Maths, Misunderstandings, and Manipulations

• Jul 29, 2014 · 7:00 PM

Maths, Misunderstandings & Manipulation

So many simple mathematical concepts seem to baffle us. Coin flips and lottery tickets often lead to confusion, but did you know there are at least 3 different kinds of probability? Did you know that Bertrand Russell took 350 pages to prove 2 + 2 = 4 and that the guy on the evening news explaining the accuracy of the survey he just told you about is highly unlikely to understand what the results actually mean?

On July 29th, Greg Solomon will help us understand some of the basics concepts of probability and will also discuss the more complicated maths that apply directly to science. He is an actuary - which means he merges the world of statistics with finance on a daily basis, for insurance clients. Come and hear him speak about Sample Size, P-values, Sigma, Bayes, false correlations, and a few other important aspects of mathematics & manipulation.

The talk is aimed at the layman - no statistical knowledge is required (although knowing that 2+2=4 might help).

Hope to see you all there

Come with questions. Leave with answers.

• Paul L.

Some skeptics point to a bit of a debate between emphasizing Bayesian or p-values in evaluating surprising claims. Here's an example from Steven Novella: http://www.sciencebasedmedicine.org/beware-the-p-value/
And here's something on it from Massimo Pigliucci: http://rationallyspeaking.blogspot.hk/2010/09/eliezer-yudkowsky-on-bayes-and-science.html I find this very interesting, but it seems to be quite a jump from Bayes theorem in predicting the draw of a playing card to evaluating acupuncture claims.

If Greg or anyone can share some insight, it would be enjoyable.

August 7, 2014

• Greg

I didn't have time to read the articles as slowly as I would have liked, but I'm not sure I agree with the author that we must make it extra hard to prove such points, and start be creating a bias because they are proving an unlikely thing. Bias in any form is wrong, start with what you KNOW - not what you believe. They seem to break their own rule this way. The nice thing about the law of large numbers is that maybe one study is enough to disprove the null at the 5% level, if you combine a whole bunch of studies then you can really see how significant the results really are.

August 10, 2014

• Greg

And one other point that comes to mind, which I didn't have time include in my lecture there the other night, is the placebo effect. We know it's real. And it can be significant. In the past drugs were deemed effective if they produced a bigger effect than "random chance". But more and more I'm seeing studies where they are showing they are effective beyond the placebo effect too.

August 10, 2014

• Greg

And here is a link to the Google Ngram search I spoke about, which lets you create your own odd correlations, like zombies & ketchup:

August 2, 2014

• Ralph

More examples of correlations of unrelated events. http://tylervigen.com/

2 · August 1, 2014

25 went

• Paul L.
• A former member
• A former member
• A former member
• A former member
• A former member
• A former member

Hong Kong, Hong Kong

Founded Apr 19, 2012

People in this Meetup are also in:

• Meat Free Hong Kong

4,822 Veg foodies

• World Explorers (HK)

5,367 Travelers

• Comedy HK

1,800 Comedy Fans

• Hong Kong Music Lovers

6,362 Music Lovers

• *❦ Coffee Break Language Exchange ❦*

10,640 Members

• Hong Kong Fun

10,415 Hongkong fun