Jordan Lord

Footium Part 1 - Auto Manager

Introduction

A small disclaimer, this is purely a Footium strategy that is used for testing in a test environment. This is not my strategy in the live game, there are stringent rules placed on the way that team members can play the game.

To make a concession, I'm poor at playing Footium. I founded Footium in 2020, as you may see from projects on this website, it's been a dream of mine to build a football simulation game for some time. Just recently in season 1, my beloved Hightock Albion missed out on promotion by four points. Meanwhile, my great colleague but also fearce rival Nick, won promotion with Barch Wanderers. This is a similar story in our test league. Nick and his hoard of clubs have managed to win TODO% of all seasons. Pock Town, my test team, have had moderate success, winning three trophies overall (TODO fact check).

It's time for that to change. My strategy is to utilise all of my test teams in the Footium test league pyramid to, not only win the league, but to dominate the entirety of Division One. In the space of four seasons, Pock Town and the associated clubs will come to dominate the league. The strategy is to use bots in order to achieve this.

League Rules

Our test league can be quite sleepy at times, especially during live seasons. A season plays out in one full week. We have 4 days of matches, 3 days of in-between season downtime. Each day has 6 matches, which make it quite intense to manage. There are at least two matches per day where it's feasibly impossible for somebody to be live managing their club. If I can write a bot, it gives me a significant advantage.

The matches play in the following schedule:

  • 9:15
  • 13:15
  • 17:15
  • 21:15
  • 1:15
  • 5:15

Each season is broken up into "Sprints" which run in parity with our development cycle, where we have one Sprint per week. The current season that is ending is Sprint 85. Pock Town have finished in seventh position after quite poor management, narrowly avoiding relegation on the final day.

There are 6 divisions, each team member is given an equal share of the 756 clubs. Here are a list of my clubs and where they currently sit:

TODO

My Strategy

The strategy is to engage every single one of those clubs with an auto-management bot. Each bot will have a slightly different strategy, which will be defined by the following parameters:

  • Target squad size - this will determine whether they sign academy players or not, for the first number of seasons I'd expect all clubs to sign academy players to make up for the gaps
  • Squad evaluation function - each match-day squad will be evaluated differently by each club
  • Training strategy - each club will have its own training strategy, some will prioritise training younger players, others will prioritise training their best players, some randomly, others their older players and so on

Another challenge is going to be that most of the clubs don't have any players. Because they've gone unmanaged for a number of seasons, almost every club has all retired players.

Each match's lineup will be set 15 minutes before the start of the game. No in-game management will happen for version 1.

Version 1

Version 1 involves just trying to choose the best team before a match, while also having a standard training strategy to train the youngest player will available potential that's not in a match day squad.

For the sake of simplicity, there will be a hot wallet that will own all of the clubs and players. The script will have access to that wallet's private key via an environment variable. To aid the transition, we'll need to write an auto registration script. We'll employ a simple rule, that rule will be to register players based on their originClubId.

  • for each club
    • if any players that are owned by the same wallet, same originClubId but don't own the player, then register that player

We'll also be automatically signing academy players if the size of the squad is less than 30 and there are academy players available to sign. I'm assuming here that I have enough Sepolia Eth to sign the players.

Season One (Sprint 86)

It was a bit of a rocky start. It turns out there was an issue with the auto manager, alongside it being release week for Footium, which meant I didn't have much time to fix. The issue was with the player cache I was using, it was out of date and wasn't repopulated. This meant that teams were submitted not using the newly signed academy players.

In addition, it meant that teams from other clubs that had been transferred weren't being used. This found a regression in our lineup validation. The validation allowed players from other clubs to be submitted in a lineup for my club. This meant that my tactics page wouldn't load, nor would matches start for my clubs that had the cross pollinated players.

At the end of the day, Project Hawking lost 5-6 full matches at the start of our first season. Pock Town sunk into the relegation zone with a few games in hand to make up for.

  • Pock Town finished 3rd
  • most teams got promoted

Season Two (Sprint 87)

Season two has started. We've now got two teams in the top division, Pock Town and Wiversing.

It wasn't until Tuesday 17:30 that the auto-select bug was found.

Then on Wednesday 09:00 another upgrade was made where we were now properly invalidating the player cache, so that the auto select should more accurately select the best team. In addition, auto select should now fail a lot less. There were instances where auto select was failing to submit lineups because player had just been put in training.

Once that was set, the teams performed a lot better. A number of clubs got promoted out of their respective division. Pock Town and Wiversing remained in Division One. Sild City and Binchellinst got promoted to Division One. Sadly in Division Two - League One, Regingsey didn't make it. They got pipped to promotion by Chitnorwore and Moodburn, both Division One giants.

This poor competitive performance can mostly be attributed to the low depth of squad of these teams, combined with the inability to react to invalid lineups.

  • final league standings

Season Three (Sprint 88)

In Season Three, we started off with 4 Division One Clubs Binchellinst, Wiversing, Pock Town, and Sild City. And 8 Division Two clubs, Baton United, Bexham Rovers, Caistord, Crough Wanderers, Helver Rangers, Lyn an Villa, Totlest Albion, Wattinstasins, and the poor Regingsey. No clubs remain in Division Six.

At this point we've now ran out of ETH with only 0.16 Sepolia ETH left in the wallet. We must be more sustainable moving forwards.

  • lineup evaluation
  • more efficient training

Season Four (Sprint 89)

  • Pock Town Champions!
  • Baton United and Regingsey relegated

Season Five (Sprint 90)

  • Pock Town second
  • Binchellinst Champions!

Season Six (Sprint 91)

  • revival of Nick
  • Beverle Champions
  • Hawkbot stops working properly

Season Seven (Sprint 92)

  • Beverle Champions again
  • Hawkbot continues to not run

Season Eight (Sprint 93)

  • TODO
    • open sources
    • revival of Hawkbot
    • cache fixes
    • formation chooser
    • squad re-balancer
    • back-to-back training