Julia Community 🟣

Discussion on: Frequentist and Bayesian coin flipping in Julia

Collapse
getzdan profile image
getzdan • Edited on

Lovely article, and I'm just reading up on these topics. My question is regarding the Conclusion section, it says "around 30 samples..." and looking at the frequentist plot, I can see 0.5 being firmly within the confidence interval the whole process. Am I missing something? Did the comment refer to a different random seed?
(also, it seems Turing.jl isn't used, so it doesn't have to be imported. The Bayesian calculation is similar to one in Turing documentation on the section of calculation without Turing).

Collapse
rikhuijzer profile image
Rik Huijzer Author

Thank you! You are right.

I've switched from loading Turing.jl to only Distributions.jl. Regarding the confidence interval, I was using a Random.seed! instead of the StableRNG, so that broke things when I updated this post to Julia 1.8. It's fixed now.