Like my friend Captain Foresight said, you should have taken at least a QB in 2012. 
—Captain Hindsight

It's been four classes since I played the Captain Hindsight game, where we go over a list of Michigan recruits going back as far as I can find crutin information (Lemming and Parade All-Americans and Sandeep's old page), and then pulling from stats and starts and awards and draft position and memory to give each guy a "results" Star rating.

But this time instead of just 1-5 stars, I quartered that to fit the same ranking system I came up with last week as a composite rating. That is…

Seth's Rating System:

And here's the results of my re-ranking survey. Please (and I'm serious about this) lodge all questions and complaints about rankings in the comments. I plan to take them all into account and adjust. Or if you want to download it and make your own rankings I'd be happy to take that. This is a feels thing so the more input the better our information. That said, unless you think I'm way off with the bulk of guys, please preserve my fragile ego, since I'm putting the sum total of my Michigan fan knowledge into those numbers and would like to continue thinking all that attention over the years hasn't been for naught.

Notes on these: Since this is just judging talent scouting, anyone I could possibly rank (including the transfers) I did so. Those not ranked were injured before we got a chance to see them on the field or compare them with players ahead of them on the depth chart.

Also to handicap things for scouts this is not about who ended up being the best PLAYER but accurately representing a guy's talent and ability to convert it to footballing. This is NOT to say every 5.0 was better than every 4.25, because some truly great players who went on to long NFL careers weren't able to help out until they were upperclassmen. I did it that way because I know the ranking systems themselves judge a player by how college-ready he is, necessarily underrating ceiling. There's no skill that would let you see a 220-pound tight end and predict he'll be the NFL Draft's first OT taken in five years. Long careers therefore can catch up to loftier ones, and the top overall groups are guys who had both.

I'll repeat that just so we can shame the guys who didn't read it in the comments: it's not about who's BEST but how accurately he was scouted.

[After the jump: we compare services, and find fun things like best class ever, most underrated guys, etc.]

------------------------------

Money Chart: What was the best class?

Can't judge 2013-'15 yet but few surprises. Those 2010 and 2011 classes though: woof! 2010 especially, even if Rich Rod was recruiting under the silly freepgate cloud, had a lot of scouting whiffs. Late Carr also saw a lot more misses than his usual, which was usually excellent. The 1997 team was built by some magnificent finds in 1993-'96.

How accurate were the recruiting services?

The r-squared was 13.3%

The way things shook out you could expect an average consensus 5-star to be at least very good, one in the top 100 or 250 to be at least serviceable, and not much difference from 3-star on down. Here's the standard deviations:

• Rivals: 0.90 stars
• Scout: 0.92 stars
• ESPN: 0.93 stars
• 247   : 0.79 stars
• 247 composite: 0.86 stars

Being accurate to within nine-tenths of a star ranking doesn't seem that helpful, but you're dealing with humans.

The thing I found most interesting were the three tiers:

Consensus 5-stars (the kind of guy who's in the top 100 and probably picked up a 5th star on two sties) usually turned out to be very good players. Then you have a group of Top 250 to 4-stars and high 3-stars whose expectations fall at about a serviceable upperclassman. The rest are indiscernible until you get below 3-stars.

Some things popped out, like Rivals consistently overrating their "6.0" guys, and 247 (their own rankings not the composite) coming out more accurate than the others. I think that's an affect of existing not as long: all of the sites got more accurate as the years progressed, and by being the last ones in the 247 ratings got to skip that 2002-2006 period of major recruiting mistakes.

Also I found ESPN's rankings past the top guys didn't match the rest, a sign perhaps that they rank and forget while the others are more diligent about parsing 3-stars from almost 4-stars.

Most underrated?

Most overrated?

Checks out.