That is a really high curve, tho.
I have tuned it to where your note appeared, and that's all that matters. each user of the algorithm will have hundreds or thousands of data points to fine tune their own custom model
It turns out, if you take out the algorithms, nearly everyone tunes into the same model.