In fact, when I am looking for "bullshit-free" explanations of statistics concepts, I am emphatically not looking for a resource that reads like lecture notes from a graph theory class.
Maybe a black box explanation would suite dev better.
I.e. Use this method to work out the probability density function from some data and these are its caveats and how to spot them.
Thank god this explantion doesn't have any of that bullshit in it.
The link to stackexchange in the comments is way more enlightening.
Then you misunderstood the complexity of the problem. It's okay, I did too when I first learned about MCMC, and I proposed the same algorithm as you did (in fact, your proposed algorithm is right there in the fifth paragraph of the post. Did you read it before commenting here?). The problem is that if X is exponentially large, and many of the probabilities are exponentially small, your algorithm takes exponential time to draw even a single example.
> The way MCMC achieves this is to "wander around" on that distribution in such a way that the amount of time spent in each location is proportional to the height of the distribution.
In fact, yes, I read the fifth paragraph, but didn't quite recognize what I proposed as a "solution" earlier before you pointed this out.