SSF indicator computation

Hi @twopirllc,

I’m reviewing the computation of the SSF indicator, and it behaves strangely at the begining of the input data.

Let’s compute and plot the SSF with a length of 10, using :

df.ta.ssf(length=10, poles=2, append=True, col_names=('SSF',))

That’s strange !

Skipping the first 10 points when plotting gives : which is (almost) what we want.

I beleive this behavior comes from this part in the computation :

    for i in range(0, m):
        ssf.iloc[i] = c1 * close.iloc[i] + b1 * ssf.iloc[i - 1] + a1 * ssf.iloc[i - 2]

Aren’t ssf.iloc[i - 1] and ssf.iloc[i - 2] undefined when i<2 ? No error is raised, so I’m confused. More generaly, I think that the SSF series should countain NaNs for i<length.

Could you please take a look ?

Thanks!

[DrPaprikaa]

Looks good to me ! Thanks @twopirllc

[twopirllc]

@DrPaprikaa,

Sounds good! No worries.

• length > poles
• for i in range(poles, m)
• ssf.iloc[:length] = npNaN

KJ