Fantasy Football Optimizer

Overview

Fantasy Football sites support games where users select virtual football teams from among actual NFL players and score points according to the actual performance of their players. These sites typically allocate a fixed budget (or salary cap) for the team that you select and also price individual players according to their expected level of performance. The objective is for users to draft their team by allocating salary in such a way as to optimize game-day scoring. This project is to generate an online web app to assist users in generating optimal teams.

Example

Consider the virtual team of Figure 1. A user has selected the team based on their expectation of point totals for each player for the upcoming week. An optimal algorithm will select the team that will generate the maximum point total while coming in under the salary cap.

Features

Part of the project is to develop a feasible optimization algorithm. It seems likely that a straightforward greedy approach can be applied in order to generate (at the least) an approximate solution. Linear programming may also provide a feasible approximation to the optimal solution.

A significant portion of the project is to be able to generate accurate point-total predictions for each player for the upcoming week. There are numerous potential techniques for making such predictions. Methods that include markov-chains or simply averaging past performance.