We don't have reliable information on this, but IN THEORY the "right" way to do the rankings is just simply in raw market value terms. The players who sign for the most money are ipso facto the best.
I suspect any changes in the rankings reflect chatter along those lines, that the players got more or less money than was previously anticipated.
Until the portal closes on Friday I suspect the market is actually growing each day. After Friday though, it will shrink rapidly.
This goes to my point about the portal ratings/ranking. In what you describe as the "ideal" version of them, they are reactionary, not prospective (even though many ratings are released
before anyone knows where the players are going and there is any chatter about for how much), and what they are reacting to is "chatter" about how much players are getting, which means that to the extent there is any real data involved, that data is unreliable and unverifiable, and even if the "chatter" turns out to be right, it is only helpful to the extent that the highest bidder on a given player made an accurate assessment of that player's value.
And that's without even getting into other nuances that would frustrate this method of calculation, such as the differences in compensation between different positions. If you were really trying to be accurate, you'd wait until you had enough data on how much players were getting in a certain position group before you issued any rating. Of course, these ratings services are not trying to be accurate, they're trying to sell a product, and the sooner you can push these ratings out the better.
But I don't even think that's how it works. If I had to guess how it actually works for the transfer portal, here is my best guess:
1. 247 puts out ratings fast, and doesn't appear to really update them, because it's about volume not accuracy.
2. On3 puts out a compilation rating, which incorporates its own assessments and 247s. It's own assessment is what you see at first (hence they're all round numbers, like 87.00), then starts incorporating others, including 247.
3. Other less known ratings outfits do their own, and a lot of them probably have a "follow the leader" approach and end up close to 247. They're probably more limited lists, probably going off 247's top 100, 200, or so on.
4. On3 starts incorporating more of these lists, and you start to see guys who 247 rated highly drift closer to their 247 rating (hence why guys like Rusk, Bowick, and Durojaye all now have ratings that are not round numbers, and are much closer to their 247 rating than they were two days ago). Since a lot of the lists are not as comprehensive (see #3) you start to see this first with guys 247 has rated higher.
I think that's both the most likely explanation, and the one that explains the ratings changes I noted previously.
