The Math of XKCD
Mar. 19th, 2009 12:00 am![[personal profile]](https://www.dreamwidth.org/img/silk/identity/user.png)
A while ago, This Story appeared in Blogspot, posted by someone who was having a problem with Verizon billing. Apparently they thought he owed them a lot of money, when in fact he had not been a customer for five years, and furthermore, they owed him money according to his calculations. After this had gone on for a long time, it came to the attention of the author of XKCD, who became ... inspired.
Here is his inspiration: http://xkcd.com/verizon/
The check is written for an amount which contains the number "i", a.k.a., the square root of negative 1. Therefore the amount of the check would be called an 'imaginary' number in mathematical terms (because there is nothing you can multiply by itself to get -1). This should not be a problem, however, because the amount Verizon was charging was also imaginary...
Edit: From the comments, we see that the equation actually resolves to a real number that Verizon can spend whichever way it wants:
"Ah, but e^(i*pi) = cos(pi) + i*sin(pi) = -1 + 0i = -1.
And the infinite sum resolves to +1
So in fact he wrote them a check for two tenths of a cent.
That, of course, would require Verizon to do math to figure out."
XKCD - brilliant as usual.
Brought to my attention by way of
lindsaybits
Here is his inspiration: http://xkcd.com/verizon/
The check is written for an amount which contains the number "i", a.k.a., the square root of negative 1. Therefore the amount of the check would be called an 'imaginary' number in mathematical terms (because there is nothing you can multiply by itself to get -1). This should not be a problem, however, because the amount Verizon was charging was also imaginary...
Edit: From the comments, we see that the equation actually resolves to a real number that Verizon can spend whichever way it wants:
"Ah, but e^(i*pi) = cos(pi) + i*sin(pi) = -1 + 0i = -1.
And the infinite sum resolves to +1
So in fact he wrote them a check for two tenths of a cent.
That, of course, would require Verizon to do math to figure out."
XKCD - brilliant as usual.
Brought to my attention by way of
![[livejournal.com profile]](https://www.dreamwidth.org/img/external/lj-userinfo.gif)
(no subject)
Date: 2009-03-19 12:16 pm (UTC)And the infinite sum resolves to +1
So in fact he wrote them a check for two tenths of a cent.
That, of course, would require Verizon to do math to figure out.
(no subject)
Date: 2009-03-19 12:46 pm (UTC)Edited to include an entire piece of the equation that my brain omitted, heh.
(no subject)
Date: 2009-03-19 02:48 pm (UTC)e^(i*pi)=-1
Cos(pi) = -1
Sin(pi) = 0
Where's my spiffy 300-function calculator
I'm still not sure how/why e^(i*pi)=-1=cos(pi), though.
i*i = e^(i*pi)? How?
(no subject)
Date: 2009-03-19 03:45 pm (UTC)"... ln(x) is the inverse function of e^(x) (they are mirror images of each other on an x-y axis). But also there is no ln of a negative number. ln(-1) = i*Pi, which is not a real number but an imaginary number. That means that e^(i*Pi) = -1 by definition of logs."
So from what I can understand of the above, ln(-1) = i*pi. This means that e^(i*pi) = -1, since that's the way ln(y) and e^(x)=y work. I still don't know why, though. The closest that explanation came was that "this is the definition of how ln(x) (the base e logarithm of some number) works".
I still didn't get it, though. I went to a more extensive web page on the subject.
http://library.thinkquest.org/27512/c5i.html
Heh. I should have known it involved infinite series. Nothing with e and i in it will be anything but infinite.
apparently it's a famous equation:
========================================================
Here is the derivation.
(no subject)
Date: 2009-03-19 04:53 pm (UTC)(no subject)
Date: 2009-03-19 04:58 pm (UTC)(no subject)
Date: 2009-03-19 05:38 pm (UTC)The author of the page immediately introduces a new question:
"So now, the question is, why is e^ix=cos(x)+isin(x) the "right" thing to define what e raised to an imaginary power means?"
Yep, he hit it on the head. I'll find the calculus derivations somewhere, but I like graphs to help wrap the brain around it.
(no subject)
Date: 2009-03-19 05:45 pm (UTC)(no subject)
Date: 2009-03-19 06:09 pm (UTC)Lemma: sinx = x - (x3/3!) + (x5/5!) - (x7/7!) + ...
Derivation:
Let f(x) = sin(x) --> f(0) = sin(0) = 0
f'(x) = cos(x) --> f'(0) = cos(0) = 1
f''(x) = -sin(x) --> f'(0) = 0
f'''(x) = -cos x → f'(0) = -1
From this, we see that:
fn(0) = 0 if n is even.
fn(0) = 1 if (n-1)/2 is even
fn(0) = -1 if (n-1)/2 is odd
Putting this all together gives us:
sin(x) = (x/1)(1) + (x2/2!)(0) + (x3/3!)(-1) + ...
QED
Lemma: cos x = 1 - (x2)/2! + (x4/4!) - (x6/6!) + ...
Derivation:
Let f(x) = cos x
f(0) = cos(0) = 1
f'(x) = -sin x → f'(0) = 0
f''(x) = -cos(x) → f'(0) = -1
f'''(x) = sin x → f'(0) = 0
fn(0) = 0 if n is odd.
fn(0) = 1 if (n/2) is even
fn(0) = -1 if (n/2) is odd
Putting this all together gives us:
cos(x) = 1 + (x2/2!)(-1) + (x3/3!)(0) + (x4/4!)(1) + ...
I can see where these series are behaving exactly like sine and cosine waves, just like the graph.
(no subject)
Date: 2009-03-19 01:11 pm (UTC)(no subject)
Date: 2009-03-20 01:56 am (UTC)That is a very cool way to mess with them. The Network = Borg, anyway.
However, I think several of my brain cells have suffered Death by Math now.
(no subject)
Date: 2009-03-20 02:07 am (UTC)Sorry about the Death by Math - I got carried away trying to solve a Puzzle...
(no subject)
Date: 2009-03-21 01:41 am (UTC)It is one reason why I went to art school. :)