>>11031276The thing that bothers me with this is how many parameters we are fitting.
First we have 6 parameters for the numbers that come from nowhere: 6, 11, 12, 4, 10, 15
Then we have (arguably) one parameters for the multiplier, as it starts from 2 and not 1.
And then the structure of the formula.
So from 8 given numbers we inferred 1 formula and 7 constants.
From a machine learning / statistics point of view, I think it's in the danger zone of overfitting.
