|
PureBytes Links
Trading Reference Links
|
Here is a simple home made SDI.
The * 100 factor is purely arbitrary (it just produces a number that
we can visually and mentally relate to - the absolute value is not
important - only the relative value.
//P_SDIRelativeROC
SDI = abs(ROC(C,1)/ROC(V,1)) *100;
//Plot(ROC(V,1),"V%change",2,1);
Plot(SDI,"SDIRelativeROC",1,1);
This is only a typical example and not the definitive indicator.
It shows the fun we can have with SD:
a)Which data point best references the price movement - close,
midpoint (averages the price move caused by SD imbalance), ATR(1),
bar range or what?
b) they are volatile indicators
c) if a frequency distribution is done I expect it would be very fat-
tailed (plenty of occasional wild extremes).
I think it also highlights one of the big challenges in TA - how do
we standardise a volatile indicator like that, where one big outlier
in the selected range can markedly where our line in the sand is
drawn (do we use StDev, MA or what to get a handle on where to put a
marker that will mean something when the SDI value hits it?)
brian_z
--- In amibroker@xxxxxxxxxxxxxxx, "brian_z111" <brian_z111@xxx> wrote:
>
> Elaborating a little further.
>
> (ROC(C,1) with ROC(V,1))win/loss, as used in the code below, is a
> simple and effective benchmarking method for rule based traders.
> On its own it doesn't tell the whole story.
> It needs to be compared to a standardised return that also includes
> ave%Won/ave%Lost (PowerFactor).
>
> The benchmark, for comparison in this example, is simply ROC(C,1) as
> PowerFactor (since the market has an upward bias the result will be
> slightly positive, reflecting a small expectancy each day we are in
> the market).
>
> ROC(Array,Periods), as PowerFactor (PoF), where Periods == the
> average time in the market for the system tested, returns the quick
> and easy measure of 'the market to beat" i.e. it measures the
returns
> on a 'dumbluck' investment.
>
> (Sorry, PoF should have a superscripted "o" - can't do it here).
>
> Someday, if time permits, I will post a fuller explanation.
>
> Back to volume:
>
> In my 2 years in the forum I haven't seen a compelling argument in
> support of volume as an indicator.
>
> Here is why I am not excited about it (as an EOD signal).
>
> If a gazillion shares change hands in one day, and at the close the
> price hasn't moved from the previous day, then all of the interest
> has been soaked up (supply == demand).
>
> Compared to that, if only 1000 shares changed hands for the day,
with
> no movement in price, do we know anything different about the
> possible future price of the stock compared to the above scenario?
>
> However, the situation could change suddenly and unpredictably if
the
> supply/demand balance shifts, for some reason.
>
> Basically, volume is somehow interrelated to supply/demand, but is
it
> causal?
>
> In my thinking so far, it seems that price movement is a more direct
> measurement, of S/D imbalance, than price/volume.
>
> So, my bias, at this stage, is to treat price momentum as a
surrogate
> for volume and not bother with the volume metric (one less to worry
> about).
>
> From there I played around with my own S/D indicators (I know plenty
> of others have done that in the past but I don't know the first
thing
> about their methods).
>
> I used things like ATR/vol, range/vol, ROC%/vol etc to try to get a
> metric of Supply/Demand imbalance i.e. how much is any unit of
volume
> actually affecting (moving) the price.
>
> Understandably, SDI indicators are biased towards days with a small
> movement (range and/or price moves) and we can already see that bias
> from the charts, without the need for additonal indicators.
>
> The problem I have with S/D indicators is that:
>
> a) price is a small number/a very large volume number
> b) %price change is small/a proportonally large % vol change
> c) we don't know, on an EOD basis, where the volume is going
> (millions of shares could change hands during the day, with no price
> movement, and then the last share of the day sells for plus 2% ==
the
> close.
>
> That is where I left off the enquiry, short of going to intra-day
> charts where the SDI (SupplyDemandIndicator) can be pinned to a
time.
>
> I would be happy to hear from anyone who can point out any
weaknesses
> in my theory that volume as a second rate EOD indicator.
>
> brian_z
>
>
>
> --- In amibroker@xxxxxxxxxxxxxxx, "brian_z111" <brian_z111@> wrote:
> >
> > Anyone?
> >
> > Here is why I can't get past first base with volume indicators.
> >
> > I have made some quick and dirty changes to Tomasz's Price *
Volume
> > code.
> >
> > If it is plotted it tallies the number of times that a (vol +
price)
> > up day is followed by a win (another price upday).
> >
> > In some old DJI components data I find that the result starts to
> > approach 50% quite quickly (approx 2600 data points).
> >
> > Given the inherent errors in such a quick and dirty method it is
> > around what I expect for a no-win signal.
> >
> > (for the deep thinkers 50/50 is the binomial equivalent of the
> golden
> > mean).
> >
> > a) Insert the code into a new indicator pane.
> > b) Scroll back to the beginning of a chart to see all of the
plots
> > and cross-check the outputs against the chart.
> > c) Scroll to the end of the chart to get symbol by symbol totals.
> > d) Use the keyboard arrow keys to scroll down the symbols list.
> >
> > //P_VolPriceSynchronization
> > //adapted from code by Tomasz Janeczko
> >
> > //Lookback = Param("Change lookback", 1, 1, 100 );
> >
> > //ChgPrice = ROC( C, Lookback );
> > //ChgVolume = ROC( V, Lookback );
> >
> > ChgPrice = ROC( C,1);
> > ChgVolume = ROC( V,1);
> >
> > BothUp = IIf( ChgPrice > 0 AND ChgVolume > 0, 1,0);
> >
> > //Period = Param("Sum period", 25, 1, 100 );
> >
> > Plot( Sum( BothUp,1), "UpVolANdPrice", colorRed );
> >
> > Upday = ROC(C,1) > 0;
> >
> > Plot(Upday, "CloseUpday", colorBlue,2 );
> >
> > TotalUpdays = Cum(IIf(Upday,1,0));
> >
> > TotalBothUp = Cum(BothUp);
> >
> > Plot(TotalBothUp, "TotalBothUp", colorBlack);
> >
> > BothUpWin = Upday == 1 AND Ref(BothUp,-1) == 1;
> >
> > TotalBothUpWin = Cum(IIf(BothUpWin,1,0));
> >
> > Plot(TotalBothUpWin, "WinFollowsVPDay", colorGreen);
> >
> > If the raw signal is not significant can we improve on that by
> using
> > some form of mathematical conditioning?
> >
> > brian_z
> >
> >
> > --- In amibroker@xxxxxxxxxxxxxxx, "brian_z111" <brian_z111@>
wrote:
> > >
> > > Jeremy,
> > >
> > > Thks for posting the link.
> > >
> > > Since I haven't found any use for Vol so far (limited to EOD
> equity
> > > trading) and you have had some success with VPCI:
> > >
> > > Have you done much work with V indicators?
> > > If so, what is your pick as the top three?
> > >
> > > Any hints on how to use VPCI or should I just follow Dormeiers
> > > recommendations?
> > >
> > > brian_z
> > >
> > > --- In amibroker@xxxxxxxxxxxxxxx, "jeremy7827110028"
> > > <jberkovits1@> wrote:
> > > >
> > > > kethek,
> > > >
> > > > I dont know if you are familiar with the VPCI indicator. It
> was
> > > > created by Buff Dormeier CMT of Wachovia Securities. Buff
> > recived
> > > the
> > > > Charls Dow award for his orignal research. I have tested the
> > VPCI
> > > > extensivly and it has a stitastiaclly signifigant positive
> impact
> > > on a
> > > > wide vierity of systems. You might find it of value. here
is
> a
> > > link
> > > > to the code and a TSC article from July 2007.
> > > >
> > > > Regards,
> > > >
> > > > jeremy
> > > >
> > > >
> > >
> >
>
http://www.traders.com/Documentation/FEEDbk_docs/Archive/072007/Trader
> > > sT
> > > > ips/TradersTips.html#amibroker
> > > >
> > >
> >
>
Please note that this group is for discussion between users only.
To get support from AmiBroker please send an e-mail directly to
SUPPORT {at} amibroker.com
For NEW RELEASE ANNOUNCEMENTS and other news always check DEVLOG:
http://www.amibroker.com/devlog/
For other support material please check also:
http://www.amibroker.com/support.html
Yahoo! Groups Links
<*> To visit your group on the web, go to:
http://groups.yahoo.com/group/amibroker/
<*> Your email settings:
Individual Email | Traditional
<*> To change settings online go to:
http://groups.yahoo.com/group/amibroker/join
(Yahoo! ID required)
<*> To change settings via email:
mailto:amibroker-digest@xxxxxxxxxxxxxxx
mailto:amibroker-fullfeatured@xxxxxxxxxxxxxxx
<*> To unsubscribe from this group, send an email to:
amibroker-unsubscribe@xxxxxxxxxxxxxxx
<*> Your use of Yahoo! Groups is subject to:
http://docs.yahoo.com/info/terms/
|