Quote Originally Posted by myles View Post
Thoughts on this style:

Attachment 10063
That one I really like. Big picture and time based. And as time moves on it should give increasing information on default variability/predictability. If you could show which risk model was being used for each cohort so much the better.