While flying back from our vacation, I was thinking about Pearson Correlation between pairs of stocks. I never did like it much. There must be a better way, eh? What I'd like to know is: "If the stock X price goes up (or down), will stock Y go up (or down)?" The usual measure of correlation don't do it. Indeed, check out the following $1K investments: See? Both stocks tend to move up/down together, eh? So what "correlation" do you expect? In fact, it's -1 (or, as some might say, -100% correlation). That's because the usual correlation does not measure whether the stocks move up or down together. It measures whether they move up or down relative to their mean return. Do I care about their relationship to their mean return? NO! Okay, so I whip out a spreadsheet and consider a different kind of "comparison" between a pair of stocks -- one that'll measure whether their returns tend to be positive or negative together. I'd also like to have this "fancy correlation" represented by a number between -1 and +1. So I start thinking. Then I realize that it seems so familiar. Then I realize that I've done this before, here. Then I realize that my senility is showing ... again! Then I realize that I definitely need more sleep. Must be the time difference and jet lag and all that cruise food ... P.S. Note that, for the chart above, gSC = 79% |
Monday, May 31, 2010
gSC
Subscribe to:
Post Comments (Atom)
How you like the sea life. Departing from Cameron, La today to take some workers off shore to jackup boat. Hope you are having a wonderful time. thanks for so many nice excel sheets. Got em all. Captain Donald(when I am not driving for a living).
ReplyDelete