Comments on: Bollinger Bands

By: Ron Pereira

Ron Pereira — Wed, 21 Feb 2007 02:02:00 +0000

Mark… are you a closet Six Sigma geek? Just joking man. Yeah there are some pretty cool tricks these market gurus use to time things. In the end though it comes down to bulls, bears, and those poor pigs who get slaughtered.

By: Mark Graban

Mark Graban — Tue, 20 Feb 2007 23:22:00 +0000

I’ve always wondered if you could use Statistical Process Control methods for evaluating stocks… to tell if an increase or decrease was “common cause” fluctuation or “special cause” change.

By: Ron Pereira

Ron Pereira — Tue, 20 Feb 2007 20:18:00 +0000

Great comment on normality. I was going to mention this but left it out… I am glad you mention it though since I do think someone could do well developing some non parametric stock market timing tools. Maybe there is an indicator with your name on it Matt! Just remember us little people when they write books about you. Cheers!

By: Matt Meyers

Matt Meyers — Tue, 20 Feb 2007 15:31:00 +0000

I find these things fascinating, as a stats geek and as a amateur trader. Bollinger Bands are based on standard deviation, and using standard deviations implies that the data has a normal distribution. Stock prices are not normally distributed. And yet, Bollinger Bands (and other quant stock techniques assuming normality) work (well, they have their own success rate < 1), despite the mathematical inconsistency, simply because the quant analysts use them. And so the tail wags the dog.