I don't make adjustments. That would defeat my entire basis that adjusting based on opinion is unnecessarry. The numbers WILL take care of themselves. Some may take longer than others but they will get there. Assuming that they are still as good as they have been.
I started my rankings 3 years ago when, as a Newbie, I was complaining about the rankings and was told that if I didn't like it, make my own. So I did. My point was that I could derive a simple calculation and let the calculation determine rankings based on the actual scores of games. It doesn't look at strength of schedule, it doesn't care what you did more than a week ago. It looks at the last game and makes an adjustment based on that.
The reason for the simplicity is this. Yes, team X may have beat team J in week 3 by 17, but team X had an injury which makes them not as good as they used to be. So even though X beat J in week 3, can they beat them again in week 7 with little Johnny Allstar sitting on the sideline with his ear in a cast? According to their past 2 weeks.... It don't look like it. So the simple fact is this. Teams get better, teams get worse, and this happens week to week. Injuries, move ins, move outs, coaches trying something new, and a big one, report cards after week 6 and 12, teams learn as they go and get better, then some kid moves to Alabama with his uncle, so now they are worse, so you put the sophomore in there and after a couple weeks, turns out he is better than the kid that moved. This stuff happens every single day. I wish I were smart enough to have a formula for it, but I went with the one that is realistic.
My respect for the Guru, Mr. Grainger Huntress has grown exponentially because of me doing this. Now I understand why he is the Guru. Why his formula (whatever it is) is second to none.
I was going to not do it anymore, and then I got an overwhelming response asking that I continue, so I will. But I will NOT be changing my formula or making adjustments. Even when I know as well as you do that SA Castle Hills is rated too low.