Overfit Wiggly Skaters?

[morash-sean]

I'm trying to implement some wiggly skaters on a data set with about 20 observations and it seems to be overfitting to the original data. Am I giving it too much foresight with the y parameter or is 20 observations just insufficient to get a meaningfully smoothed outcome?
y= random.sample(range(10, 30), 20)
def posteriors(f,y):
s = {}
x = list()
k=1
a=0
for yi in y:
xi, xi_std, s = f(yi,s,k,a)
x.append(xi[0])
return x

[microprediction]

I can't tell from your code which f is being used. Can you share a colab?