I think some sites only score 20 players, so if you have 25 and lose 5, you only lose points if is player 1-20 and nothing if it is 21-25. And then it is the difference in ranking. If you have 20 and lose 2, then you lose all the points and it is a much bigger hit.I think you may be so invested in debunking "this horrendous exercise" that you missed the point. I suggested that if you have a 20th ranked recruiting class, but a third to half of them leave, your class is probably no longer deserving of the ranking of 20 since the talent that "earned" the ranking left the team. I think that is self-evident. But perhaps I am missing something.
Your beef seems to be about the methodology of measuring how far the ranking would fall. For what its worth, if one was going to attempt this exercise, then yes, it would be necessary to adjust Oklahoma's 2018 ranking. In addition, you would have to do every teams "attrition ranking" with the same methodology and then see how the numbers fall. This whole thing is of course a mostly meaningless exercise/ But, it is the off season. We have time.
That would also have to be considered in scoring for every team, if not done above. I would bet in the above, they just scrubbed all the points and it isnt even accurate to start with as they would then need tonfillnguys back in.