little is, but the approximation isn’t going to mislead.

If you want to attack, attack the ±5%.
PhilW
Dalva Golden White Colheita 1952
Posts: 3503
Joined: 14:22 Wed 15 Dec 2010
Location: Near Cambridge, UK

Re: The Scottish Independence Referendum, 18th Sept 2014

Post by PhilW »

DRT wrote:please define standardlynormally distributed.
It's exponential.
User avatar
jdaw1
Cockburn 1851
Posts: 23613
Joined: 15:03 Thu 21 Jun 2007
Location: London
Contact:

Re: The Scottish Independence Referendum, 18th Sept 2014

Post by jdaw1 »

PhilW wrote:
DRT wrote:please define standardlynormally distributed.
It's exponential.
For those new to this rant, start here.
PhilW
Dalva Golden White Colheita 1952
Posts: 3503
Joined: 14:22 Wed 15 Dec 2010
Location: Near Cambridge, UK

Re: The Scottish Independence Referendum, 18th Sept 2014

Post by PhilW »

jdaw1 wrote:Odds of 1/4 and 3/1 imply probabilities of 80% and 25% (the total exceeding 100% by the bookies’ profit). So divide by 1.05 to get implied probabilities of 76% and 24%.
Agreed to there; and that is the margin which surprises me.

I admit that you've then lost me when you proceeded to somehow apply a normal distribution to the voting, since it is a straight yes/no.
Last edited by PhilW on 18:08 Sun 14 Sep 2014, edited 1 time in total.
PhilW
Dalva Golden White Colheita 1952
Posts: 3503
Joined: 14:22 Wed 15 Dec 2010
Location: Near Cambridge, UK

Re: The Scottish Independence Referendum, 18th Sept 2014

Post by PhilW »

jdaw1 wrote:
PhilW wrote:
DRT wrote:please define standardlynormally distributed.
It's exponential.
For those new to this rant, start here.
But for once, it actually is (normal distributions follow an exponential function, unless limited by boundary conditions), and Derek was missing his chance to say so!
User avatar
jdaw1
Cockburn 1851
Posts: 23613
Joined: 15:03 Thu 21 Jun 2007
Location: London
Contact:

Re: The Scottish Independence Referendum, 18th Sept 2014

Post by jdaw1 »

PhilW wrote:I admit that you've then lost me when you proceeded to somehow applying a normal distribution to the voting, since it is a straight yes/no.
But our knowledge of the Yes is that it will be μ±σ. By assuming σ, and using the observed probability of 76%, we can deduce μ, which isn’t far from ½.
User avatar
jdaw1
Cockburn 1851
Posts: 23613
Joined: 15:03 Thu 21 Jun 2007
Location: London
Contact:

Re: The Scottish Independence Referendum, 18th Sept 2014

Post by jdaw1 »

PhilW wrote:But for once, it actually is (normal distributions follow an exponential function, unless limited by boundary conditions), and Derek was missing his chance to say so!
The equation for a normal distribution isn’t exponential, even though it contains an exponential. As you well know.

(Are we off topic?)
User avatar
DRT
Fonseca 1966
Posts: 15779
Joined: 23:51 Wed 20 Jun 2007
Location: Chesterfield, UK
Contact:

Re: The Scottish Independence Referendum, 18th Sept 2014

Post by DRT »

PhilW wrote:
jdaw1 wrote:
PhilW wrote:
DRT wrote:please define standardlynormally distributed.
It's exponential.
For those new to this rant, start here.
But for once, it actually is (normal distributions follow an exponential function, unless limited by boundary conditions), and Derek was missing his chance to say so!
I didn't want to embarrass him in public.
"The first duty of Port is to be red"
Ernest H. Cockburn
User avatar
djewesbury
Graham’s 1970
Posts: 8165
Joined: 20:01 Mon 31 Dec 2012
Location: Gothenburg, Sweden
Contact:

Re: The Scottish Independence Referendum, 18th Sept 2014

Post by djewesbury »

jdaw1 wrote:
PhilW wrote:But for once, it actually is (normal distributions follow an exponential function, unless limited by boundary conditions), and Derek was missing his chance to say so!
The equation for a normal distribution isn’t exponential, even though it contains an exponential. As you well know.

(Are we off topic?)
Could an admin please move all this twaddle to Lies, Damned Lies and Meaningless Drivel?
Daniel J.
Husband of a relentless former Soviet Chess Master.
delete.. delete.. *sigh*.. delete...
User avatar
DRT
Fonseca 1966
Posts: 15779
Joined: 23:51 Wed 20 Jun 2007
Location: Chesterfield, UK
Contact:

Re: The Scottish Independence Referendum, 18th Sept 2014

Post by DRT »

Please link to the post where you want to make the split.
"The first duty of Port is to be red"
Ernest H. Cockburn
User avatar
djewesbury
Graham’s 1970
Posts: 8165
Joined: 20:01 Mon 31 Dec 2012
Location: Gothenburg, Sweden
Contact:

Re: The Scottish Independence Referendum, 18th Sept 2014

Post by djewesbury »

DRT wrote:Please link to the post where you want to make the split.
Somebody appears to doing it already. I think the point where mathematicians start discussing bookies' odds is probably a good point to administer the scalpel.
Daniel J.
Husband of a relentless former Soviet Chess Master.
delete.. delete.. *sigh*.. delete...
User avatar
jdaw1
Cockburn 1851
Posts: 23613
Joined: 15:03 Thu 21 Jun 2007
Location: London
Contact:

Re: Lies, Damned Lies and Meaningless Drivel

Post by jdaw1 »

Above posts moved here from The Scottish Independence Referendum, 18th Sept 2014 by jdaw1.
LGTrotter
Dalva Golden White Colheita 1952
Posts: 3707
Joined: 17:45 Fri 19 Oct 2012
Location: Somerset, UK

Re: Lies, Damned Lies and Meaningless Drivel

Post by LGTrotter »

Maths is nice. What happened to the link to the maths jokes website? I'm sure there was one.
PhilW
Dalva Golden White Colheita 1952
Posts: 3503
Joined: 14:22 Wed 15 Dec 2010
Location: Near Cambridge, UK

Re: The Scottish Independence Referendum, 18th Sept 2014

Post by PhilW »

jdaw1 wrote:(Are we off topic?)
Yes, but it's a good topic.
jdaw1 wrote:
PhilW wrote:But for once, it actually is (normal distributions follow an exponential function, unless limited by boundary conditions), and Derek was missing his chance to say so!
The equation for a normal distribution isn’t exponential, even though it contains an exponential. As you well know.
With σ defined/assumed as a constant, then the normal distribution function is of the form f(x)=k.eg(x)
Does g(x) being of quadratic form therefore preclude this equation from being described as of exponential form?
User avatar
flash_uk
Graham’s 1977
Posts: 4081
Joined: 20:02 Thu 13 Feb 2014
Location: London

Re: Lies, Damned Lies and Meaningless Drivel

Post by flash_uk »

LGTrotter wrote:Maths is nice. What happened to the link to the maths jokes website? I'm sure there was one.
Here are a couple of cartoons which made me chuckle
Attachments
cartoon_maths_1.jpg
cartoon_maths_1.jpg (29.36 KiB) Viewed 6418 times
User avatar
flash_uk
Graham’s 1977
Posts: 4081
Joined: 20:02 Thu 13 Feb 2014
Location: London

Re: Lies, Damned Lies and Meaningless Drivel

Post by flash_uk »

and the other
Attachments
cartoon_maths_2.jpg
cartoon_maths_2.jpg (28.41 KiB) Viewed 6418 times
Post Reply