When I first started working, I listened in on a few engineers discussing a problem involving predictive modeling of a system with many nonlinear variables. The junior guy was very confident in success, the senior guy simply told the junior “good luck with that.” I dont think they ever got it working as promised.

Now that Ive gotten a bit older, I can see some systems are simply too complex with too many unknown variables to properly model beyond very wide degrees of confidence. You can decompose some of the unknown variables with regression analysis, but there are limits with that to.

One thing I realize is those who dont really understand or have never practiced real science see it, or what passes for it these days, as a religion. They dont really understand it, so they attribute magical power to it. Those who actually do it see science simply as a powerful tool, and will be the first ones to admit its limitations.