"You will notice a JR at 20 is likely going and a SO at 20 is OTF. "
YES...because CLASS helps determine outlook AND position on the board.

Would you feel better is I changed this to :
Id say that the following occurs in this order "
First : class + ratings = outlook
Second : ratings + outlook = position on BB
Third : outlook + position on BB = probability of EE

Again ...chicken and egg on which happense 1st/2nd. But for your original post you really only care about that last piece (probability of EE). The rest is just discussion while we are stuck at home. Anyone want a conference call ? lolol

"I think the core of the small disagreement here is when you said outlook plays a part in determining position on the big board."
Thats one possibility on chicken/egg arguement. But you are right, I more firmly believe below.

Id say that the following occurs in this order "
First : ratings + class = position on BB
Second : class + position on BB = outlook
Third : outlook + position on BB = probability of EE

This is fun stuff, I like talking through these things.

I think a little bit of what we're saying is just overlapping and is just how we're saying it.

Id say that the following occurs in this order
First : ratings + class = position on BB
Second : class + position on BB = outlook
Third : outlook + position on BB = probability of EE

I agree with this. I think you could just bypass the last step as the outlook is more meant to just help the player (I think). However, there are very large changes in probability of leaving within the outlook and it's definitely not linear. For example, #28 Soph Likely Staying has about 40% chance of leaving while #40 through #100 Soph Likely Staying has about 10%.

So that's one of the big takeaways here. You can see a big drop within an Outlook but it doesn't go to zero, it basically drops then stays flat through the majority of the big board.

Well the parts we are left to disagree with I can only provide my opinion because ADMIN wouldn't serve up ALL the details on how this works.

"However, there are very large changes in probability of leaving within the outlook and it's definitely not linear."
I ran 2 separate models for about 5 seasons each. One assumed the chance to leave went down the lower you were on the BB (linear), the other assumed what you have above. The problem is with limited data and NOT seeing what the engine "rolled" for each player I honestly dont know for sure which is correct.

Maybe you can hook us up and check with ADMIN?

Ha, I know how that road goes. You'll get an ambiguous response that raises more questions than answers.

But I'm pretty confident in my assessment. I'll continue to collect data but right now I have 440 Sophs listed as Likely Staying and 577 Juniors listed as On The Fence so I'm using those sample sizes for my conclusions. I doubt the needle will really change beyond a couple % points if those sample sizes were doubled but might as well keep tracking.

Confirmed by ADMIN :

Id say that the following occurs in this order
First : ratings + class = position on BB (cubcub is correct. I think I already knew this but had not really talked/thought about EEs for a long while) )
Second : position on BB + class = outlook
Third : outlook + position on BB = probability of EE

I'm glad you two called me out on that. Not really sure what I was thinking.

Posted by mullycj on 4/22/2020 3:35:00 AM (view original):
Confirmed by ADMIN :

Id say that the following occurs in this order
First : ratings + class = position on BB (cubcub is correct. I think I already knew this but had not really talked/thought about EEs for a long while) )
Second : position on BB + class = outlook
Third : outlook + position on BB = probability of EE

I'm glad you two called me out on that. Not really sure what I was thinking.

Ah nice, thanks for finding out for sure.

Okay, so what was the original question?