Hacker News new | past | comments | ask | show | jobs | submit login
Mathe Carlo (observablehq.com)
62 points by tosh on Jan 5, 2020 | hide | past | favorite | 8 comments



Thank you for this. I've been curious about Monte Carlo techniques ever since I was strongly influenced by Taleb's ideas. He uses Monte Carlo techniques via Mathematica* to produce curves for very complex probability functions. He prefers using Monte Carlo (practice) over using math theory. (He calls it "tinkering" not "Pinkering", a reference to Steve Pinker, whom Taleb criticizes because Taleb argues that Pinker relies excessively on theory, not enough on practice).

My request to experienced Monte Carlo tinkerers on HN: please share "open source + free" tools and resources that you'd recommend to learn Monte Carlo.

*Mathematica is very expensive software, without a free tier.


Separating Monte Carlo methods from math theory is an odd distinction to make. There's a lot of sophisticated theory underlying Monte Carlo methods as well. MC also is more of a loose term referring to a class of techniques, rather than any specific algorithm. In that sense, there aren't really any [specific] tools or software you would use.

A textbook on MC applications to your area of interest is probably the best place to start, there are a ton in practically every field. If theory-heavy presentations aren't your flavor then you can likely find a text or blog series to match.


I wouldn’t call myself an “experienced MC tinkerer” but I liked this series on MC using Python https://towardsdatascience.com/monte-carlo-simulations-with-...


Interesting, where does Taleb discuss this?


Does any one know something similar as a primer on Markov Chains? I get the idea of simulating the pattern of behaviors or states with the Markov chain, and how the Monte Carlo is useful for that, but how to actually make a Markov chain and then test it mathematically with the Monte Carlo is something that I don't know how to do.


Hey, author of the Mathe Carlo essay here.

Markov Chains is on my list of things to write these kind of explorable essays for, good to know there's some interest for it!


Funny coincidence! Literally just made my own write up about Monte Carlo methods! https://www.towardssoftware.com/blog

It’s not as long and detailed, but it’s still very interesting to see how you can derive pi with random data!


Well done!




Consider applying for YC's Spring batch! Applications are open till Feb 11.

Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: