🇳🇱
People never seem to realize that a great many of the logical fallacies they engage in every day could be avoiding if they just understood Regression to the Mean!
@freemo I don't know what this means but it sounds easier than memorizing everything from here:
https://changingminds.org/disciplines/argument/fallacies/fallacies_alpha.htm
@hrisskar It basically means if your responding to problems by being reactive, you probably arent solving anything at all.
More specifically if you put your attention on the worst examples of a problem, and try to fix those first, you cant accurately judge if you fixed the problem at all.
For example, lets say you are in charge of pool safety for a whole country. If you take the 10 pools int he nation that have had the most drownings, and enact new policies at those 10 pools to help reduce drowning, and then suddenly you see those pools have a decrease in the number of people that drown, you can not make the claim that your policies were successful in fixing the problem.
The reason being that the most extream cases will have the strongest tendency to "regress to the mean" so they would have almost certainly seen a reduction in the number of drownings even if you did nothing at all.
@hrisskar Not the same, but can be related, especially when your focus is on those who survive but came closest to dying.. For example if you look at only the planes of the survives with the most bullets, and then you tweak something about the plane and find they have less bullets the next go around.
@hrisskar This might be it, I could have been mistaken though: https://youtu.be/1tSqSMOyNFE
@freemo Mmm, brain candy. 🤤 Thanks for the link!
@freemo I remember that example from the following:
https://youarenotsosmart.com/2013/05/23/survivorship-bias
I wish this blog kept posting more like this!
@hrisskar As I recall there was a veritassium video that talked about Survivorship bias in relationship to regression to the mean.. let me find it.
@freemo >More specifically if you put your attention on the worst examples of a problem, and try to fix those first, you cant accurately judge if you fixed the problem at all.
This reminds me a bit of Survivorship Bias, though I doubt they're the same thing at all.