Crystal Ball Tips

Submitted by Steve Lorenz on

Hey guys. Hope all is well here, and hope you've enjoyed the radio/podcast stuff. (darts talk). It's been a lot of fun. 

I lurk the board relatively frequently still, and noticed some threads being started based on a couple Crystal Ball picks or even one pick in certain instances, so I figured I'd post some pointers for those that do follow it, or want more clarity on it. 

First things first: it's not a perfect tool. It's a great one, and almost always gives an accurate glimpse into a recruitment, but like most things involving recruiting, has its flaws (late flips, uninformed analysts making predictions like the Chris Evans one, etc.). For those of us that take it seriously (most users), it's a tangible way to put our names behind where we think someone will go. A couple quick things I'd consider regarding it: 

1. Our national guys (Steve Wiltfong, Barton Simmons, etc.) won't usually put an initial pick in unless the information is strong in one direction. Steve's prediction on David Long today is a good example. When we expanded the ability to choose to outside sites, it opened up the door for people to make completely speculative picks based on nothing (the Chris Evans pick/thread is a perfect example). Usually the given team writer's pick + national analysts almost always = probably what's going to happen. There have been a few instances (Damon Webb) where that didn't happen, but for the most part it's solid. 

2. We're measured on a couple different thresholds regarding the CB...which are basically the percentage of accurate picks, and the average length (days) our average prediction is worth. So while flipping a pick late to get the pick right does help the accuracy average, it hurts the average prediction length, so we're pushed to get it right as far ahead of the commitment as we can. 

Figured this may give people a better idea on how to view the Crystal Ball going forward and heading into Signing Day. Again, hope all is well here. 

 

FormerlyBigBlue71

January 6th, 2016 at 9:25 PM ^

Even if we get Gary, Long, Hill, and Crawford we are still likely going to finish in the 4-6 range on 247. Mostly due to the other teams having multiple 5 stars. No fault to Michigan as their is just not enough 5 star kids in the region. Still will be the most talented class on paper I have seen Michigan bring in.

M-Dog

January 7th, 2016 at 1:59 AM ^

I like Les because he bleeds Michigan blue, but LSU recruiting keeps him employed more than his coaching acumen.  

Had he come to Michigan, he would not have been able to replicate anywhere near what he did at LSU.  And forget the notion that Louisana kids would follow him up here, they wouldn't.

 

IncrediblySTIFF

January 6th, 2016 at 9:34 PM ^

Which is the Crystal Ball a bigger tool for: the website making money off clicks or the "picker" making money off correct predictions? for the record I support both causes

SwitchbladeSam

January 6th, 2016 at 9:52 PM ^

So, does this mean if you predicted someone to go to Michigan 300 days ago, and then 1 week out, you learn there's a 90% chance he's going to Slippery Rock... It's still a better bet for you to not switch because of the way you're measured?



Sent from MGoBlog HD for iPhone & iPad

LSAClassOf2000

January 6th, 2016 at 10:12 PM ^

Thanks to Aquaman for stopping by, first and foremost, and providing this insight.

I really was not aware how the Crystal Balls on Rivals worked, to be honest. I assumed it was just something that the analyst go to do as part of their role in providing information, so it was neat to see there's a bit more to it and the ratings than that. Interesting indeed.

Farnn

January 6th, 2016 at 10:41 PM ^

247 needs to stop counting an cloudy/unsure pick as a wrong pick.  Currently there is no reason to change an old pick that you don't have faith in because the foggy ball is always a wrong pick.  And votes on the day of announcement shouldn't even be counted as they are just used to pad success rate.

HHW

January 6th, 2016 at 11:32 PM ^

Can someone statistically and realistically smarter than me do a diary on the accuracy of crystal ball picks over the past 5 years?  With data points maybe at 12 months, 6 months and 1 month from signing day.  I'd be interested to see how accurate they are.

MeanJoe07

January 6th, 2016 at 11:33 PM ^

Crystal balls aren't 100% accurate predictions of the future!??!  WHAT!??!?!  My reality is shattered!!!!!  OR should I say . . . My crytal ball HEY LOOK A PUG EATING BUGLES!!!!

C-Bar

January 7th, 2016 at 12:12 AM ^

I've been a long time reader but only recently gotten an account but it seems this site has launched quite a 'tree' of its own with various past contributors moving on populating various other sites. (not to mention some of the more colorful account stories in the past like RDT, etc.) Is there some link or 'about' somewhere with the past genealogy or history of mgoblog somewhere?