talib.RSI gives incorrect values

ds55118

I have tried every possible way to calculate RSI, but the talib.RSI calculation disagrees with all other packages, libraries and methods to calculate RSI. Is there anybody who knows how talib.RSI calculates RSI values? Please let me know.

ptunney

TA-Lib is an opensource product, I am sure you can find the info out there on how to use it, how it works and as a final resort, get the code and look at what it is doing. https://www.ta-lib.org/

RSI itself is not a pinned down method, there are lots of different ways to calculate RSI.

This seems to work well for me.. It is important to pass it a lot of history.

    close = md.bar.daily(start=-100,include_empty=False).close
    RSI = talib.RSI(close, timeperiod=4)

ds55118

Thank you for the link, but there is no contact information there, and the community link is broken, so I cannot post a question there. The author claims that the calculation is based on Wilder's original formula, which is most likely a false claim.