Intro:
With the new football season just around the corner, now is the perfect time to start planning your Fantasy League journey. Dust off your managerial cap, embrace the strategizing, and dive headfirst into the world of Fantasy Football.
Rules:
The aim is to pick a squad of 15 players, including 2 goalkeepers, 5 defenders, 5 midfielders, and 3 forwards for a budget caped at 100 Million. The objective is to build the best possible team within the budget and earn points based on the performance of the selected players in actual Premier League matches.
The Excel solver has a limit of leveraging only 200 decision variable to had to do a bit of feature engineering.
DECISION MAKING : EXCEL SOLVER:
a) The Decision variable would be binary to select a player for the team or not (Highlighted in Yellow cells)
b) The object function is to minimise the total points scored if the players is selected
c) Constraints:
i) Cost: Sum of all players selected < 100 Million
ii) Count of position for GKP = 2
iii) Count of position for DEF = 5
iv) Count of position for MID = 5
v) Count of position for FWD = 3
vi) Sum the count of player selected from a team < 3
Setup in Excel Solver:
Result:
| name | team | position | cost | total_points |
|---|---|---|---|---|
| Raya | BRE | GKP | 5 | 166 |
| Alisson | LIV | GKP | 5.5 | 162 |
| Gabriel | ARS | DEF | 5 | 146 |
| Mings | AVL | DEF | 4.5 | 130 |
| Mee | BRE | DEF | 5 | 143 |
| Schär | NEW | DEF | 5 | 139 |
| Trippier | NEW | DEF | 6.5 | 198 |
| Martinelli | ARS | MID | 8 | 198 |
| Ødegaard | ARS | MID | 8.5 | 212 |
| Douglas Luiz | AVL | MID | 5.5 | 142 |
| Gross | BHA | MID | 6.5 | 159 |
| Eze | CRY | MID | 6.5 | 159 |
| Watkins | AVL | FWD | 8 | 175 |
| Wilson | NEW | FWD | 8 | 157 |
| Kane | TOT | FWD | 12.5 | 263 |
Summary:
Having learned on how to optimize the chance to score more points, enjoy the exciting way to immerse ourselves in the world of football management. May our Fantasy teams reach new heights, and may the spirit of competition unite us in the love for the beautiful game.
Reference:
FPL Player Status Data
Fantasy Premier League
Top comments (3)
No Man City players.... Interesting.
No Nottingham Forest players..... Unforgivable 😉
with the constraint on budget Man City players were not value for the money based on the model and the data ...
Can share the model with bias to have nottingham player if you need :-)...
the decision making should not have emotion attached.
I could imagine default human behaviour to go straight to winning team for first picks, and then their team.
Removing that bias is what makes models so powerful, though would still have Gibs-White in there 😉