I was just doing a pracspec X-Y plot to convince myself that RSI is RANDOM ( and it IS regradless of implementation) the nice thing about rsi is that is has less correlation with reality that say momentum... ----- Original Message ----- From: "Chris Cooper" <chris@xxxxxxxxxxxxxxxxxx> To: <tech-spec@xxxxxxxxxxxxx> Sent: Thursday, October 14, 2004 1:22 PM Subject: [tech-spec] Re: Correctness of R functions and models > Yes, RSI has historically had different implementations. Most of > the differences are due to interpretation of Wilder's original > writings, and to substitution of exponential average in place > of simple moving averages, etc. If you really care, you have > to look at the code itself. If you really care, then perhaps > your system is overfitted and you should reconsider the > whole thing! > > > -----Original Message----- > > > > I have been testing the rsiTA implementation of R and it > > seems to differ > > from the one I wrote and the one in wealthlab developer. It > > seems that there > > are discrepancies between the definition of RSI. > > > > when i do a X/Y diff they give completely different results > > ...(not a linear > > relation) > > > > How do we know all these functions dont have errors? > > > > > > > > > >