On skills-based subjects, end-of-year testing is fine. Basically you're going with Julian Stanley's DT/PI (diagnostic testing followed by prescriptive instruction) model. Find out what they know and don't waste your time on that. Find out what they don't know and teach that.
I agree with you that I'd rather see 90% proficiency on a skill. The 80% is probably coming from Winebrenner's book, and it's the only thing in the book I disagree with. I taught math. If I could insist on 95% or 98% proficiency, I would. My feeling is that if you don't have really solid skills in the skills-based subjects (math, grammar, handwriting, foreign language, reading decoding (that's by contrast to the content-based subjects -- science, social studies, literature, written composition, etc)), you are eventually going to get overwhelmed and bog down and be miserable.
The easy-to-difficult scores are related to W scores, which are basically linearizing item difficulty as measured not by the items themselves but by how kids at different ages in the norming sample did on them. That's also where the RPIs come from -- basically, the easy-to-diff are based on subtracting or adding a certain number to the W score, while the RPIs are based on comparing the W score to the typical W score for that kid's age.
Grade equivalents in general tend to be stupid for a large number of reasons (W scores attempt to reduce some of the stupidity but can't solve the underlying problem that development is fundamentally nonlinear), but educators often like them. I personally rarely put them in reports if I have a better alternative (which on the WJ I always do -- I sometimes use them in specific situations for other tests). I prefer to use RPIs and GRPIs (turning an RPI upside down to reflect relative proficiency when the kid is a lot *better* than the typical kid their age at the measured skill).
I think there's a technical bulletin on the Riverside website that explains W scores and RPIs and such -- check there on
www.riverpub.com. It's really technical, but hey, you asked a technical question.