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

Re: GEN - PFE



PureBytes Links

Trading Reference Links

Neal,

Thank you very much for the ela files. I've spent a few minutes playing
around with them already. It seems you are right about the imput that
smooths the curve. The curve changes and becomes very interesting below 1
however - try entering 0.1 for EMALEN (on daily data). The curve looks like
a digital representation of a sound wave. It forms a triangle for each trend
(wave) that terminates when the trend stops. Anyway, now that I have it as a
function I can start working on some systems.

Thanks
Simon Trevor


>Simon,
>I have converted the PFE to a user function and am attaching two ELA
>files, one with an indicator and function and the second with just the
>function (PFEUSER).
>This function has two inputs: 1. Lookback period (default=9) and 2.
>EMALEN.
>EMALEN is the length of the exponential moving average that smooths the
>raw PFE.
>The standard PFE uses .333 as a smoothing factor, that corresponds to an
>exponential moving average of 5 days (2 / (5+1)).
>Both of these inputs can now be changed, but in changing them the
>indicator does not change to any great degree.
>Regarding the loss of the sawtooth waveform as the time frame increases, I
>tend to agree with you regarding the "movement in the price."  The input
>parameters are not a factor.
>My guess is that the sawtooth waveform is produced by the lower relative
>volatility of intraday versus daily data.  I will be doing some research
>into this to check whether intraday data has a lower relative volatility
>than daily data.
>Neal
>
>----------
>> Only on some of my daily charts does the sawtooth pattern exists. It
>seems
>> to be related to how much movement there is in the price and how smooth
>the
>> movement is.
>> 
>> I would like to be able to system test the PFE and this cannot be done
>from
>> an indicator. As I only have Supercharts, I cannot convert it to a
>function
>> that a system can use. Would someone be kind enough to change it to a
>User
>> function?