tomm1111 197 posts msg #66066  Ignore tomm1111 
8/11/2008 1:36:04 PM
Anybody know the actual rsi(2) equation SF uses? Using only 2 bars returns alot of 0's and 100's, however, that is not what a SF rsi(2) filter returns.
The following rsi equation doesn't work.
RSI = 100  100 / (1 + RS)
RS = AG / AL

tomm1111 197 posts msg #66094  Ignore tomm1111 
8/11/2008 8:33:06 PM
Bump.
Smart guys, SF. What's the actual RSI(2) equation used?

tomb 261 posts msg #66097  Ignore tomb 
8/11/2008 9:42:20 PM
Hi,
The RSI is a cumulative measure where the current value is based on all previous values. On StockFetcher, a measure such as this is "seeded" with at least 2 years of historical values to compute the most recent value. In other words, an RSI(2) is not just based on the last 2 periods (as a simple moving average would be.)
As you noted, Wilder's RSI formula is:
RSI = 100  (100 / (1 + (AG/AL)))
However, note that the 'AG' and 'AL' components are not derived from simple moving averages, but each are computed using Wilder's smoothing method:
X(i) = [X(i1)(n1) + Y(i)] / n
Where X(i) is your current average, and Y is your input data series. The above is very similar to an EMA, and as you can see requires all previous values to compute the current value.
Tom
StockFetcher.com Support

tomm1111 197 posts msg #66099  Ignore tomm1111 
8/11/2008 10:52:31 PM
Cool. Thanks Tomb for the reply.
I think I remember running into that equation at some point. I'm going to look into it further.
Tom
