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

Re: Discretion vs Mechanical



PureBytes Links

Trading Reference Links

Hi Charles and Peter

For me Charles has made a good fist of the replies. 

The only comment I'd add is before we can conclude
we can trade is you'd have to trade real-time. Paper
trading, no matter how rigorous, at best, tells us 
our system has some sort of edge. We still need to 
find out if we, as traders, have an edge.

regards

ray

R Barros
101/25 Market Street
Sydney NSW 2000
Australia

Voice:   61 2 92673470  
Fax:       61 2 92673478
E-Mail:  rbarros@xxxxxxxxxxxxxxxxxx

----------
> From: charles meyer <chmeyer@xxxxxxxx>
> To: RealTraders Discussion Group <realtraders@xxxxxxxxxxxxxx>
> Subject: Re: Discretion vs Mechanical
> Date: Wednesday, August 05, 1998 12:50 PM
> 
> Peter:
> 
> I'm on a slippery slope here so I'll just ask the more experienced guys
> to jump in and correct anything I say that is wrong.
> 
> > Okay.. when you develop this system, it can be on paper, not
programmed,
> > right?
> 
> It should be in writing but does not have to necessarily programmed. 
Just
> write out the 'rules' that you intend to follow to trigger a trade entry.
> 
>  And your profit-taking could be based on support/resistance, rather
> > than just an arbitrary percentage?
> 
> Yes, Douglas says this should all be 100% mechanical.  No judgment
> involved.......
> 
>  So it has some level of discretion in
> > it... If you use H&S or other classical formations, is that mechanical?
> 
> Only if you quantify what constitutes a H&S so you do not have to
'guess'.
> It either meets your criteria or it doesn't.....
> 
>  And
> > backtesting, for the edge, how do you do this... by buying historical
> data
> > and running a backtest?
> 
> Yes.
> 
>  Wouldn't you have to real-time test to actually
> > know if it's possible ie. call your broker when you get the signal, get
a
> > quote, record it, subtract reasonable slippage & comm. and log it as a
> > paper-trade?
> 
> I would say yes; all the above appears acceptable, BUT be intellectualy
> honest with yourself and the log entries.  You could real-time test on
> paper
> first if you were uncertain about 'taking all the trades'.  And that's
the
> rub.
> 
>  I have certain concerns about the "strictly mechanical"
> > approach, but you both sound as if you're talking about a disciplined
> > rule-based approach, rather than a "black box" Club 3000 type of
> > methodology eg. Catscan, Aberration.
> 
> System should NOT be black box.  You cannot build trust in a black box
> that you did not build yourself.  Your system, and it's accompanying
rules
> should be customized to you--by you and for you.  Everyone has concerns
> about
> 'strictly mechanical' because we try to outguess the market and our
system.
>  The
> idea is to develop discipline and trust in oneself by 'taking all the
> trades'.
> 
> Charles
> ____________________________________________________________________
>  Below is my response to Peter Timaratz
> > regarding my differentiation between the two.
> > 
> > At 05:19 PM 8/4/98 -0500, you wrote:
> > >Peter G,
> > >
> > >I'm curious as to how you interpret the word discretionary. It seems
to
> me
> > >that if you are successful trading then you aren't just winging it and
> you
> > >are following a set of rules. I'd like to hear what for you is the
> > >difference between the two.
> > >
> > 
> > For example, Larry Connors and Linda Raschke's stuff has set rules, but
> is
> > not mechanical. Some people program every aspect of their trading into
a
> > computer system. If you trade off charts or trail your stop by visual
> cues,
> > this is discretionary. If you use a black box (some do!) and take every
> > signal, that is completely mechanical. You cannot use Elliot Wave
unless
> > you are discretionary; it just doesn't program. If you use a 7 and 21
> > period EMA crossover, with a 10% trailing stop after a floor of $400
has
> > been reached and a money management of $500, that is mechanical. I have
> > always been wary of putting a linear computer program into a non-linear
> > environment, but some people say that the rules keep you on track.. you
> > should never deviate. 
> > 
> > P