Kase Peak Oscillator

[[Deleted]]

I was wondering if anyone has written an indicator in MQL4 of the Kase Peak Oscillator (KPO). I have found mql4 codes based on the following algorithm:

KPO = RWI(high) - RWI(low)

where

RWI(high)=(High[0]-Low[n])/ATR*sqrt(n),

RWI(low) =(High[n]-Low[0])/ATR*sqrt(n),

and ATR is the average true range over n cycles.

However her published work Article 1, and Article 2 suggest that the formula should be sth like this:

KPO = (KSDI(up) - KSDI(down))/volatility*sqrt(n)

where

KSDI(up) = Log(High[0]/Low[n]),

KDSI(down) = Log(High[n]/Low[0]),

and volatility is the standard deviation of the logarithmic price change over n cycles (ie. STDEV(log(C[n+1]/C[n]) ).

The main differences between the two algorithms is that the first uses the linear change in price while the second uses the logarithmic change in price. In addition the first divides by the ATR which is an average of the price fluctuation while the second divides by the standard deviation.

Has anyone used the second algorithm?

Another part that I haven't seen in any code or perhaps I haven't understood is the self optimization of the Kase Peak Oscillator (KPO). Cynthia says that the KPO is self-optimized over 50 different cycle lengths. This is done by looping back over an internal n and finding the maximum values of KSDI(up) and KSDI(down) over that interval. How do you define the interval? In the following links

http://www.aspenres.com/Documents/pdf/KaseStudies.pdf

http://www.aspenres.com/Documents/AspenGraphics4.0/Kase_Peak_Oscillator.htm

which are manuals for her indicators they list 3 parameters, cycle range low-8, cycle range high-65 and scaling factor-50. I guess the cycle range low-high are the indices of the lookback loop that maximizes KDSI(up) and KDSI(down) but what is the scaling factor then?

Thank you in advance,

Nikos