Desert Fox wrote:Fancy Pants wrote:Stringer Bell wrote:You are completely wrong. A more useful data set might be '09 NLJ plus a 3 year average of clerkship data. You could also throw in a 3 year average of academia and make assumptions on prestigous PI too.
The point is that in the real world people do analysis with incomplete numbers numbers and assumptions instead of saying "eff it, I'm not looking at this until I have actuals." hth
The fact that "the real world" (whatever that is) uses incomplete data to make predictions really doesn't change how you should feel about using wrong data to make predictions.
Also, you mention "assumptions" but fail to realize that the whole point of what we're saying is that the assumption is that the 2009 data is going to be very different from 2008.
http://en.wikipedia.org/wiki/Subprime_m ... it_ratings
Oh come on. How can you expect anyone to build a model with the crazy idea that an asset could decrease in value?