While running this notebook, the delta between the 1st and 2nd best fit, is smallet than 3*std of 1st best fit.
Therefore the comment was not valid.
Is this expected?
Perhaps because of variations associated to the random values picked for the grid?
Indeed, we set the seed such that the explanation stands but it would require running the notebook linearly and not repeating some cells. We tried to make sure that the randomization would not have any effect but this is not always easy.
Thanks for reporting.