1725: "Linear Regression"

This forum is for the individual discussion thread that goes with each new comic.

Moderators: Moderators General, Prelates, Magistrates

User avatar
orthogon
Posts: 2879
Joined: Thu May 17, 2012 7:52 am UTC
Location: The Airy 1830 ellipsoid

1725: "Linear Regression"

Postby orthogon » Fri Aug 26, 2016 1:34 pm UTC

Image

Alt-text: The 95% confidence interval suggests Rexthor's dog could also be a cat, or possibly a teapot.

Next time somebody asks me what star sign I am, I'm totally going with Rexthor.
xtifr wrote:... and orthogon merely sounds undecided.

richP
Posts: 176
Joined: Wed Aug 17, 2011 3:28 pm UTC

Re: 1725: "Linear Regression"

Postby richP » Fri Aug 26, 2016 1:39 pm UTC

Rexthor's teapot? Randall is trolling philosophers again?

User avatar
Lode
Posts: 67
Joined: Thu May 29, 2008 5:16 pm UTC

Re: 1725: "Linear Regression"

Postby Lode » Fri Aug 26, 2016 1:50 pm UTC

Now somebody needs to write a tool to do Rexthor-the-dog-bearer regression, finding the best fitting Rexthor-the-dog-bearer on any data set
Last edited by Lode on Fri Aug 26, 2016 4:09 pm UTC, edited 1 time in total.

kateract
Posts: 5
Joined: Wed Nov 03, 2010 6:30 pm UTC

Re: 1725: "Linear Regression"

Postby kateract » Fri Aug 26, 2016 2:00 pm UTC

It looks more like a chicken to me. Rexthor the Chicken Bearer. "The eyes of Rexthor and his Chicken are upon us brothers! He favors us in the coming battle!"

User avatar
cellocgw
Posts: 1888
Joined: Sat Jun 21, 2008 7:40 pm UTC

Re: 1725: "Linear Regression"

Postby cellocgw » Fri Aug 26, 2016 2:00 pm UTC

Lode wrote:Now somebody needs to write a tool to do Rexthor-the-dog-bearer regression, finding the best fitting Rexthor-the-dog-bearer-regression on any data set


Strange as it sounds, there are tools that let you fit to arbitrary formulas. See, e.g., R nls-package . It does get hairier when the function isn't single-valued, as would be the case with Rexthor here.

Anyway, I'm not sure how this true-life bit of statistics snuck into xkcd. You have to keep beating your students/coworkers over the head with goodness-of-fit values until they stop thinking that a reported linear fit line has any relationship to the dataset.
https://app.box.com/witthoftresume
Former OTTer
Vote cellocgw for President 2020. #ScienceintheWhiteHouse http://cellocgw.wordpress.com
"The Planck length is 3.81779e-33 picas." -- keithl
" Earth weighs almost exactly π milliJupiters" -- what-if #146, note 7

User avatar
Flumble
Yes Man
Posts: 1996
Joined: Sun Aug 05, 2012 9:35 pm UTC

Re: 1725: "Linear Regression"

Postby Flumble » Fri Aug 26, 2016 2:08 pm UTC

If at first you don't succeed, scale the axes.

very linear.png
very linear.png (8.01 KiB) Viewed 7512 times

ThemePark
Posts: 450
Joined: Fri Jun 27, 2008 5:42 pm UTC
Location: Århus, Denmark

Re: 1725: "Linear Regression"

Postby ThemePark » Fri Aug 26, 2016 2:27 pm UTC

kateract wrote:It looks more like a chicken to me. Rexthor the Chicken Bearer. "The eyes of Rexthor and his Chicken are upon us brothers! He favors us in the coming battle!"

Nope, Rexthor is a waiter, and that's a glass of ice cream with a waffle and a spoon.
I have traveled from 1979 to be a member of the unofficial board Council of Elders. Phear M3

stilettoblade
Posts: 24
Joined: Tue Mar 17, 2009 6:10 pm UTC

Re: 1725: "Linear Regression"

Postby stilettoblade » Fri Aug 26, 2016 2:33 pm UTC

I'm going with "Rexthor, the Martini-Drinker"

User avatar
da Doctah
Posts: 873
Joined: Fri Feb 03, 2012 6:27 am UTC

Re: 1725: "Linear Regression"

Postby da Doctah » Fri Aug 26, 2016 3:55 pm UTC

That's not Rexthor at all. It's Arrrrrgus, the Pirate, and what you're calling his dog, cat, teapot, etc is Poliwanacrackus, the Parrot.

User avatar
Quizatzhaderac
Posts: 1530
Joined: Sun Oct 19, 2008 5:28 pm UTC
Location: Space Florida

Re: 1725: "Linear Regression"

Postby Quizatzhaderac » Fri Aug 26, 2016 4:32 pm UTC

Lode wrote:Now somebody needs to write a tool to do Rexthor-the-dog-bearer regression, finding the best fitting Rexthor-the-dog-bearer on any data set
Actually, I'd say a large number of different constellations. If any constellation fits better than a line, it's a mess not a graph.
The thing about recursion problems is that they tend to contain other recursion problems.

chenille
Posts: 409
Joined: Tue Dec 20, 2011 7:25 pm UTC

Re: 1725: "Linear Regression"

Postby chenille » Fri Aug 26, 2016 5:26 pm UTC

Quizatzhaderac wrote:If any constellation fits better than a line, it's a mess not a graph.

With some exceptions like Hydra.

User avatar
Copper Bezel
Posts: 2416
Joined: Wed Oct 12, 2011 6:35 am UTC
Location: Web exclusive!

Re: 1725: "Linear Regression"

Postby Copper Bezel » Fri Aug 26, 2016 6:13 pm UTC

Though that could simply be an example of the converse, that if your constellation is just a jaggy line, it's kind of a shit constellation. Still, I guess it comes to the same thing, that we really have to run the test in reverse first and identify those constellations that don't have any business being constellations.
So much depends upon a red wheel barrow (>= XXII) but it is not going to be installed.

she / her / her

User avatar
ShuRugal
Posts: 75
Joined: Wed Jan 26, 2011 5:19 am UTC

Re: 1725: "Linear Regression"

Postby ShuRugal » Fri Aug 26, 2016 9:34 pm UTC

da Doctah wrote:That's not Rexthor at all. It's Arrrrrgus, the Pirate, and what you're calling his dog, cat, teapot, etc is Poliwanacrackus, the Parrot.


NO! It is Shiiingus, the ninja! He is holding the head and spinal column he has just cut from your unsuspecting corpse!

User avatar
keithl
Posts: 633
Joined: Mon Aug 01, 2011 3:46 pm UTC

Re: 1725: "Linear Regression"

Postby keithl » Fri Aug 26, 2016 10:55 pm UTC

Physicists are not immune, especially if they offset 3 sigma results with the logarithm of the megaeuros spent acquiring them, or of the hundreds of papers written about them. FYI, after six more months of intense data gathering, delaying data taking for other scheduled projects at LHC, the 750 GeV "digamma resonance" shown below disappeared back into the statistical noise. The boring old standard model still explains everything we can observe, though it does not specify a suitable punishment for thousands of three-sigma results chasers.
Perhaps they should be sentenced to a career change to pharmaceutical research, where cherry-picked 3 sigma results are considered the gold standard.
Attachments
digamma.png

xtifr
Posts: 322
Joined: Wed Oct 01, 2008 6:38 pm UTC

Re: 1725: "Linear Regression"

Postby xtifr » Sat Aug 27, 2016 12:39 am UTC

kateract wrote:It looks more like a chicken to me. Rexthor the Chicken Bearer. "The eyes of Rexthor and his Chicken are upon us brothers! He favors us in the coming battle!"

Her chicken! She favors! :P ;)
"[T]he author has followed the usual practice of contemporary books on graph theory, namely to use words that are similar but not identical to the terms used in other books on graph theory."
-- Donald Knuth, The Art of Computer Programming, Vol I, 3rd ed.

User avatar
CorruptUser
Posts: 9874
Joined: Fri Nov 06, 2009 10:12 pm UTC

Re: 1725: "Linear Regression"

Postby CorruptUser » Sat Aug 27, 2016 12:33 pm UTC

In cases like this, generally speaking the linear regression will fail the p test anyway. That and SSR will be tiny compared to the SSE, that is, the variance explained by the model is tiny compared to the variance not explained by the model.

NemeSys
Posts: 10
Joined: Mon Dec 17, 2012 4:24 pm UTC

Re: 1725: "Linear Regression"

Postby NemeSys » Sat Aug 27, 2016 1:11 pm UTC

Damn it! Now when I look at Orion all I can see is him throwing a frisbee ("Fetch, Canis Major, fetch!")

All_¥our_Bass
Posts: 13
Joined: Thu Jan 03, 2013 1:17 am UTC

Re: 1725: "Linear Regression"

Postby All_¥our_Bass » Sun Aug 28, 2016 8:38 am UTC

bearholdingshark.jpg

User avatar
Sableagle
Ormurinn's Alt
Posts: 1652
Joined: Sat Jun 13, 2015 4:26 pm UTC
Location: The wrong side of the mirror
Contact:

Re: 1725: "Linear Regression"

Postby Sableagle » Sun Aug 28, 2016 8:43 am UTC

ThemePark wrote:
kateract wrote:It looks more like a chicken to me. Rexthor the Chicken Bearer. "The eyes of Rexthor and his Chicken are upon us brothers! He favors us in the coming battle!"

Nope, Rexthor is a waiter, and that's a glass of ice cream with a waffle and a spoon.


I started with "cocktail waitress" and the extra stars were going to be falling sparkles, but I couldn't really place a sparkler so ...

Constellation.png
Oh, Willie McBride, it was all done in vain.

User avatar
Sableagle
Ormurinn's Alt
Posts: 1652
Joined: Sat Jun 13, 2015 4:26 pm UTC
Location: The wrong side of the mirror
Contact:

Re: 1725: "Linear Regression"

Postby Sableagle » Sun Aug 28, 2016 8:44 am UTC

xtifr wrote:
kateract wrote:It looks more like a chicken to me. Rexthor the Chicken Bearer. "The eyes of Rexthor and his Chicken are upon us brothers! He favors us in the coming battle!"

Her chicken! She favors! :P ;)

Reginathor, then?
Oh, Willie McBride, it was all done in vain.

User avatar
Znirk
Posts: 174
Joined: Mon Jul 01, 2013 9:47 am UTC
Location: ZZ9 plural Z α

Re: 1725: "Linear Regression"

Postby Znirk » Sun Aug 28, 2016 11:05 am UTC

Sableagle wrote:Reginathor, then?

As opposed to Reginator, which is one of the titles of Geoffrey Fisher.

User avatar
somitomi
Posts: 585
Joined: Fri Nov 06, 2015 11:21 pm UTC
Location: can be found in Hungary
Contact:

Re: 1725: "Linear Regression"

Postby somitomi » Sun Aug 28, 2016 7:40 pm UTC

Copper Bezel wrote:Though that could simply be an example of the converse, that if your constellation is just a jaggy line, it's kind of a shit constellation. Still, I guess it comes to the same thing, that we really have to run the test in reverse first and identify those constellations that don't have any business being constellations.

I never really understood how people decided this for example is a weighing scale. To me it could be anything from a really shabby shed to a kite. Despite this, I did come up with two constellations, one named Great Triangle (guess what it looked like) and another one, which included every star except one for some reason.
—◯-◯

teelo
Posts: 759
Joined: Thu Apr 08, 2010 11:50 pm UTC

Re: 1725: "Linear Regression"

Postby teelo » Mon Aug 29, 2016 8:11 am UTC

But what-if its actually a connect-the-dots of a Pokemon?

User avatar
orthogon
Posts: 2879
Joined: Thu May 17, 2012 7:52 am UTC
Location: The Airy 1830 ellipsoid

Re: 1725: "Linear Regression"

Postby orthogon » Mon Aug 29, 2016 2:45 pm UTC

somitomi wrote:[
I never really understood how people decided this for example is a weighing scale. To me it could be anything from a really shabby shed to a kite. Despite this, I did come up with two constellations, one named Great Triangle (guess what it looked like) and another one, which included every star except one for some reason.

How many stars were in the Great Triangle?
xtifr wrote:... and orthogon merely sounds undecided.

User avatar
somitomi
Posts: 585
Joined: Fri Nov 06, 2015 11:21 pm UTC
Location: can be found in Hungary
Contact:

Re: 1725: "Linear Regression"

Postby somitomi » Mon Aug 29, 2016 8:20 pm UTC

orthogon wrote:
somitomi wrote:[
I never really understood how people decided this for example is a weighing scale. To me it could be anything from a really shabby shed to a kite. Despite this, I did come up with two constellations, one named Great Triangle (guess what it looked like) and another one, which included every star except one for some reason.

How many stars were in the Great Triangle?

The quite surprising answer is
Spoiler:
3
But I haven't a clue, which ones they were, I recall picking three of the brighter ones.
—◯-◯

User avatar
addams
Posts: 9869
Joined: Sun Sep 12, 2010 4:44 am UTC
Location: Oregon Coast: 97444

Re: 1725: "Linear Regression"

Postby addams » Tue Sep 06, 2016 3:21 pm UTC

Somehow; This seems to belong in this thread.
Sableagle wrote:
Diadem wrote:
Chen wrote:www.dailymail.co.uk
And god that article was terrible.

https://www.youtube.com/watch?v=h4MhbkW ... tu.be&t=15
At 10 min and 25 sec The Graph appears.
At about the same time hope for humanity evaporates.

keithl seems to understand what drives us to stare off into space...
And, see Rexthor the Dog bearer.
keithl wrote:
Spoiler:
Physicists are not immune, especially if they offset 3 sigma results with the logarithm of the megaeuros spent acquiring them, or of the hundreds of papers written about them. FYI, after six more months of intense data gathering, delaying data taking for other scheduled projects at LHC, the 750 GeV "digamma resonance" shown below disappeared back into the statistical noise.
The boring old standard model still explains everything we can observe, though it does not specify a suitable punishment for thousands of three-sigma results chasers.
Perhaps they should be sentenced to a career change to pharmaceutical research, where cherry-picked 3 sigma results are considered the gold standard.

Look for the missing data or look for the obvious pattern?

oh..What to do?..What to do?
Look for the obvious pattern!

We might see what Sableagle sees.
It might be "One Heck of a Party."
Life is, just, an exchange of electrons; It is up to us to give it meaning.

We are all in The Gutter.
Some of us see The Gutter.
Some of us see The Stars.
by mr. Oscar Wilde.

Those that want to Know; Know.
Those that do not Know; Don't tell them.
They do terrible things to people that Tell Them.

Bratmon
Posts: 23
Joined: Sun Aug 30, 2009 7:54 pm UTC

Re: 1725: "Linear Regression"

Postby Bratmon » Wed Mar 01, 2017 1:54 am UTC

538 just tried to use this graph as a key argument for an article:
Image

I'm seeing a crossbow:

Image

User avatar
pogrmman
Posts: 503
Joined: Wed Jun 29, 2016 10:53 pm UTC
Location: Probably outside

Re: 1725: "Linear Regression"

Postby pogrmman » Wed Mar 01, 2017 5:37 am UTC

Bratmon wrote:538 just tried to use this graph as a key argument for an article:
Image

I'm seeing a crossbow:

Image


I agree with you, but the article does mention that while this isn't a particularly strong correlation in finance or hard sciences, it is pretty strong for political science.

However, the paper the article cites seems to contradict that directly:
"One interesting result is the fact that that the addition of the regional instability index REGREV has almost no effect the magnitude of the coefficient on the domestic instability index REV."

Yeah, the paper does have some OK correlations between regional stability and GDP growth. The author was stupid to pick this graph. She should have used figure 3 which actually shows somewhat of a relationship between regional instability and GDP growth. Even then, it still doesn't fit with her point of weakened security.

EDIT: Also, their "regional instability index" is just the average number of coups or revolutions per year in the neighboring countries. Neither that nor the domesticic one has anything to say about security as a whole.


Return to “Individual XKCD Comic Threads”

Who is online

Users browsing this forum: No registered users and 27 guests