Peer Review- rz98

[ZRich098]

The Dream Team project's main goal is to create a model that can best predict the performance of fantasy football players. The group plans to use data provided by Fantasy Football Data Pros, which includes the projected and actual performance of players from 1970 to 1999. By doing so, the group hopes to generate an accurate model which can be used to maximize the chances of winning fantasy football for bragging rights or gambling with money.

Three things I like about the project are:

1. If successful, this project has the possibility of making serious amounts of money for the creators. Fantasy Football is very popular and has prize pools of over a million dollars for winners on certain platforms.
2. The dataset for the project is very easily accessible and likely contains very well formatted and accurate data.
3. The dataset is very large and there will be plenty of data to learn and train the model on.

Three areas I think could be improved:

1. The dataset being examined is not incredibly feature rich. There are not many relevant features, so it may be hard to predict actual performance based on only features such as projected performance or position. I would suggest looking into datasets with more complete information such as number of touchdowns or performance in kickoffs, etc.
2. The group mentioned only using the data from 1970-1999, which is not very current. Fantasy Football Data Pros provides data from 1970 to 2019, so I would suggest using the more complete dataset instead.
3. Fantasy Football predictors are comparable to stock market predictors in the sense that it is very hard to create ones that have a high performance. Even though player performances can be modeled, actual fantasy football has lots of arbitrary rules to prevent the system from being abused for profit, or else lots of people would be cashing in.