Over the first few weeks of the season I've experimented with different methods of lineup building and I've found the most successful method for me is not to be building lineups too early, unless it is a cash game lineup.
My reasoning for this is too often you will find yourself over adjusting once rosters come out. This can be for players you had your eye on and end up starting, or players you had in your lineups that end up benched. Too often I've completely overhauled line ups when a player I wanted isn't playing, or not gotten enough exposure to players I was really high on who ended up getting starts.
So my process going forward for line up selection will be this:
- Identify the fixtures I want to target. This is done by looking at odds as well as using Tango method for regression to predict expected shots.
- Look through my predictive model see what players rate the highest from the teams in the fixtures from step 1.
- Compile a list of these players separating them into differing price brackets as well as identifying which players I am most bullish on.
- When lineups get released an hour before game time look for non usual starters and cross check their stats in case they aren't included in my spreadsheets due to not having enough minutes played. This cost me missing Willian when they played Villa.
- Begin building lineups based around the games and players I want to target.
Handling things this way tends to cut down on me making too many lineups. It also allows you to start from scratch and build your lineups around the players you are most bullish on. When you do things this way instead of trying to insert them into your already made lineups you will find that you are over adjusting far less to accommodate the players you want.