Week 2/Question 9

I have a question… I hope it does not violate rules.

in the feedback of question 9, it was referred “If your dataset was infinitely big, 2.2% would be a perfect estimate of the improvement”

I thought it would be maximum not perfect. please provide a more intuitive explanation that why when the dataset is infinitely big, it would perfectly be 2.2%.

That basically means – in the best case scenario. Observing all the data usually infeasible for real-world tasks. This question in some way related to your other question here.