Programming Leftovers

posted by Roy Schestowitz on May 18, 2025

• Ken Koon Wong ☛ From Complete Separation To Maximum Likelihood Estimation in Logistic Regresion: A Note To Myself

  Look at the difference in the summary output. The estimates are much smaller, and the standard errors are reasonable. The residual deviance is also much larger, indicating that the model is not perfectly fitting the data. Also with lower iterations.

  This made me really curious how does glm find these coefficients to begin with? Yes, I’ve heard of maximum likelihood estimation and I know that it uses that to find the estimate and standard error, but… how does it actually do that? 🤔

  Also, if we have a perfect prediction, shouldn’t our standard error be very very small instead of very very big !?! Maybe the answer lies in how these coefficients are estimated!

• [Old] Boyd Kane ☛ Experts have it easy

  Something that’s painfully understudied is how experts are more efficient than novices while achieving better results. I say understudied and not unstudied, because it’s common knowledge that charging people for their time results in experts being paid less since they work faster, which is why experts charge more for their time.

• [Old] Jacob Tomlinson ☛ Most stale bots are anti-user and anti-contributor, but they don't have to be

  If you’ve been around open source projects on GitHub you may have encountered a project with a stale bot.

  Here’s how a common stale bot interaction goes; You’ve found a problem and you open an issue, but nobody responds. Then 30 days later you get a notification from a bot saying “Beep boop, there hasn’t been any activity here for a while. I’m going to mark this as stale”. Then a month after that you get another notification from the bot saying “Closing this issue due to inactivity”.

  In the meantime you’ve either worked around the bug or pivoted entirely to avoid it. As a user this experience sucks.

• Rlang ☛ spuriouscorrelations: An R package to show examples about spurious correlations

• Rlang ☛ From Complete Separation To Maximum Likelihood Estimation in Logistic Regresion: A Note To Myself

• Perl / Raku

  • Arne Sommer ☛ Average Backspace with Raku - Arne Sommer

    You are given an array of numbers with even length.

    Write a script to return the count of distinct average. The average is calculated by removing the minimum and the maximum, then average of the two.

• Python

  • SANS ☛ xorsearch.py: Python Functions, (Sat, May 17th)