Do you have adblock enabled?
If you can read this, either the style sheet didn't load or you have an older browser that doesn't support style sheets. Try clearing your browser cache and refreshing the page.

(Ars Technica)   Computer program correctly predicts SCOTUS decisions 70% of the time, has more empathy than Scalia   (arstechnica.com) divider line 28
    More: Interesting, Supreme Court, Supreme Court decisions, heat map, fantasy sport, Josh Blackman  
•       •       •

647 clicks; posted to Politics » on 31 Jul 2014 at 2:37 PM (26 weeks ago)   |  Favorite    |   share:  Share on Twitter share via Email Share on Facebook   more»



28 Comments   (+0 »)
   
View Voting Results: Smartest and Funniest
 
ZAZ [TotalFark]
2014-07-31 02:20:58 PM  
With enough free parameters you can fit any curve. They have over 90.
 
2014-07-31 02:42:20 PM  
Big deal.  I can do the same thing.  All you need to do is decide what is unethical or unamerican and you'll get the same results.
 
2014-07-31 02:42:43 PM  

ZAZ: With enough free parameters you can fit any curve. They have over 90.


reminds me of in school when people would plot a set of points in excel, and then fit a line by increasing the order of the polynomial until the R^2 was close to 1.
 
2014-07-31 02:43:10 PM  
Empathy is not found in the Constitution.
 
2014-07-31 02:43:22 PM  
Well a baseline of randomly picking gives you 50%. Given that, I could probably hit 70% accuracy and I have no background in law, just a general familiarity with the overall left/right bias of each judge and some familiarity with their voting record on a handful of significant cases.
 
2014-07-31 02:45:30 PM  
People are always coming up with algorithms that predict past stock market behavior with great accuracy too. Then they try using them to predict the future, which some of them can do for a while until they don't.
 
2014-07-31 02:46:13 PM  

ZAZ: With enough free parameters you can fit any curve. They have over 90.



"With four parameters I can fit an elephant, and with five I can make him wiggle his trunk" - von Neumann
 
2014-07-31 02:49:10 PM  

jjorsett: People are always coming up with algorithms that predict past stock market behavior with great accuracy too. Then they try using them to predict the future, which some of them can do for a while until they don't.


I always remember that one group that created an algorithm that successfully predicted the outcome of the last 10 presidential elections (retroactively) and predicted a Romney win in 2012.
 
2014-07-31 02:49:50 PM  
It also has more of a soul than Scalia.
 
2014-07-31 02:55:33 PM  

Delta1212: Well a baseline of randomly picking gives you 50%. Given that, I could probably hit 70% accuracy and I have no background in law, just a general familiarity with the overall left/right bias of each judge and some familiarity with their voting record on a handful of significant cases.


Apparently poeple who are very knowledgeable are only 75% accurate.
 
2014-07-31 03:02:55 PM  

BMFPitt: Delta1212: Well a baseline of randomly picking gives you 50%. Given that, I could probably hit 70% accuracy and I have no background in law, just a general familiarity with the overall left/right bias of each judge and some familiarity with their voting record on a handful of significant cases.

Apparently poeple who are very knowledgeable are only 75% accurate.


Thereby disproving Sturgeon's Law, which says that 90% of everything is crap.
 
2014-07-31 03:04:09 PM  

give me doughnuts: It also has more of a soul than Scalia.


Don't gingers have a bit more of a soul than Scalia?
 
2014-07-31 03:15:50 PM  

jjorsett: BMFPitt: Delta1212: Well a baseline of randomly picking gives you 50%. Given that, I could probably hit 70% accuracy and I have no background in law, just a general familiarity with the overall left/right bias of each judge and some familiarity with their voting record on a handful of significant cases.

Apparently poeple who are very knowledgeable are only 75% accurate.

Thereby disproving Sturgeon's Law, which says that 90% of everything is crap.


90% of everything is everythin.
 
2014-07-31 03:22:30 PM  
might as well put 'puters on the U.S. Surpreme Court Bench.   the human ones currently there do what their Corporate Masters tell them to do.
 
2014-07-31 03:23:32 PM  
Yeah, but it still acquits Ben Reich.
 
2014-07-31 03:57:16 PM  

ZAZ: With enough free parameters you can fit any curve. They have over 90.


It's interesting when you click through to the pdf of the 90 variables. What variable has the strongest correlation? The month the decision was issued, at 2%.
 
2014-07-31 04:02:11 PM  

Vacation Bible School: give me doughnuts: It also has more of a soul than Scalia.

Don't gingers have a bit more of a soul than Scalia?


Depends on haw many they've stolen.
(count the freckles)
 
2014-07-31 04:05:30 PM  

kidgenius: ZAZ: With enough free parameters you can fit any curve. They have over 90.

reminds me of in school when people would plot a set of points in excel, and then fit a line by increasing the order of the polynomial until the R^2 was close to 1.


i.imgur.com
 
2014-07-31 04:08:11 PM  

Jiro Dreams Of McRibs: Empathy is not found in the Constitution.


3/5 of it is.
 
2014-07-31 04:09:57 PM  

ZAZ: With enough free parameters you can fit any curve. They have over 90.


To be fair, they're not just backfitting, they're evaluating future performance by leave-out cross validation.
 
2014-07-31 04:24:14 PM  

worlddan: It's interesting when you click through to the pdf of the 90 variables. What variable has the strongest correlation? The month the decision was issued, at 2%.


Actually, the strongest weight is "Justice Lower Court Direction Abs. Difference" at 0.02494.  (Though I can't replicate it with their data; I get 0.02035 when I average the weights for that feature.)
 
2014-07-31 04:25:47 PM  

kidgenius: ZAZ: With enough free parameters you can fit any curve. They have over 90.

reminds me of in school when people would plot a set of points in excel, and then fit a line by increasing the order of the polynomial until the R^2 was close to 1.


I have absolutely no idea what this means because I can't math.

What I do know is that if Scalia and Black Mr. Potato Head can find any way to be assholes, they'll do it. If there is a decision to be made that has ANY sort of grey area, they'll err on the side of the corporation that is funding their wife's livelihood. That seems to be a mathematical certainty.
 
2014-07-31 04:41:19 PM  
Heck, the US chamber of commerce beats that handily.
 
2014-07-31 04:50:03 PM  
It sounds like it was poorly programmed
 
2014-07-31 06:37:43 PM  
All hail the new Justice Bender, who invites you to bite his shiny metal briefs and promises to defy predictability by not disclosing the winning bribe argument which influences his decision.

www.superbwallpapers.com
 
2014-07-31 06:49:24 PM  
Given that another computer somewhere has figured out how to gerrymander districts to make them all but failure proof for one party or another, I don't think it is all that big a deal.
 
2014-07-31 08:30:37 PM  
I was under the impression that the Supreme Court was there to interpret the law and the Constitution.

If you want "empathy", go find a grandma and get a hug.
 
2014-08-01 09:25:23 AM  
I thought the mission of the SCOTUS was to interpret and apply the COTUS.

Is "empathy" in their charter?
 
Displayed 28 of 28 comments

View Voting Results: Smartest and Funniest


This thread is closed to new comments.

Continue Farking
Submit a Link »
On Twitter





In Other Media


  1. Links are submitted by members of the Fark community.

  2. When community members submit a link, they also write a custom headline for the story.

  3. Other Farkers comment on the links. This is the number of comments. Click here to read them.

  4. Click here to submit a link.

Report