I am in the middle of working on a book that deals with financial mathematics. The intention was to write the book without using any 'real' math, and when in doubt, throw any real math under the bus, as I am of the opinion that anything beyond a plus or minus sign makes most people jump pages, and I want it to be understandable by the maximum audience.
So here's the question. I am writing a chapter on 'Relationships.' Particularly the expectation of how one-month, intra-product, should move with another month.
Now, I always ran a regression with a few filters, took the output, and did my thing with it. Problem is, throughout my career, I have always, and everyone around me has always referred to these as correlations. Something I still see prevalent in most analyses. i.e., A moves 70% of B, then the correlation is 70, which entirely ignores they move in the same direction every time, so the actual correlation is 1. (For my purposes the 1 is irrelevant).
Not being much of a writer, nor a stickler for semantics, I never gave it much thought. Now that I am forced to write about it, this ‘feels’ a bit thorny for me, as any slip in credibility will be extremely damaging, I want to get this correct so I am seeking the advice of those more knowledgeable and opinionated. Is there some understanding that calling this correlation is acceptable, is this a different definition of correlation? Am I missing something? Do I just write one sentence addressing I am going to call it by the wrong name (as is my intention, and why I am seeking confirmation I am not going to get this wrong)? Is anything or everything I am saying nonsensical?
-TIA
Edit: Incorrectly typed towards the end of the 3rd paragraph that 70% of B is 1... Now corrected.
[link] [comments]
from math https://ift.tt/2HBvBrs https://ift.tt/eA8V8J
Post a Comment