|
PureBytes Links
Trading Reference Links
|
Adam Hefner pointed out:
> Owen,
> It might help if I could see your exact code.
>> Today I tried to average three of these things, and it
>> just doesn't work. The indicators (based on the RSI)
>> give the expected wiggly lines. Add them up, and divide
>> by three, and what I get is a constant, a nice, straight
>> line! In fact, simply adding them up produces a constant
>> -85. Using the Add() function unstead of a simple plus
>> produces the same thing -- not that I doubted it would.
I didn't bother including them, because I suspect this is a
general problem. However, the indicators I was playing with
this time were simple %B of the RSI. To use pseudo-code:
100 * (RSI - BollingerBandBottom) / (BBTop - BBBottom).
All I wanted to do was:
(%B1 + %B2 + %B3) / 3, where the period of the RSI and
the Bollinger band length varied from one %B to the next.
It seemed little enough to ask.
What came out was a straight line, which was outside the
range of %B values I was trying to average. I tried varying
PerN and BBLenN, without effect. Even the value of the
constant output remained the same.
The same thing happened the other day when I was playing
with combining several different oscillators in the same way.
Eventually, I stopped editing that indicator and set up another
in which to play with the various factors that entered into the
equation. That time it worked, with no obvious explanation,
as I hadn't done anything different. Unfortunately, I don't even
remember which oscillators were involved. It wasn't anything
serious, just a way of clearing my mind when overburdened
with other work.
I appreciate whatever help anyone can offer.
Thanks again.
Owen Davies
|