On Tue, Jun 14, 2016 at 8:16 AM, Eric Chadbourne <[email protected]
> wrote:

>
> I'm reading up on being "unfalsifiable" now.  Fascinating


​Indeed. Science is based on hypothesis-testing. Only a "falsifiable"
hypothesis is testable.

If evidence against a conspiracy theory is taken to show how effective the
conspiracy is,​ the hypothesis in unfalsifiable.

If the statistical model is guilty of over-fitting (too many degrees of
freedom aka too many parameters), the model is non-falsifiable in the
short-term. (But eventually enough data will show that adding 5th order
epicycles is guff.)



-- 
Bill Ricker
[email protected]
https://www.linkedin.com/in/n1vux
_______________________________________________
Discuss mailing list
[email protected]
http://lists.blu.org/mailman/listinfo/discuss

Reply via email to