Dan Luu

I thought at first that you were presenting mostly-original analysis and didn't realize that the actual stats you cite were (it sounds like) taken from a much more detailed model than the one you present

Anyway, it's a totally legitimate move to write a completely caveat-free analysis as an example and then be like "please take your nitpicks to the original source, I promise it addresses them" + maybe this would have saved you time defending from nitpicks?

Oh no, is the appendix new or did I miss it the first time around? If I missed it, that's really embarrassing since it basically addresses the thing about variance :\

Yeah I haven't edited the post yet :-). I've noticed from talking with folks that pretty much no one reads appendices, so maybe I should label them as something else? Not sure what I should call it.

I've tried mentioning that I'm doing a simplified analysis in other posts. It's possible that this forestalls some nitpicking, but people often will attack the exact thing I mentioned. I get so much static on https://danluu.com/new-cpu-features/ … from people who haven't read my 2nd sentence

I really don't like putting disclaimers up front since I don't think of those as "content" and it pushes content further down, and it also doesn't seem to work? The most frequent comment on that CPU post is "this guy doesn't know anything about CPUs, X was around before the 80s"

Hmm. For me, I think pushing the content down by 1 sentence is a small price to pay for making that content much easier to parse. I read the entire baseball example with "so when is he going to talk about variance/hitter skill?" in the back of my head

Also, what happens if you try to optimize for "people I know are good readers understanding" instead of "minimize Internet static"? I'd worry that the latter is too noisy (hah) for most improvements to be noticeable