Monday: That doesn't make sense.

I was able to get it to partially work, but the thing I've been avoiding at work while reading up on the code still isn't finished.  I can't figure out how it's doing a mapping, and I don't see any place that's defining the mappings it is getting right.

Tonight was also board game night, which I came out of with a pretty decent 50/50 split.  I could have won one of the games of Sabotage if I'd waited slightly longer to betray everyone else.

It's muggy and hot, but I should probably make this again once it gets cold.  I think some sort of pork would help flesh it out into more than just a side.

• This was my local grocery store.
• Yeah.
• So Baylor is just the absolute worst.
• Am I sufficiently not lazy to do the math for this?
• From this data, you can assume that this is reasonably described by a three-component mixture model (using Gaussians because I'm lazy and they're good enough).
• Feeding the individual values to my GMM code yields all modes with approximately sigma = 1, and means and mode fractions:
• F: m = (1.89, 6.74, 9.12); P = (0.04, 0.49, 0.46)
• M: m = (1.36, 6.28, 8.85); P = (0.23, 0.56, 0.21)
• The bias for males is then something like a half point.  It's smaller for the highest ranked mode, but that might just be because of the fact that the range and sampling is limited.