One of the problems that science has created for itself is the overuse of statistical modeling. This approach to science, allows science to cheat compared to rational modeling, since the former allows fudge factors. If someone found an exception to a rational model; even one data point out of place, the model would need to be revised. If space-time did not change with mass everywhere on earth, General Relativity would be subject to revision.
With statistical modeling none of the data points have to touch the curve, since the curve was formed by averaging experimental data and then using fudge factors. This is a watered down standard. The rational approach has to first build a logical conceptual framework, then create the math, and then prove it with experiments, with the data all needing to touch the curve; Newtonian Gravity.
This higher standard would kill most statistical based theory, such as evolution. However, the watered down standard allows even half baked theory too linger and not be forced to revise, since the fudge can used as a defense for the bad fit.
One of the worse application of this watered standard is risk analysis. According to this second tier science approach, we had the risk of getting COVID. After several years, the data shows that not all people caught the virus. They never had any risk. That was all based on fortune telling and not science. Why does this theory still if it got so many bad data points out of place? Risk is based on the emotion of fear. Emotions are not useful for rational thinking, but it can allow one to use emotions to fill in the curve for you. To appease the fear you will need to accept the theory thereby preventing the trash can where it belongs.
Say a mother tells her son to clean his room. She come back several hours later and the room is still a mess. She is mad and asks him, why didn't you clean your room? Her son says, yesterday I was talking to Joe and he said he may be able to borrow his dad's car and drive the gang to the beach. The odds looks very good to me, so I was preparing for the breach, since Joe does not like to wait. The mother looks at him and says, well how can I argue with science and statistics. Bull crap is able to grow legs.
Casino science needs to come with a disclaimer that it is useful for emotional thinkers; risk and jackpots, but is a watered down standard and that any broad based conclusions that result need to be taken with a grain of salt. It is a stepping stone approach that is useful until someone can use the more advanced rational science approach and get a solid correlation that does not need fudge.