uh yea what he said no really i think you answered the question which was meant for me - your explanation was far better than the one i would have taken the time to give. i appreciate your response and it's clear you know what's going on and have a deep understanding many lack. mb
I do trade off of statistical stuff, including correlations, but I've been doing it for a real long time, and I know what the weaknesses are, as elaborated partially by QuantPlus. But what really gets these guys is that both in 1998 and just before this recent blowup, volatility was low. If you think of this in terms of Bollinger Bands, which are based on deviations off the trend, the bands just prior to either of these events from a longer-term POV were narrow. Which means that it would take less price movement than otherwise to get outside of "normal", statistically defined as two deviations away from the trend. So, you get a price movement that's far out of "normal", but as defined only by recent data. If you trade based on that, you have to be aware that what looks "normal" currently, may not be. You can't categorically decide that it isn't, but you equally can't just glibly assume that it is. And that, I think, is where they go wrong.
smarter trading - kaufman tackes this very problem and supplies an answer in which he excludes the anomaly which causes the out of normal, from the future calculation. the indicator resets itself. percentages are really something that these quants should study. mb
Seems like this dilemma transcends trading. The Colorado River is central to civic planning/growth management planning for Las Vegas - Since the Colorado provides much if not all of the water for Vegas. There are rumors that when they looked at how much water the Colorado could provide to Las Vegas, they looked at 50 years of data. Since then, with the advancements in science, they have been able to get a feel for the highs and lows of water supply from the Colorado going back well beyond 50 years. There are rumors that some geologists feel the 50 year period used for the study was a period of abnormally high water supply. So, it is the same thing - using the wrong "normal." I hope the rumor is wrong - I dig Vegas.