TE
TechEcho
Home24h TopNewestBestAskShowJobs
GitHubTwitter
Home

TechEcho

A tech news platform built with Next.js, providing global tech news and discussions.

GitHubTwitter

Home

HomeNewestBestAskShowJobs

Resources

HackerNews APIOriginal HackerNewsNext.js

© 2025 TechEcho. All rights reserved.

World Is Not Normal

12 pointsby amitminaover 3 years ago

3 comments

noduermeover 3 years ago
The &quot;great&quot; insight here should be that anything normal can also be mapped as 80&#x2F;20 and vice-versa. How you decide to map it is more of a political choice, or maybe an emotional one, based on where you fall on any given economic or social curve.<p>The medium, in other words, is the message. Every one of these things can be plotted on a bell curve if you frame it differently:<p>&gt;&gt; A handful of cities inhabit most of the world population;<p>Change to distribution of land per person on earth<p>&gt;&gt;majority of revenue for organizations comes from a handful of products;<p>Change to products bought per consumer, or revenue as a whole<p>&gt;&gt;a handful of nations win most of the gold medals at the Olympics;<p>Change to average gold medals per nation<p>All the 80&#x2F;20 does in the above cases is lop off the bottom end of the curve.
derbOacover 3 years ago
The problem, as it were, alluded to by the piece, is that many traits <i>are</i> normal. They give the example of height, but this applies to many other things as well. So understanding how you get from one distribution to the other is the real issue.<p>Survivorship bias also can be modeled using a pareto distribution (search for &quot;pareto survival&quot;, or just notice the mention of insurance modeling), but of course the piece doesn&#x27;t spin the pareto that way.<p>Costs can also be modeled as a pareto as well (<a href="https:&#x2F;&#x2F;www.sciencedaily.com&#x2F;releases&#x2F;2016&#x2F;12&#x2F;161212115702.htm" rel="nofollow">https:&#x2F;&#x2F;www.sciencedaily.com&#x2F;releases&#x2F;2016&#x2F;12&#x2F;161212115702.h...</a>). So if you have a right skewed pareto on one end, and a left skewed pareto on the other...<p>It&#x27;s disturbing to me how statistical distributions get used in this way as political bludgeons.<p>Statistically there are other reasons to use the normal even when it&#x27;s known to be false, having to do with how to handle various forms of uncertainty. Is it overused? Probably. But if you had to pick something by default you&#x27;d get much further with the normal. Pretty far actually.
thriftwyover 3 years ago
Been removed?
评论 #30203810 未加载