Silicon Investor (SI) -- The First Internet Community

STOCKTALK

We've detected that you're using an ad content blocking browser plug-in or feature. Ads provide a critical source of revenue to the continued operation of Silicon Investor. We ask that you disable ad blocking while on Silicon Investor in the best interests of our community. If you are not using an ad blocker but are still receiving this message, make sure your browser's tracking protection is set to the 'standard' level.

Technology Stocks : Dell Technologies Inc. -- Ignore unavailable to you. Want to Upgrade?

To: BGR who wrote (83321)	12/2/1998 3:47:00 PM
From: Lee	Read Replies (2) \| Respond to of 176387

Hi Apratim,..Re:.Substantial research has been done which show that percentage increase in equity prices follow a "random" (same as Brownian motion) path. Thanks for your reply. I am aware that many studies totally debunk the validity of TA; however, there are newer studies which may show some correlations. Just can't find the references right now. <vbg> One is from an MIT professor but I can't remember the name, otherwise I could provide the reference. Will keep looking, Lee

To: BGR who wrote (83321)	12/2/1998 8:41:00 PM
From: Moominoid	Read Replies (1) \| Respond to of 176387

Substantial research has been done which show that percentage increase in equity prices follow a "random" (same as Brownian motion) path. Absolutely true, but first there are the following qualifications: 1. DlnP is nonintegrated (I(0) in econometricians parlance) but there are cycles in it - it isn't perfect white noise. 2. Reasonably large segments of the series lnP are a random walk (or BM in continuous terms). Shorter chunks, from my research are not necessarily. There may be some fractile type self similarity as well. 3. The traditional TA methods which work for me decompose the series into integrated and non-integrated components. At least that's what it seems to me currently. Any moving average approach decomposes the series to the moving ave, M, + (P-M). P-M is serially correlated and I(0). The shorter the moving ave the more stationary it is. The method seems to work because we now have a component M which has a much narrower forecast confidence interval than the original series P and a series that tends to revert to mean P-M. The Bollinger Band method just sticks a useful yardstick to assess whether P-M is near the top of its range or not. Of course this isn't the interpretation placed on this method by many TA afficionados. Somebody maybe can prove that this decomposition really doesn't lead to smaller forecast errors than those of the original series. Show me that and then I'll drop using the method. I would guess that by imposing some structure you do get some gain. Stochastics are pretty similar too. David