[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: Shifted Mov. Avgs.//optimized moving average oscillators



PureBytes Links

Trading Reference Links

Lionel,

The fib #'s always seem to add special "spin" when applied to
any formula I use.  I currently substitute fib numbers in the
MACD(dema) formula and I feel it's better than the defaults.  In
1978/79, I started fooling around with what I called "optimized
moving average oscillators" (at the time, even the word
oscillator was just creeping into technical analysis terms). 
Perry Kaufman thought it was interesting enough to call once a
week to monitor my proress for his first or second book.  
 
Anyway, I applied this systematic approach to the markets with
decent success.  Overall, I thought that I was onto something,
but moved on to other approaches.  So, Onno and others exploring
MA's here's a down and dirty explanation.  (If you develop this
test and make a fortune, send me a Superbowl ticket as a tip.)

1.  I started by taking a three day MA and comparing
(subtracting) it to a three day moving average developed 10 days
previous.
2.  The difference was plotted on a zero basis graph.
3.  If the oscillator turned up, below the zero line, for two
consecutive days, 
then a buy signal was generated.
4.  If the oscillator turned down for two consecutive days,
above the zero basis line, then a sell signal was generated.
5.  I simulated this simple approach (3/10) on ten years of
commodity data and found that it actually posted profits (best
results: beans, silver, cocoa).  Quite honestly, I was shocked! 
6.  I then "optimized" the numbers.  Instead of using 3 days
lagged 10 periods, I took 2 to 12 day averages and lagged them 5
to 15 days (developing a 10 x 10 profit/loss matrix).  I looked
for groupings of profits within the matrix.  Each commodity had
it's own set of numbers (as I recall: live hogs was a 4 day MA
lagged 6 days; cocoa was 2 days lagged 14; etc.)

All this was accomplished on a TRS 80 model I.  It became the
basis for  the core of my trading for a couple of years. 
Simple?  Stupid?  I don't know...it made money in the late
'70's.  At the time I thought the work was clever.  Any other
ideas on taking a MS indicator (i.e. MACD) and building an
oscillator out of it by comparing it to itself?  

Steve Karnish
CCT
----------
> From: Lionel Issen <lissen@xxxxxxxxxxxxxxxx>
> To: metastock@xxxxxxxxxxxxx
> Cc: Alton Stephens <astephen@xxxxxxxx>
> Subject: Re: Shifted Mov. Avgs.
> Date: Sunday, August 16, 1998 11:52 AM
> 
> Alton:
> 
> Using a price oscillator (difference of a long and short mov)
will give you
> similar results.  I have found that there is nothing special
about this use
> of a shifted mov.  I do use shifted mov's to rpoduce trading
bands.
> 
> Richards suggestion of using the square root of the length of
the mov is
> interesting. Some people have suggested using Fibonacci
intervals.
> 
> Lionel
> 
> Lionel
> -----Original Message-----
> From: Alton Stephens <astephen@xxxxxxxx>
> To: metastock@xxxxxxxxxxxxx <metastock@xxxxxxxxxxxxx>
> Date: Sunday, August 16, 1998 10:02 AM
> Subject: RE: Shifted Mov. Avgs.
> 
> 
> >Richard Estes says try shifting a horizontally by the square
root [nearest
> >whole number] of the length of the MA.  He also has other
ideas and
> >suggestions regarding shifting MAs- perhaps he will chime in.
> >
> >Tom Joseph [Advanced GET] has built a trading system based on
two 6 period
> >moving averages shifted 4 periods horizontally.  The two MAs
are based one
> >on the HIGH for the period, and the second on the LOW for the
period.  The
> >resulting bands make for a workable trading band system.  he
uses it in
> >conjunction with his proprietary XTL[eXpert Trend Locator].
> >
> >Al
> >
> >Al Stephens
> >
> >> -----Original Message-----
> >> From: owner-metastock@xxxxxxxxxxxxx
> >> [mailto:owner-metastock@xxxxxxxxxxxxx]On Behalf Of Onno
Goedknegt
> >> Sent: Sunday, August 16, 1998 9:41 AM
> >> To: metastock@xxxxxxxxxxxxx
> >> Subject: Shifted Mov. Avgs.
> >>
> >>
> >> Hi all,
> >>
> >> Please give me your opinion about the usage of
> >> (vertical and horizontal) shifted moving averages ?
> >>
> >> Can you use them for graphic purposes only or also
> >> as a basis for a (end-of-day) trading system ?
> >>
> >> Thanks!
> >>
> >> Regards,
> >> Onno
> >>
> >>
> >>
> >
> >