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

Re: [amibroker] Re: Short system advice?



PureBytes Links

Trading Reference Links

Brian

You've lost me, I'm afraid. This is clearly way ahead of where I would hope to be with AFL and system design, even some time from now, but the ideas are intriguing.

Regards.

ChrisB

Brian <brianrichard99@xxxxxxxxxxx> wrote:
So far it looks like I will be weighting signals based on 1) how
consistently leading the indicator is, and 2) various ratios used to
identify profitability during the optimization process. Currently my
signals are all equal weight. As I get used to how the system
performs in papertrading, I will tweak the weights accordingly. I
will be writing some code to give each indicator a ratio relative to
all other signal ratios in its signal group. That way I don't have
to assign a number, which would force me to become refamiliar with a
new signal range (1-100 is now 1-120, etc.). The results in AA would
then come out as percentages instead of integers, that way.

If you come up with a better weighting method please send me an
email at brian (at) brianrichard (dot) com.

Thanks!

Brian


--- In amibroker@xxxxxxxxxxxxxxx, kris45mar <kris45mar@xxx> wrote:
>
> Brian
>  
>   Good feedback, thanks.
>  
>   Yes "feeler trades" is basically where I am at at the moment.
Keep probing the market, if good, add more, if bad, too bad.
>  
>   A recent presentation in Perth by a systems developer (who sells
his  signals commercially so I am not interested: I am too pig
headed for  this I suppose) is based on this principle. He runs 10
systems over the  top 30 US stocks. If on any day there are 7 or
more buy signals from  the individual systems, this is his buy
recommendation. Sort of a  weight of evidence theory. Seems based on
sound statistics, thousands  of trades in his stats, stop and
reverse system, in and out of sample  testing with smooth equity
curve and manageable drawdowns. EOD only and  entry/exit on open.
>  
>   Similar I guess to what you are doing by weighting your signals.
This may ultimately where I am headed.
>  
>   Sounds like I have more coding to learn: half the fun, though.
>  
>   Regards
>  
>   ChrisB
>  
> Brian <brianrichard99@xxx> wrote:          ChrisB,
>  
>   I just completed another "system" that's comprised of about 15
>   different proprietary indicators, all of them optimized for both
EOD
>   and Weekly timeframes, as well as for a specific group of 500
>   stocks, ETFs and CEFs. I used to look at just two or three
>   indicators (a "system"), but found each indicator has its unique
>   weakness. So now I just go with looking at all of the signals
that
>   all of my best indicators generate. I weight each signal. I also
>   look at signals that are up to 3 bars old, and weight those
signals
>   less.
>  
>   I will eventually try to automate all of this optimization. Not
>   there yet.
>  
>   I also separated out trending signals from daily buy/sell
signals,
>   so I really have a separate trend system as well.
>  
>   I've read more than once, from reading passages written by very
>   profitable traders, that their systems generally take the same
form
>   as the one I've created. Very few boil everything down to one
type
>   of trade that they do over and over. Problem with those single
setup
>   trades is you still need to look at all important variables
outside
>   of the setup that can affect the setup. A mentor statistican
friend
>   of mine recommended I build in an additional "trade
cancellation"
>   system that lets me know when outside variables are building
against
>   the system signal. This helps me sort the best trades out from
the
>   bunch.
>  
>   I am using no equity curve. IMO that would just seem to add
another
>   layer of unnecessary complexity. Focus on money management --
>   scaling and scaling out, user "feeler" trades, etc. That will
likely
>   get you farther down the road. My personal goal is to use my
system
>   to identify good trades for my discretionary style of trading.
>  
>   Just my toe scents.
>  
>   ~Brian
>  
>  
>  
>   --- In amibroker@xxxxxxxxxxxxxxx, kris45mar <kris45mar@> wrote:
>   >
>   > Phew, Yuki.
>   >  
>   >   Honoured to humbled to receive your lengthy reply. Please be
>   warned:  this inspirational, supportive ( and midly cajoling )
reply
>   ( thank  you! ) may be  transferred into my "Yuki says"
handbook! 
>   Everything you say strikes a resonant note though, and is taken
in
>   good  spirit.
>   >  
>   >   I have 25% DD with 4 wins out of the last 30 discretionary
>   trades.
>   >   Looking through my last two years' trades tells me that what
>   worked in  2004 is not working in 2005/6. This brings me to the
>   point in my 2005  trading plan where I defined conditions to
stop
>   trading. I now need a  change of direction: the plan is to
continue
>   to explore AB, AFL and the  superb posts on this board towards
>   developing a mechanical system. It  can't be that hard for me to
>   develope one that does better than my 2005  trading year.
Whether I
>   can then actually trade it is a whole different  ball game.
>   >  
>   >   You said:
>   >  
>   >   "
>   >   And thank goodness not everyone can do this.  We need some
>   productive
>   >   members of society, too.  ^_-
>   >   "
>   >  
>   >   LOL.... and yet the lesson we learn about ourselves by
trading
>   can make us more productive in other areas!
>   >  
>   >   In summary:
>   >  
>   >   You will never avoid drawdowns: agreed.
>   >   Sharper gains (with a reliable system) may come when the
equity
>   curve is below its MA. Sounds logical, and worth exploring.
>   >  
>   >   All I am asking is this:
>   >  
>   >   Markets change over time (that is why there is no Holy
Grail)
>   and so  should our systems, or the ones we choose to trade with,
not
>   respond to  this? Or we may choose to stand aside for a while.
Or
>   just trade  different markets with concurrently different
systems to
>   create a  smoother equity curve overall?
>   >  
>   >   Could you comment on whether you trade with one system only
or
>   more  than one? And if more than one, what would be a trigger to
>   change if  the Equity curve is not the signal to do so?
Drawdowns?
>   Sleepless  nights? Declining expectancy? This has to part of our
>   business plan  after all. In 2004 I achieved my trading goals,
2005
>   was not a  successful one. Message: time to stop doing what I am
>   doing: it is not  working. Do something else. The goal then is
to
>   replace what I am doing  with something that does work.
>   >  
>   >   I realise the answers to these questions are personal, but
it
>   is  invaluable to get some insight to the philosophies of
others, in
>   an  attempt to know where to start.
>   >  
>   >   Regards
>   >  
>   >   ChrisB
>   >  
>   >  
>   >
>   > Yuki Taga <yukitaga@> wrote:          Hi kris45mar,
>   >  
>   >   Monday, March 13, 2006, 11:35:06 PM, you wrote:
>   >  
>   >   k>   b. When the Equity is above the MA, then take the
signals.
>   >    
>   >   k>   c. when the equity curve falls below its MA, then
either.
>   >    
>   >   k>       i. stop trading that system until such time as the
>   curve goes back above the MA.
>   >    
>   >   k>       ii. or severely reduce position size.
>   >    
>   >   k>       iii. and/or swap over to another system that is now
>   above its MA.
>   >    
>   >   You will get various opinions on this, however I think it
really
>   >   boils down to just how logical you suspect your system
>   methodology
>   >   is, and whether you suspect it is actually and finally being
>   >   arbitraged out of existence.
>   >  
>   >   If the system has worked for several business cycles in
various
>   >   market modes, and has never really gotten into serious
trouble --
>    in
>   >   other words, it's a system you can trade -- then it would
seem
>   to me,
>   >   and indeed is what I do, that the time to be more careful is
when
>   >   equity has been running well above the MA for some rather
lengthy
>   >   period of time. I'm inclined to bump *up* position size a
little
>   bit
>   >   when I start experiencing a losing streak -- in other words
when
>   I
>   >   get mean reversion or worse of the equity curve.  I would
>   certainly
>   >   not stop trading when that happens.  I think your gut
feeling is
>   >   exactly opposite of what you should do.
>   >  
>   >   The sharpest gains and nicest times you are likely to ever
have
>   are
>   >   when equity is making the swing from below average to above
>   average.
>   >   This is much more fun than the opposite, and you are going to
>   >   experience both. So why would you consider stopping trading
when
>   >   equity dips below average? Immediately, you would then be
>   preparing
>   >   to cheat yourself out of your best performing part of the
cycle,
>   and
>   >   you would be ready to embrace the worst cycle segment of
your
>   system:
>   >   when equity moves from above average to below average.
>   >  
>   >   In the end, it all boils down to confidence.  You either
have a
>   >   viable system, or you don't.  If you have one, follow it. 
If you
>   >   can't stand the drawdowns ... IMHO, you don't have a viable
>   system,
>   >   and probably should not be trading it.  No one should trade
any
>   >   system that has drawdowns they cannot stomach, and stomach
>   >   comfortably, probably max system percentage drawdown times
two,
>   maybe
>   >   times 2.5 or three.
>   >  
>   >   But if you really do have a system, take every signal.
Period.
>   If you
>   >   want to "play" your system a little bit, consider something
like
>   >   *lightening* position size -- slightly -- when equity has
been
>   >   running above average for some period of time, and
*increasing*
>   it
>   >   ... again, slightly ... when equity has been running below
the
>   line
>   >   for some time.  You have to judge when these conditions
might
>   apply
>   >   after carefully analyzing your system yourself.
>   >  
>   >   But using the MA of equity to flatly refuse or take signals
is
>   simply
>   >   a different form of "Holy Grailism".  It is a fear of taking
>   losers,
>   >   or an attempt to altogether avoid taking losers, which
absolutely
>   >   must be taken in any systematic trading. You simply have to
have
>   a
>   >   system in which you can *stand* to take the losers, and be
>   >   comfortable with them.  If you don't, you can't trade it,
and
>   playing
>   >   around using the equity curve as an ultimate filter is not
>   likely to
>   >   made a dangerous system safe, or an uncomfortable system
>   comfortable.
>   >  
>   >   You will never, ever, find a system that has an equity curve
that
>   >   doesn't dance on both sides of a MA.  Life doesn't work that
>   way. But
>   >   if the curve is obviously solid, in other words, a real
curve or
>   >   slope, and not an amusement park thrill ride, and all the
metrics
>   >   look nice over thousands of trades and many years, you may
want
>   to
>   >   think about doing exactly the opposite of what your gut
tells you
>   >   when you hit a soft patch.
>   >  
>   >   OTOH, if the last sentence above applies ... why stress
yourself
>   at
>   >   all?  Take the *&$% signals as they come, and relax.  ^_^ 
If you
>   >   cannot stand a loss the magnitude of which would tell you
that,
>   >   indeed, your system is no longer functioning, you are
probably
>   not
>   >   well enough capitalized to be in this business.  Not
everybody
>   is.
>   >   And thank goodness not everyone can do this.  We need some
>   productive
>   >   members of society, too.  ^_-
>   >  
>   >   Yuki
>   >  
>   >            
>   >
>   >     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 other support material please check also:
>   >   http://www.amibroker.com/support.html
>   >  
>   >    
>   >
> >                       
>   >
>   >           SPONSORED LINKS   
> >                                                         
Investment
>   management  software                                          
Real
>   estate investment 
software                                          
>   Investment property
>
software                                                            

>                Software
> support                                           Real estate
>   investment analysis
>   software                                           Investment
>  
software                                                            

>                  
>   >        
>   > ---------------------------------
>   >     YAHOO! GROUPS LINKS 
>   >  
>   >      
>   >     Visit your group "amibroker" on the web.
>   >     
>   >     To unsubscribe from this group, send an email to:
>   >  amibroker-unsubscribe@xxxxxxxxxxxxxxx
>   >     
>   >     Your use of Yahoo! Groups is subject to the Yahoo! Terms
of
>   Service. 
>   >  
>   >      
>   > ---------------------------------
>   >  
>   >  
>   >          
>   >
>   >            
>   > ---------------------------------
>   >  Yahoo! Mail
>   >  Use Photomail to share photos without annoying attachments.
>   >
>  
>  
>  
>  
>            
>
>     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 other support material please check also:
>   http://www.amibroker.com/support.html
>  
>    
>
>                      
>
>           SPONSORED LINKS   
>                                                         Investment
management software                                          Real
estate investment software                                         
Investment property
software                                                            
             Software
support                                          Real estate
investment analysis
software                                          Investment
software                                                            
               
>        
> ---------------------------------
>     YAHOO! GROUPS LINKS 
>  
>      
>     Visit your group "amibroker" on the web.
>     
>     To unsubscribe from this group, send an email to:
>  amibroker-unsubscribe@xxxxxxxxxxxxxxx
>     
>     Your use of Yahoo! Groups is subject to the Yahoo! Terms of
Service. 
>  
>      
> ---------------------------------
>  
>  
>          
>
>            
> ---------------------------------
> Yahoo! Mail
> Bring photos to life! New PhotoMail  makes sharing a breeze.
>






YAHOO! GROUPS LINKS






Yahoo! Mail
Bring photos to life! New PhotoMail makes sharing a breeze.

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 other support material please check also:
http://www.amibroker.com/support.html





SPONSORED LINKS
Investment management software Real estate investment software Investment property software
Software support Real estate investment analysis software Investment software


YAHOO! GROUPS LINKS