re: The Big 10 and ACC are setting the stage for locking up 3/4 playoff spots

quote:

---

That's inherently subjective. To have any amount of objectivity in data analysis, we analyze the data we have

---

I completely get where you are coming from here, but I do have a disagreement.

The formulas these data points are incorporated into are also to some degree subjective and serve various purposes with their construction, and while these cross purposes might be fine in a normal season as things normalize and converge, they are going to suffer, I fear, in a year where sample sizes are very disparate between teams.

Like Sagarin for example was tuned to offer predictive power about which team would beat other teams, which says nothing about quality of resume.

I’m simply saying I think I probably disagree slightly with the formulas used in this context. This particular disagreement isn’t likely to manifest very often outside of 2020, but it is a concern I’ve got. Again I’d point to the fact that people usually tend to laugh at how ridiculous computer rankings are until halfway through a season because they get kind of weird without sufficient data.

Perhaps we can track how many games it takes for computer rankings to mostly converge in a boring season without many updates and create a minimum game threshold based on this to maximize faith that the computer rankings have expelled noise.