Spooky I am talking about actual talent as shown again and again on the field by play, in the games, this year. Game starting rosters. So no forecasting at all about it. This team did poorly in post season awards because they earned it with their play in games this year.
As for using supposed pre-season "talent" Phil Steele does it well. He does not include star power for players not on the team for example (because duh guys not on the team never make good contributions
). Returning starters are worth more cause on average they outperform the others. His predictions like all predictions will become less and less valid as important players miss time for whatever reason throughout the year or there's a 3* Johnny football surprise at QB.
I share your preference for 247 ranking.
But signed talent is less and less useful as a proxy for actual talent on a team now, over time (one team retains and develops most 4-5 starts and another gets low performance out of em (poor for NU's 2011 class...Texas classes 2010 2011 and 2012 IIRC their NFL drafts hit the sh*t*r the last few years and that is mostly to do with the the old coaches recruiting not with the current coaches coaching).