$(4$ points$)$

An analyst claims that any adjustment to a model that reduces the percentage error on the training data is always beneficial to the model. Is the analyst correct? Why or why not?

$(5$ points$)$

Consider the following fitting curve. Which level of model complexity would you recommend? Explain. $($Report the approximate level of complexity based on visual inspection of the figure.$)$

Figure $3.$ A fitting curve showing model accuracy for training data $($blue$)$ and testing data $($orange$)$