chapter 01-g-h-filter.ipynb - wrong results demonstration for filters with the same data

[afanasyspb]

Hi!
Compare the plots in the fields: in [17] (Building a filter) and in [22] ("Solution and Discussion" for g-h filter). They use the same dataset. But if in the previous plot the measurements start at day 1 and "Filter" value goes from 160 (day 0), the latter plot shows "actual weight", "filter" and "measurements" at the step 0. It is better to shift the "filter" and "measurements" values to the Step 1 on the plot, or at least change the "actual weight" value from 160 to 161 at the step 0.

[rlabbe]

Good catch, fixed.

[fpertl]

It appears that now (current version) the estimates no longer fall between the prediction and the measurement for a given day. In the 1st plot above they still do.

[ydixon]

Jesus. This was driving me crazy as I'm totally new into this topic. I was scratching my head why we are making estimate using measurement from Tk-1. I'm glad someone mentioned it.

[rlabbe]

Fixed via #343.