Or, more specifically, I think I fixed the last big problem in the stuff at work, and it turns out to be a logic error. The math is due to the fact that I was going to do some math stuff here, but my test code doesn't quite work due to the fact that I have quantization issues in my images, and I'm too lazy to put in a proper way to handle that. I think a drizzle like algorithm will do the trick, as I don't need to have insane precision, just enough to smooth out the bumps.
|
"Damn it, I forgot the burgers!" |
- I think Krugman is misunderstanding the Princeton team's two predictions. The today number is the current snapshot. The election day one is a "regression to the mean" one, which attempts to smooth large outliers by assuming things trend to the Gaussian mean of one of the parameters. Since the news yesterday was not just a random discrepant poll but a systematic change in the Kansas election, there isn't enough time for that shift to be significant in that estimate.
- Also, this jackass, and his wife, are both going to jail for a long time.
- I went googling for a recipe that I saw a gif of, and wanted to know what they were doing, but instead, I found this. I'm calling it a win.
- Friendly squirrel.
- Sleepy squirrel.
- Boom boom boom..
No comments:
Post a Comment