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

Re: CL_Modifying system´s parameters



PureBytes Links

Trading Reference Links

> I am a "pasive" omega-list´s reader, I mean, I dont use to send
> too much e-mails. Maybe because I am Spanish and sometimes I cant
> express correctly what I want to say. 

Jose, your English is *excellent*.  I think you express yourself very 
well.

> The question is: how can we determine when we have scenario 1 or
> scenario 2. I mean, if you have backtested you system and know that
> when volatility raises, the optimal parameters of the system are
> (for example) 1 and 2, and in choppy markets the optimal
> parameters are 2 and 1.5, how can we do to estimate the current
> market scenario and so to determine which parameter to choose. 

Do you have a good programmed way to detect the difference between 
"volatile" and "choppy"?  If you do, then you could do something like 
this:

Inputs:  K1vol(1.0), K2vol(2.0), K1chop(2.0), K2chop(1.5);

Vars:  K1(0), K2(0);

{ VolOrChop returns True if volatile conditions, False if choppy }

if VolOrChop(x,y,z,whatever) then begin
  K1 = K1vol;
  K2 = K2vol;
  end
else begin
  K1 = K1chop;
  K2 = K2chop;
  end;

Or, if you have a a continuous-function result that determines the 
volatility vs. choppiness of the market, maybe you could get better 
results by computing optimal values of K1/K2 for the current market 
conditions.  For example, if VolChop returns an index from 0 to 1.0, 
where 0 = totally volatile and 1.0 = totally choppy, you could use it 
in an expression to compute K1/K2.  One way would be to predetermine 
the vol/chop values of K1/K2 as in the example above, and use those 
in the equation:

  K1 = (1-VolChop)*K1vol + VolChop*K1chop;
  K2 = (1-VolChop)*K2vol + VolChop*K2chop;

...or maybe you could figure out a better way to compute K1 and K2 
based on the value of VolChop.  You'll need to figure out what works 
for your specific situation.

I'm sure there are better ways to do this, but there are a few ideas 
for you to think about.

> I am testing the system I enclosed as an ELA file. 

This code looks at the highest high, lowest low, etc for the last 
Mday or Nday bars.  Since you said you traded this intraday, I'm 
wondering if you have an error here.  You use values of Mday=1 and 
Nday=1.  That means you're only looking back one BAR.  Did you intend 
to look back one DAY, or did you perhaps want the highest high &etc 
that has appeared so far today?

Gary