Website development

[Henry Harrison]

I maybe mentioned in other post that a website was more of a long-term goal but as I've been giving it more thought that may not be the case. I think there is a real possibility of offering something new with a website and then incrementally adding the more common features of other websites. Here's a rough outline of a potential strategy:

0. A tool that displays player information in a table and allows the user to add players to lineups and create a saved portfolio of lineups for one particular slate.

1. Provide simple information about the portfolio's composition, e.g. the percentage of lineups the user has made that include each player. This seems to be "low-hanging fruit": it requires no data access or statistical analysis and I know of no websites with this feature.

2. Provide an interface for projections, either scraped from some other site (FantasyPros or Numberfire, for two easy options) or uploaded by the user and include these projections in the interactive player table.

3. Implement a linear optimization tool to generate lineups, a la FantasyCruncher or LineupLab.

4. Allow the user to specify in advance the desired percentage owned of each player, add these choices as constraints to the linear solver (alternatively this could be done with a probabilistic technique, with the likelihood of selecting each player adjusting as each lineup is created), and thereby generate portfolios of lineups.

5. Provide realtime information on the lineup's performance during games. This is actually relatively easy for NFL thanks to nflfan (https://github.com/BurntSushi/nflfan).

Just thought I'd put his on paper while it was in my head.