PureBytes Links
Trading Reference Links
|
I'm guessing that the basic concept that underlies TD's "Duration Analysis"
term, is the idea of how time relates to "value". Taking ideas from Market
Profile concepts and my own, I sort of equate time to a recurring sine wave
sort of sentiment cycle. Put more clearly, the amount of time a market
SPENDS at a price level gives important sentiment information, that rotates
from bullish-bearish-bullish-bearish-neutral (or vice versa). So for
example, suppose a market makes a new high; initially, the longer it stays
there the more bullish it is, as it is indicating that prices are being
accepted (since it is not being sold down). However, after a while, if the
market is continuing sideways, the sentiment sine wave peaks, and moves the
other way. So relating to my market example, the market has made a new
high, but has not moved higher, and is starting to stall. Good traders
sense this, and start to exit, and maybe shorts start to come in. So
suppose that time continues to pass, but the market is still trading
sideways. Pretty soon, the shorts get discouraged, and cover, and maybe
some others go long getting new confidence that the market is accepting the
higher price level. This rotation can go on for a while. However, the
longer the market moves sideways, the less effect this sine wave of
sentiment will exert, until eventually, its effect will be negated, at which
time we look to the next higher timeframe for influence. Naturally, the
market will usually move instead of going sideways for an extended period of
time, but by being in sync with this cycle, you can better time your
entries.
The big question I have is what oscillator Demark is using. More
importantly, what is the oscillator measuring?
Chris Lober
-----Original Message-----
From: Ron Augustine <RonAug@xxxxxxxxxxxxx>
To: omega-list@xxxxxxxxxxxxxx <omega-list@xxxxxxxxxxxxxx>
Date: Saturday, November 07, 1998 10:57 AM
Subject: Duration Analysis?
>
>Has anyone worked with this concept? Any specifics as to how it's
>calculated and quantified?
>
>The following was posted to one the TSC forums:
>
>"Tom DeMark's book, New Market Timing Techniques, explores a concept called
>"duration analysis." He uses duration analysis to determine whether or
not
>an overbought/oversold condition on an oscillator is indicating price
>reversal or price continuation. According to his work, a "mild" overbought
>condition (i.e., one that lasts less than five to six consecutive days)
>precedes price reversal, whereas a "severe" condition (i.e., one that lasts
>more than six consecutive days) precedes price continuation.
>
>Obviously, there's a little more to it than that, but you get the idea.
>(Interestingly, my charting of the S&P just fulfilled the severe criteria
>last Friday). It's a fairly simple concept, although the book itself is
>pretty darn complex."
>
>
>
|