Re: [amibroker] Re: Absolute value ATR?, AmiBroker Email List Archive

2008/2/25, brian_z111 <brian_z111@xxxxxxxxx>:

Howard's book has a chapter on issue selection (filtering by
liquidity). The concepts are the same so it should help in that
regard.

The main function of this forum is to help learn and apply AmiBroker
so most of the time you will get an answer if you have a
specific "how to do such and such in AB".

I can't show you how to do volatility bands at the moment.
Maybe another day or I will post on it at the UKB (I found it useful
for benchmarking signals - Howard also gives a method for
benchmarking signals in his book).

He is writing another AB training book that will be out soon.

He has talked about a third book.
Those of us who have been around awhile are waiting for that one.

brian_z

--- In amibroker@xxxxxxxxxxxxxxx, "Louis Préfontaine"
<rockprog80@xxx> wrote:
>
> Hi again Brian,
>
> I think I understand what you mean. Do you think that reading
Howard's book
> will help me splitting the stocks betweeen high volatility and low
> volatility and then be able to work with that? I sure would like
to get a
> different stop-loss % for each of those groups.
>
> I tried to make groups but ran into some problems. I experimented
with
> "setoption" to build a custom backtesting procedure but I just
can't filter
> groups/industry/index so I don't have every stocks in the
backtesting (it
> isn't useful to get NYSE trin as a buy/sell signal...).
>
> I tried
>
> Filter=0;
> Filter = IndustryID(0) == 254;
>
>
> (I want to exclude the Industry number 254 (that is, the one before
the last
> one in the group) from the scan... I think I really get this
wrong... Will
> Howard's book help me with this kind of issues?)
>
> I had a look to the UKB to the books you recommanded. I might buy
the first
> one on the list as well. I really need to catch up very fast! ;-)
>
> Louis
>
>
>
>
>
> 2008/2/25, brian_z111 <brian_z111@xxx>:

> >
> > I won't look at your code too closely.
> >
> > > I get only results of ATR > 3-4. Only Indu.x is under 2; all the
> > >other one
> > > are over 3-4, and sometimes 6-7 and more.
> >
> > 2% was an off the cuff example.
> > Move that wherever you want to achieve whatever it is you want to
do
> > (if you can).
> >
> > > However, are you sure about the idea of entering/exiting a stock
> > >when
> > > volatility gets too high?
> >
> > No, I didn't say that (that is another story).
> >
> > I gave one example of using volatility in trading i.e. volatility
> > stops (Tomasz uses an ATR example as his trailing stop in the help
> > manual so it synched with that).
> >
> > > Wouldn't it be better simply to scan and >avoid
> > > high volatily stocks to reduce the drawdown possibility?
> >
> > You might have one strategy for low vol stocks and another for
high
> > or you might find volatility too hard to handle and avoid it.
> >
> > The key factor, in theory, is Reward/Risk.
> > With volatility stops in place a higher volatility stock might
have,
> > say twice the risk. If it has three times the Reward then it can
> > absorb the extra Risk (there is more to it than that but that is
the
> > basic starting point).
> >
> > brian_z
> >
> > --- In amibroker@xxxxxxxxxxxxxxx <amibroker%
40yahoogroups.com>, "Louis
> > Préfontaine"
> > <rockprog80@> wrote:
> > >
> > > Hi,
> > >
> > > Thanks for the suggestion. I must do something wrong however
> > because from
> > > the code I wrote
> > >
> > > _SECTION_BEGIN("ATR");
> > > periods = Param( "Periods", 15, 1, 200, 1 );
> > > Plot( ATR(periods)/Ref(C,-1)*100
> > > , _DEFAULT_NAME(), ParamColor( "Color", colorCycle ), ParamStyle
> > ("Style") );
> > >
> > > LongPer = Param("Long Period", 50, 10, 100, 5 ); /* select
periods
> > with
> > > parameter window */
> > >
> > > MidPer = Param("Mid Period", 20, 0, 50, 1);
> > > ShortPer = Param("Short Period", 10, 3, 10, 1 );
> > >
> > > LongEMA = EMA( ATR(periods)/Ref(C,-1)*100, LongPer );
> > > MidEMA = EMA (ATR(periods)/Ref(C,-1)*100, MidPer);
> > > ShortEMA = EMA( ATR(periods)/Ref(C,-1)*100, ShortPer );
> > >
> > > _SECTION_END();
> > >
> > > _SECTION_BEGIN("ema-ATR");
> > > Plot( LongEMA, "EMA(ATR(periods)/Ref(C,-1)*100, "+WriteVal
> > (LongPer,1)+")",
> > > colorBlue,
> > > styleLine|styleNoRescale );
> > > Plot( MidEMA, " EMA(ATR(periods)/Ref(C,-1)*100, "+WriteVal
(MidPer,1)
> > +")",
> > > colorBrown,
> > > styleLine|styleNoRescale );
> > > Plot( ShortEMA, " EMA(ATR(periods)/Ref(C,-1)*100,
> > > "+WriteVal(ShortPer,1)+")", colorGreen,
> > > styleLine|styleNoRescale );
> > >
> > >
> > > _SECTION_END();
> > >
> > > I get only results of ATR > 3-4. Only Indu.x is under 2; all the
> > other one
> > > are over 3-4, and sometimes 6-7 and more.
> > >
> > > However, are you sure about the idea of entering/exiting a stock
> > when
> > > volatility gets too high? Wouldn't it be better simply to scan
and
> > avoid
> > > high volatily stocks to reduce the drawdown possibility?
> > >
> > > Thanks,
> > >
> > > Louis
> > >
> > > 2008/2/24, brian_z111 <brian_z111@>:
> > > >
> > > > If you want a relative measure of range then you could use
> > ATR%, as
> > > > suggested by Graham.
> > > >
> > > > High volatility stocks would be filtered by e.g. ATR% > 2 etc.
> > > >
> > > > In that case your stops would be something like:
> > > >
> > > > ProfitStop = Ref(C,-1) * (1 + ATR%/100);
> > > > StopLoss = Ref(C,-1) * (1 - ATR%/100);
> > > >
> > > > If you want to standardise range then you could use:
> > > >
> > > > StandardATR = StDev(ATR(1),Periods);
> > > >
> > > > An example of a stop would then be:
> > > >
> > > > ProfitStop = Ref(C,-1) + StDev(ATR(1),20);
> > > >
> > > > Or maybe (for live work):
> > > >
> > > > ProfitStop = Ref(C,-1) + Ref(StDev(ATR(1),20),-1);
> > > >
> > > > Or something like that.
> > > >
> > > > That is only one way of doing it.
> > > >
> > > > You can try whatever you like, within the boundaries of your
> > > > knowledge (plus a little bit more).
> > > >
> > > > brian_z *;-)
> > > >
> > > >
> > > > --- In amibroker@xxxxxxxxxxxxxxx <amibroker%

40yahoogroups.com><amibroker%40yahoogroups.com>,

> > > > "brian_z111" <brian_z111@> wrote:
> > > > >
> > > > > Louis,
> > > > >
> > > > > > only thing I need to know is simply to set the STdev at 2
or 3
> > > > (if
> > > > > it's what
> > > > > > I want to do) and then automatically ATR will be use that
new
> > > > StDev?
> > > > > >
> > > > >
> > > > > No.
> > > > >
> > > > > ATR and StDev are both measures of volatility but they
measure
> > it
> > > > in
> > > > > different ways. Generally you would use one or the other.
> > > > >
> > > > > StDev has special uses.
> > > > >
> > > > > If you want to use them it would pay off to study them
closely
> > > > first.
> > > > >
> > > > >
> > > > > > I like your idea to make two groups; one with high
volatility
> > and
> > > > > one with
> > > > > > low volatility. Would you consider it would be possible to
> > > > adjust
> > > > > the
> > > > > > stop-loss differently for each group?
> > > > >
> > > > > You could try it e.g. the stop loss can be the close - StDev
> > (C,20).
> > > > > You can vary the stop loss for one group by using a
multiplier
> > so
> > > > the
> > > > > stop loss could be close - StDev(C,20) * 1.5 for one group
and
> > > > close -
> > > > > StDev(C,20) for the other.
> > > > >
> > > > >
> > > > >
> > > > > > And how do you filter the top
> > > > > > performers?
> > > > >
> > > > > It depends on what you have chosen as your favourite metric
for
> > > > > evaluating systems. As I said in an earlier post I like
Power
> > > > Factor
> > > > > (I will be explaining this at the UKB soon) but you would be
> > better
> > > > > served choosing your own.
> > > > >
> > > > > If you are not sure on evaluation, and use of the metrics,
then
> > > > > Howard Bandy's book is a good place to start.
> > > > >
> > > > > Sorry, I can't help you any further with this.
> > > > > I have a couple of posts for the UKB I want to get finished.
> > > > >
> > > > > BTW did you see the answer I gave you yesterday on "Trying
to
> > > > compare
> > > > > market and industry" ?. see message # 120270
> > > > >
> > > > > I hopoe that helps you a little and good luck with your
trading.
> > > > >
> > > > > brian_z
> > > > >
> > > > >
> > > > > --- In amibroker@xxxxxxxxxxxxxxx <amibroker%

40yahoogroups.com><amibroker%

> > 40yahoogroups.com>, "Louis
> > > > Préfontaine"
> > > > > <rockprog80@> wrote:
> > > > > >
> > > > > > Hi Brian,
> > > > > >
> > > > > > Thanks for those explanation. I will experiment with this
> > > > tonight
> > > > > and
> > > > > > tomorrow. However, I am not sure about something: are you
> > saying
> > > > > that the
> > > > > > only thing I need to know is simply to set the STdev at 2
or 3
> > > > (if
> > > > > it's what
> > > > > > I want to do) and then automatically ATR will be use that
new
> > > > StDev?
> > > > > >
> > > > > > I like your idea to make two groups; one with high
volatility
> > and
> > > > > one with
> > > > > > low volatility. Would you consider it would be possible to
> > > > adjust
> > > > > the
> > > > > > stop-loss differently for each group? And how do you
filter
> > the
> > > > top
> > > > > > performers?
> > > > > >
> > > > > > As always, thanks for your help!
> > > > > >
> > > > > > Louis
> > > > > >
> > > > > > 2008/2/24, brian_z111 <brian_z111@>:
> > > > > > >
> > > > > > > Sorry Louis, a mistake there.
> > > > > > >
> > > > > > > I am getting my standard deviations mixed up between
> > programs.
> > > > > > >
> > > > > > > In AB StDev is 1 by default and is in $values.
> > > > > > > To use AB's StDev at 2,3 deviations etc just multiply
StDev
> > > > (C,10)
> > > > > * 2
> > > > > > > etc
> > > > > > > To use it as StDev%
> > > > > > > StDevPercent = StDev(C,Periods)/MA(C,Periods) * 100;
> > > > > > >
> > > > > > > For STANDARD measures of deviation use StDev.
> > > > > > > For relative measures of deviation use ATR as % or
StDev as
> > %
> > > > > > >
> > > > > > > One example:
> > > > > > >
> > > > > > > Say you want to compare the performance of a fast horse
and
> > a
> > > > slow
> > > > > > > horse. If they both travel 1 StDev in the same time
(number
> > of
> > > > > > > periods) their performance is equal but the VALUE
(QUALITY)
> > of
> > > > the
> > > > > > > fast horses performance is higher - it's a grade one
horse
> > > > > compared
> > > > > > > to the other horse, which is a grade 2 (using speed as
the
> > > > > criteria).
> > > > > > >
> > > > > > > In practice - profit/loss stops might be set at +- 1
> > standard
> > > > > > > devation and then filtered for the top performers. The
top
> > > > > performers
> > > > > > > could then be segregated into two watchlists - those
with 1
> > > > stdev
> > > > > >
> > > > > > > 2% (high volatility stocks) and those with stdev <=2%
(low
> > > > > volatility
> > > > > > > stocks) - this would allow a comparison of the
performance
> > of
> > > > that
> > > > > > > trading signal/stop loss combination on high and low
> > volatility
> > > > > > > stocks.
> > > > > > >
> > > > > > > brian_z
> > > > > > >
> > > > > > >
> > > > > > > --- In amibroker@xxxxxxxxxxxxxxx <amibroker%
40yahoogroups.com><amibroker%
> > 40yahoogroups.com><amibroker%40yahoogroups.com>,
> > > > > > > "brian_z111" <brian_z111@> wrote:
> > > > > > > >
> > > > > > > > Louis,
> > > > > > > >
> > > > > > > > >Does anyone know if it is possible to get an absolute
> > value
> > > > > ATR?
> > > > > > > >
> > > > > > > > The Abs() function serves that purpose but I think you
> > mean
> > > > > > > something
> > > > > > > > else.
> > > > > > > >
> > > > > > > > ATR is a measure of volatility and it is specific for
each
> > > > > stock (or
> > > > > > > > instrument). The whole idea of it (AFAIK) is to use
it on
> > an
> > > > > > > > individual
> > > > > > > > stock basis.
> > > > > > > >
> > > > > > > > It can be useful to compare volatility:
> > > > > > > >
> > > > > > > > 1) internally e.g. against an average of the last (x)
> > days OR
> > > > > > > against
> > > > > > > > the StDev (standard deviation) of the volatility
measure
> > OR
> > > > > just use
> > > > > > > > StDev of the Close etc on its own.
> > > > > > > >
> > > > > > > > StDev() function does allow to change the setting
between
> > 1
> > > > or
> > > > > 2 etc
> > > > > > > >
> > > > > > > > 2) externally to the volatility of the market OR a
sector
> > > > that
> > > > > the
> > > > > > > > stock is a member of OR compared to another stock in
the
> > same
> > > > > sector
> > > > > > > > etc.
> > > > > > > >
> > > > > > > > brian_z
> > > > > > > >
> > > > > > > >
> > > > > > > >
> > > > > > > > --- In amibroker@xxxxxxxxxxxxxxx <amibroker%
40yahoogroups.com><amibroker%
> > 40yahoogroups.com><amibroker%
> > > > 40yahoogroups.com>,
> > > > > Graham
> > > > > > > <kavemanperth@> wrote:
> > > > > > > > >
> > > > > > > > > you could try a percentage type
> > > > > > > > >
> > > > > > > > > ATR(10)/ref(c,-1)*100
> > > > > > > > >
> > > > > > > > > --
> > > > > > > > > Cheers
> > > > > > > > > Graham Kav
> > > > > > > > > AFL Writing Service
> > > > > > > > > http://www.aflwriting.com
> > > > > > > > >
> > > > > > > > >
> > > > > > > > > On 24/02/2008, louisprefontaine <rockprog80@> wrote:
> > > > > > > > > > Does anyone know if it is possible to get an
absolute
> > > > value
> > > > > > > ATR?
> > > > > > > > I
> > > > > > > > > > already use the ATR, but it changes from stock to
> > stock,
> > > > > > > > depending on
> > > > > > > > > > the value of the stock. Would it be possible to
get an
> > > > > > > absolute
> > > > > > > > value
> > > > > > > > > > indicator, like CMF, RSI, etc.? Thanks!
> > > > > > > > > >
> > > > > > > > > >
> > > > > > > > > >
> > > > > > > > > > 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
> > > > > > > > > >
> > > > > > > > > >
> > > > > > > > > >
> > > > > > > > > >
> > > > > > > > >
> > > > > > > >
> > > > > > >
> > > > > > >
> > > > > > >
> > > > > >
> > > > >
> > > >
> > > >
> > > >
> > >
> >
> >
> >
>