Trump has 97% chance of beating Hillary!

Discussion in 'Politics, Religion, Social Issues' started by 1458279, Mar 20, 2016.

  1. 1458279 Suspended

    1458279

    Joined:
    May 1, 2010
    Location:
    California
    #1
  2. jerwin macrumors 65816

    Joined:
    Jun 13, 2015
    #2
    His actual model inspires little confidence. He's a New Hampshire booster, nothing more.
     
  3. mrkramer macrumors 603

    mrkramer

    Joined:
    Jul 11, 2006
    Location:
    Somewhere
    #3
    1. This article was already posted in a different thread a long time ago.
    2. He puts way too much emphasis on New Hampshire.
    3. Trump has already broken a lot of other political rules, it would seem that if this projection is actually usually accurate this is the cycle it will break down since all the other predictions have.
     
  4. 1458279 thread starter Suspended

    1458279

    Joined:
    May 1, 2010
    Location:
    California
    #4
    Except that in 100 years, he's never been wrong except Bush/Gore which was exceptionally close.
    --- Post Merged, Mar 20, 2016 ---
    It seems that Trump breaking rules has always favored Trump.

    Unless you're saying that the Trump rule breaking has changed the whole thing that it's almost as if Trump were not in a regular party. In that case, ok, I can see that. In which case, it's really a matter of will the standard Right voters come out and vote in addition to others that Trump brings in.
     
  5. mrkramer macrumors 603

    mrkramer

    Joined:
    Jul 11, 2006
    Location:
    Somewhere
    #5
    And how many of those were before he made up the model? It's easy enough to make a model that will predict results that you already know, especially when there aren't a lot of events, there aren't a huge number of presidential elections to go off of.
     
  6. 1458279 thread starter Suspended

    1458279

    Joined:
    May 1, 2010
    Location:
    California
    #6
    Well it doesn't give much info on the model he's using. I'd like to see some more details because it sounds kinda like a "no candidate has ever..." kinda thing.

    Anyone have any details about his model?
     
  7. jerwin macrumors 65816

    Joined:
    Jun 13, 2015
    #7
    If I told you in 2008 that voters would never elect Obama, based on the fact that they had never elected a black president, and *never will*, I would have been foolish.

    Predicting the election based on political science models is a game. And because he presents a incredibly simplistic model coupled with a very early prediction, Norpath serves as a bit of a goad.

    Anyway, you can read some of his papers:

    Comfort for Obama: History is on His Side in 2012
    Time for a Change? Forecasting 2008
     
  8. thekev macrumors 604

    thekev

    Joined:
    Aug 5, 2010
    #8

    What matters is that the model was able to predict a sufficient number of past elections before their actual data was added. That would be an indication of a good predictor. If you're just looking for something that fits all the points without any disturbing singularities, then you run into the possibility of overfitting. (links are wiki and wolfram definitions)

    The article itself doesn't give you enough to scrutinize it. It's just more clickbait. I did a tiny bit of digging.

    This is his university page.

    One of the links on his page was broken, and the wayback machine had vandalized links on some of the snapshots. This is the best one I could find.

    I also dug out one of the embedded links here.

    I can't find any concrete details regarding the amount of test data. It matters whether the model can predict new data that wasn't used to construct it in the first place. I don't have anything against the guy, but I haven't seen enough to really support his case. If I was really interested, I might drop him an email. Some professors like to talk about their work.
     
  9. 1458279 thread starter Suspended

    1458279

    Joined:
    May 1, 2010
    Location:
    California
    #9
    It is interesting that he's only been wrong once and that was a the Bush/Gore race and that was as close as it gets.

    Maybe the 97% number comes from the fact that Hillary lost early and the model didn't take into account that the area was Bernie country. IDK.
     
  10. Zombie Acorn macrumors 65816

    Zombie Acorn

    Joined:
    Feb 2, 2009
    Location:
    Toronto, Ontario
    #10
    Model probably factored in criminality and auto dropped Hillary.
     
  11. cfedu, Mar 20, 2016
    Last edited: Mar 20, 2016

    cfedu macrumors 65816

    cfedu

    Joined:
    Mar 8, 2009
    Location:
    Toronto
    #11

    This race has a dynamic that has not been seen in over 50 years. Once both candidates are nominated, only then will be be able to get a good grasp of what will happen. I think both Hilary and trump both have huge potential to crash and burn.
     
  12. thekev macrumors 604

    thekev

    Joined:
    Aug 5, 2010
    #12
    I can't tell if that's a joke or a typo.

    It's not that interesting if you only have that information. That was my point. You can choose some least squares model for whatever features you would like to generate the data. This will not necessarily make it an accurate predictor with new data. As I mentioned overfitting can be an issue. I find it a little disturbing that he didn't show any prior case studies made before the results were actually in.
     
  13. cfedu macrumors 65816

    cfedu

    Joined:
    Mar 8, 2009
    Location:
    Toronto
    #13
    Typo!!!
     
  14. lparsons21 macrumors 6502

    lparsons21

    Joined:
    Jun 3, 2014
    Location:
    Southern Illinois
    #14
    One can only hope!!

    We haven't seen this level of stupidity on display at the Presidential level in my lifetime, or at least 60+ years. Every time Trump runs his yap I'm amazed that he keeps winning primaries.

    And Hillary's potential legal issues have got to come into play I would think.

    With Trump I'm constantly flabbergasted when I see people I know that are generally intelligent, well read and successful that buy into his particular brand of BS. I fully understand the majority of his supporters are angry because they know that those in power bailed out the fat cats, that we have a congress that collectively won't do their job, all the while the 'common man' is told to go punt.
     
  15. hiddenmarkov macrumors 6502a

    Joined:
    Mar 12, 2014
    Location:
    Japan
    #15
    Generally when you see academia on mainstream sites not academic or technical ones you need to take that with a grain of salt.

    Ideally the preferred outlet is established and accredited journals. To be peer reviewed. Since you have to at least provide the algorithmic and mathematical basis to the model. Source code and test data nice additions as well to have something called reproducibility. Have a statistical model that foresees the future....can we get 1000's to verify this with the source code?

    Sample of data used, how was is cleaned, does it have potential issues? this we like to see as well. In R for example I can clean data to my hearts content. To the point data I don't like is removed. It is common to do this with say NA values. An accepted removal of data only because it will break the analysis fast. Or, tbh, I can improperly in a stupid attack (or do it intentionally if I wanted to for some reason) clean the data to make the analysis flawed before the analysis portion of the code is even hit.

    That and well...I am a gamer. Many games I fall under the whims of the RNG (random number generator) gods. X-com players will know my pain well. Probability is one of its most made fun of aspects.

    Got the point blank shot lined up. You are in their face....this can't miss. It hits the alien its going down. It misses, your character is exposed and is probably royally screwed in the enemy's turn. Probability of hit is 85%. Press the button and the shot misses.

    Welcome to X-Com (and RNG based gaming in general), we have all been here before if gamers. RNG gods are cruel at times.
     
  16. Jess13, Mar 20, 2016
    Last edited: Mar 20, 2016

    Jess13 Suspended

    Jess13

    Joined:
    Nov 3, 2013
    #16
    GOPe Paul Ryan said a few days ago, on the subject of Trump presidency: “We’ll make it work if it happens.” Yesterday, we witnessed round two of liberal organizations-directed anti-Trump civil unrest slash major protests. GOPe effectively conceding that Trump presidency is likely; and liberal freak out, from realization that Trump could steal the election from Hillary or Bernie. (If Trump wasn’t perceived as legitimate threat, wouldn’t be spastic attempts to counter, silence and shut down.) Earlier today, I tabulated likes of the [at that time] 100 most recent tweets by Trump and Hillary, and found: Trump had over 1,000,000 more likes. Trump is way more popular than polls conducted with very limited samples of usually only a few hundred likely voters would have you believe.


    1.png
     
  17. LizKat macrumors 68040

    LizKat

    Joined:
    Aug 5, 2004
    Location:
    Catskill Mountains
    #17
    You estimate the actual popularity and potential political success of Trump and Clinton based on who loved their tweets? You deserve to be surprised when the real ballots get counted. :)

    I hardly ever bother to like (love? -- the stupid upvote is a heart now) tweets, I just scroll through and sometimes clink on a link in a tweet from someone I follow.

    Anyway if I "loved" all the tweets I have enjoyed even in my limited twitter feed, I'd have some sort of carpal tunnel syndrome by now. I read all different viewpoints in media of longer form, but my Twitter feed is only from people whose tweets I really enjoy, plus a few breaking news feeds which I vary from time to time.

    So the list of people that I follow is basically one big love fest. Trump's definitely not on that list and neither are the Clintons.
     
  18. Jess13 Suspended

    Jess13

    Joined:
    Nov 3, 2013
    #18
    I’m not saying with any certainty that Trump’s vastly more popular Twitter is going to translate into him receiving more real votes if he is the nominee, but it has to be factored in and seriously considered because it just may. And my comment was more than just Trump having 1,000,000+ more likes than Hillary in only 100 tweets, it had to do with GOPe leader’s comment and major liberal freak out, as well. I can’t say what’s going to happen or not, in the general election. But I wouldn’t be surprised if Trump wins.
     
  19. LizKat macrumors 68040

    LizKat

    Joined:
    Aug 5, 2004
    Location:
    Catskill Mountains
    #19
    OK :) and if Trump wins I may congratulate you from my new perch as an expatriate somewhere. I'll grant that this is a weird year in American politics. It's hard for any pollster to swear right now by their general election matchups, because of these populist surges from both left and right, plus the clear split that already exists in the Republican Party as evidenced by the House situation.

    Also, either of the two front runners may eventually run into the same roadblocks Obama has done with this gridlocked Congress, if neither candidate has coattails to affect Senate races in the general election. It's all enough to give God a headache.
     
  20. Technarchy macrumors 603

    Technarchy

    Joined:
    May 21, 2012
    #20
    There's a slight difference. Trump can buy votes in congress through countless means due to his high level of wealth. Through PAC's, direct contributions and speaking/contract gig promises, he can make gridlock vanish in a way that Obama could never dream of.
     
  21. oldmacs macrumors 68040

    Joined:
    Sep 14, 2010
    Location:
    Australia
    #21
    If thats the case get ready for disaster. Time to move to Mars.
     
  22. APlotdevice macrumors 68040

    APlotdevice

    Joined:
    Sep 3, 2011
    #22
    Perhaps there will be a more vested interest in getting money out of politics if that happens?

    I know, fat chance.
     
  23. Eraserhead, Mar 21, 2016
    Last edited: Mar 21, 2016

    Eraserhead macrumors G4

    Eraserhead

    Joined:
    Nov 3, 2005
    Location:
    UK
    #23
    Taylor Swift and Selena Gomez for president.
    --- Post Merged, Mar 21, 2016 ---
    It's predicting 25 elections with a 50:50 chance each time with some weighting towards the other party after 8 years and with presidents more likely to win two consecutive terms and with longer term cycles as well - which I believe (if you count 2000 as a Gore win) is up for the Democratic Party based on historical precedent.

    So the longer cycles were 7 out of 9 for the republicans, followed by 7 out of 9 for the Democratic Party followed by 5 out of 6 for the republicans followed by 5 out of 6 (including al gore in 2000) for the Democratic party.
     
  24. Thomas Veil macrumors 68020

    Thomas Veil

    Joined:
    Feb 14, 2004
    Location:
    OBJECTIVE reality
    #24
    That comes across as if you think it's a good thing. I trust you didn't mean it that way.
     
  25. bent christian Suspended

    bent christian

    Joined:
    Nov 5, 2015
    #25
    2016 is where political careers go to die.

    Norpoth developed his model in 1996, so he has had 5 elections without previous data. His model was wrong in one of those elections.

    Attaching his name to Trump in this way means that we probably won't ever be hearing from him ever again past December 2016. Donald Trump will not be our next president, and the name Helmut Norpoth will basically be forgotten.
     

Share This Page