Hey Allbuffs (long time lurker, first time poster). I hate to start this up again, and maybe this is due to my ignorance on how the ratings work, but can someone explain why Wray and Lewis recently dropped in the rankings and lost their composite 4 star rating on 247? I mean, Wray dropped 670 spots on 247...
I'm going to be the sucker who thinks there's a chance you are not someone's sock trying to be funny and give you a real answer.
The recruiting "ratings" are set by curve. There's pretty much the same total number of 5* and 4* recruits every. single. year.
Ask yourself: does that match your observed reality of how football talent ebbs and flows?
The NFL has good draft years, and bad draft years too - but every single year there are the same number of first round picks. We all know that there are some years that a lot guys go in the second round that would have gone in the first round in other years.
If recruiting rankings were "objective" from year to year, there would be a lot of variation in the number of 5* recruits each year. But every year, there's about 32 +/- 2 5* recruits.
That should tell you something. If a scout scores a player at, say a 47.5 on whatever scale they use, some years that might be a 4*, and some years it might be a 3*. Let's say it = a 4* when they put it into the database. Let's also say that there is another player that gets scored at a 47.3, and that right now, given all the scores in the database, the cutoff between a 4* and 3* is 47.4. Well, the second guy is now a "high 3*."
Now, another scout does an evaluation of the second recruit, and he scores the second player at a 48.1. If they average the two scout's scores (I don't know if they do this, or if they give more weight to more recent evaluations), that means the second guy now gets scored at a 47.7. Which is better than the first guy. Guy #2 is now a 4*.
Except the fact that we now have a player with a higher score means that the curve gets moved too. Now the cutoff for 4* is at 47.6. Boom, guy #1 is now a 3*.
It's even possible for no new scoring to be done on guys 1 and 2, but instead a whole bunch of scoring got done on some other guys that moved the curve up or down - viola, both of their ratings got changed, without there being any new information specific to them, their raw scores didn't change at all.
The main point to understand is that recruiting ratings are not a raw score. They are a rating that is created based on their raw scores
in relationship to the raw scores of all the other recruits.
When people move "beyond the star ratings" to ratings like 5.5 or 5.7, or 91 or 89, they often think that these numbers represent a raw score. They don't. These are still ratings (albeit more granular than star ratings) that are based on some sort of raw scoring that is then compared, normalized, curved (whatever verb you want to use to describe it) to the raw scores of all the other recruits.
So, what can move a recruit's rating?
Updates in their raw scoring based on new information.
Updates in their raw scoring based on a different scout scoring them a little differently.
Updates in their raw scoring based on a scout taking a much closer look and doing a re-evaluation (yes, a P5 offer can trigger this, and yes, a "blue blood" offer can trigger this too).
Them actually getting a real evaluation and a true raw score instead of a basic score assigned by an algorithm (yes, P5 offers often trigger this, and it's why 2* often move to 3* after a P5 offer).
And of course the big one:
Any of the above happening to
other recruits and the curve getting shifted as a result.
If you're in the high 3* and low 4* range, there's going to be a lot of movement in the recruit's ratings, and it's easy to tell yourself conspiratorial stories about why that movement is happening.
Duff's response is the right one: at the end of the day, if you fill your team with this level of recruits, you're going to win a lot of games, so don't worry about odd movements that you see at this level.
What you worry about is filling up with the low 3* - they're like C students in an old-fashioned grading curve: there's a whole lot of them, and while they each (presumably) have some great aspects as individuals, as a group they're... average.
