Anger, disinformation, populism, the sorry state of the world right now... and truth
This post should be seen through the lens of self therapeutic writing, as I angrily work through a few ideas on my mind.
And now, for something completely different.
Julia - the magic of composability
I've been getting into the Julia ecosystem a little (turns out there's a lot of tutorials you get to watch when doing the dishes), and I've been meaning to get some hands on experience. Today, I'd like to reproduce Mosè Giordano's gist example, which I first saw in this JuliaCon talk by Stefan Karpinski. I think it really illustrates Alan Edelman's points on Julia's composability from the talk I linked last time.
What we'll do today is combine a simple example of an ODE solution with an uncertainty package, and see how easy or difficult it becomes using Julia.
Short updates: Bayes, Master's, PlasmaPy, Protoss, Julia
No real new material today, as I'm still in the thick of it. A few updates:
Parsing SC2 replays for later analysis
I've realized I owe you an explanation on how to parse your own SC2 replays for the series of posts on Bayesian SC2 replay data analysis. Let's go through it here!
Learning my per-matchup MMR in Starcraft II through PyMC3
In this post we'll continue our SC2 replay research, started last time. You may want to go back to that and pick up on the terminology!
To recap: we used replay data from my SC2 games over 2019 to estimate a "true MMR" value and infer the size of per-game fluctuations. This time, we'll redo that analysis, except to get something more useful: we'll look at the three matchups I played and infer separate MMR values for each of those. Let's dig into it!