Talk:Sam Wang (neuroscientist)

Latest comment: 7 years ago by SnowFire in topic Wang's election prediction misses

Please add at least 5 photos edit

Please add at least 5 photos

  1. official
  2. lecturing
  3. showing brain parts or psepho-diagrams
  4. happy with students or friends
  5. signing books — Preceding unsigned comment added by 2A02:587:4109:FA00:8DD8:AF26:5823:633E (talk) 05:34, 18 January 2016 (UTC)Reply

Wang's election prediction misses edit

The previous version, "PEC's calls were more of a miss", is an obvious use of weasel words. Let's just admit he was wrong — Preceding unsigned comment added by 156.111.111.113 (talk) 13:50, 28 March 2017 (UTC)Reply

On the contrary, it's accurate and far from weasel words. Suppose a Las Vegas oddsmaker assigns 10:1 odds on the underdog, and the underdog wins anyway - some sort of Leicster City situation. Were these odds "incorrect"? Well, they might be, but the act of assigning odds inherently says "Yes, the underdog will win X% of the time." If the underdog wins, that doesn't necessarily mean the oddsmaker screwed up. Or, put another way, if the #1 seeds in the NCAA don't all make the Final Four, that doesn't mean that the rankings committee were wrong. Now, did Wang's odds underestimate the probability of a Trump win? Sure, he's said as much himself. But were they "incorrect"? Wrong term, unless he'd claimed 100% odds of a Clinton victory. SnowFire (talk) 15:32, 28 March 2017 (UTC)Reply

@snowfire, that logic is specious. By your reasoning, no poll prediction (or any prediction) is incorrect if it assigns any non-zero probability to anything. If I predict that tomorrow the sun will not rise with 99.99% probability and it does, it would be reasonable to say I am incorrect. Wang himself basically admits he was wrong, here.2016/11/11/politics-polls-20-what-just-happened/. I'm all for expressing complexity of how wrong, but "more of a miss" is a very imprecise term. — Preceding unsigned comment added by 156.111.111.113 (talk) 16:41, 28 March 2017 (UTC)Reply

Well... *yes*. If you asked someone in June 1914 what the odds were of the biggest war to date starting in the next 3 months that would kill a zillion people and last at least 4 years, they might reasonably say "less than 0.1%." And they would be right, even with the benefit of retrospect, but that war would still happen.
Look, I partially came to this article to rag on Wang too, but let's get our terms correct. Something other than "a miss" is fine, but "incorrect" is not, lest a 49% A /51% B prediction be "incorrect" if A wins. SnowFire (talk) 18:01, 28 March 2017 (UTC)Reply