Meant as a prompt for discussion, of course, not for any one specific answer. Share your thoughts!
-
-
Prikaži ovu nitHvala. Twitter će to iskoristiti za poboljšanje vaše vremenske crte. PoništiPoništi
-
-
-
I mean, the standard deviation for a number of coin flips is sqrt(N)/2, so I would set a threshold of say 3 or 4 standard deviations, and if exceeded that I would get suspicious. 7/10 isn't suspicious. 70/100 is 4 sigma, that's suspiciousish. 700/1000 is nearly 13 sigma. Rigged.
-
thats what I thought, but if you flip one coin and it comes 700/1000, its not so crazy, yes, its suspicious, but statistically you can have that proportion I mean, 13 sigma it's a heck low probability, but one experiment doesn't mean the coin is charged
- Još 1 odgovor
Novi razgovor -
-
-
What if I flip a sea turtle?
-
If you flip a sea turtle 2 times, can it still walk? What if you flip it 700 times? And if you flip it an even number N of times, N big enough, can it still go? At what N the turtle simple give up and tell you to get a life?
- Još 1 odgovor
Novi razgovor -
-
-
Are we talking Bayes or Frequentist? If Bayes, plug it into a Beta distribution and you will see that the more observations you get, the narrower our credible interval becomes.https://www.desmos.com/calculator/tvrt09gxrm …
-
Python implementation from scratch that finds a high credible interval. For anyone interested in this subject, be sure to read "Bayesian Statistics The Fun Way" by
@willkurt and "Empirical Bayes: Examples from Baseball Statistics" by@drobhttps://gist.github.com/thomasnield/d770b6afa908bddeffd4fe3d22e23ba3 … - Još 1 odgovor
Novi razgovor -
-
-
I think we need to pick a p-value.
-
Or a prior distribution. Which could be different for coins and turtles, depending on what we know.
- Još 3 druga odgovora
Novi razgovor -
-
-
I have to admit the sex ratio of newborn sea turtles would have to be quite skewed before I started considering it unfair.
-
Thanks Notch, Very Cool!
Kraj razgovora
Novi razgovor -
Čini se da učitavanje traje već neko vrijeme.
Twitter je možda preopterećen ili ima kratkotrajnih poteškoća u radu. Pokušajte ponovno ili potražite dodatne informacije u odjeljku Status Twittera.