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

Re:New indicator



PureBytes Links

Trading Reference Links

Dear Tomasz,
I understand that this is a convention.
Since we do not have value for tomorrow, we a g r e e to 
give to ref(c,+1) the last known close value.
The good fortune for all of us, Amibroker users, is that you have 
overcomed such conventions with moving averages.
I used for a while a very nice program, with many and interesting
built-in functions, (trend, sintrend, fft etc), with a friendly 
Language, quite professional, which made a terrific mistake in 
calculation of moving average
" pads the series with zeros at the beginning and end before 
computing the average" as referred in the well prepared manual.
The result was very strange : The close and the average graph
were beginning and ending at same points, making the use
of this common function totally inapplicable.
May be unbelievable, but it is true.
(The cost of this software was 2500 USD one year ago).
Could anybody pay for Amibroker 2500 USD and have ma(c,10)
or ema(c,10) wrong for the last 10 days by default??
And properly explained in the Manual??
It is better sometimes to estimate what we have got here.
 
Best Regards
Dimitris Tsokakis
--- In amibroker@xxxx, "Tomasz Janeczko" <amibroker@xxxx> wrote:
> Hi,
> 
> >Because, whithout data for tomorrow, the program gives for
> >ref(close,1) the available today´s price.
> >(In my opinion, it should indicate an {empty} value, because
> >ref(close,+1) exists for all previous days except the last one)
> >I do not know how it is designed, I know how it works.]
> 
> Yes it gives today's price for the last bar as we don't know the 
future yet :-)
> The reason why it doesn't return {empty} is that some of the other
> functions may be confused if they see {empty} at the end of the 
array
> (normally {empty} values appear only at the beginning and all AFL 
functions
> are aware of this fact)
> 
> Also: I agree with Dimitris that writing trading systems using Ref 
with positive
> offsets (looking into the future) would give results that could not 
be 
> reproduced in real life. (unless you have time-machine :-)))
> 
> 
> Best regards,
> Tomasz Janeczko
> amibroker.com