Can someone help me answer and understand the "why"? I am trying to teach myself.

In each case below, indicate whether a more flexible model (like a high- degree polynomial) would tend to perform better or worse on a test set than a more inflexible one (like a low-degree polynomial). Explain your reasoning. (a) The number of observations n is large and the number of predictor variables is small. (b) There is a very large number of predictor variables, but a small number of observations n. (c) The apparent dependence of the response on the predictor vari- ables is highly non-linear. (d) The variance of the error terms is very high