Experiment documentation & adding a constant -- how to handle at forecast time?

[afogarty85]

In this example, we add a constant to avoid computational problems: https://github.com/Nixtla/statsforecast/blob/main/experiments/ray_ets/experiment.py

Noting:

# add constant to avoid numerical errors
# in production settings we simply remove this constant
# from the forecasts

Given a forecast and a prediction interval, how do we remove this constant?

Consider this forecast:

1. Forecast: 8, Low 90: 14, High 90: 22
2. Forecast: 0, Low 90: 0, High 90, 21

To correct the effect of the constant, do we subtract -10 from all values, to include the prediction intervals, so long as to not yield a negative answer?