Bronze Strategy
My Bronze bot was still mostly heuristic, but it was no longer just a basic harvest-and-drop bot. At this stage, the strategy started to combine several mechanics: farming, training, fruit management, planting, chopping, and a first attempt at converting fruits into wood.
The main idea was to build a small economy around useful fruits. The bot tried to decide whether a fruit should be brought back to the shack for training, or whether it could be used as planting fuel.
Core loop
The basic economy was still based on collecting resources and bringing them back to the shack:
move to tree → harvest → return to shack → drop
However, Bronze added an important distinction: not every fruit had the same purpose.
If a troll was carrying a fruit that was needed for the next training step, the bot tried to bring it back to the shack. If the fruit was not immediately needed for training, the bot could use it for planting instead.
So the logic became closer to:
harvest fruit → if needed for training, drop it → otherwise plant it
This was my first attempt at making the economy less passive. Instead of only collecting what already existed on the map, the bot started to use fruits to create new trees.
Training
Training was already present before Bronze, but in this version it became more connected to the economy.
The bot checked the shack inventory and tried to train additional trolls when enough resources were available. The goal was to grow from a small number of workers into a more stable multi-troll economy.
The training was still basic. I was not using a sophisticated build order or adapting troll profiles deeply to the map. The main idea was to create useful workers and improve resource production.
The bot therefore had to decide which fruits were important for training and should be preserved. This is where fruit management started to matter.
Fruit management
One of the main improvements in Bronze was that the bot started to reason about fruits.
A carried fruit could have two different roles:
-
Training resource: bring it back to the shack.
-
Planting fuel: plant it on a useful cell.
This meant the bot did not blindly drop every fruit. If a fruit was useful for the next troll, it was brought back. If it was not needed, the bot could try to plant it instead.
This was still a simple rule-based system, but it introduced an important strategic idea: fruits were not only points or resources, they could also be converted into future value.
Planting
The bot started to plant trees when it had a fruit that was not required for immediate training.
The preferred planting spots were useful cells, especially cells near water when possible. Planting near water was valuable because it could improve the growth cycle and make the tree more useful later.
At this stage, planting was still opportunistic. It was not yet a fully planned economy, but it was the beginning of a plant-based strategy.
Chopping and wood
Bronze also started to exploit wood.
If a troll had chopPower > 0, it could move to a tree, use CHOP, collect wood, return to the shack, and drop it.
The wood loop was:
move to tree → CHOP → collect wood → return to shack → DROP
This was important because wood was more valuable than basic fruits. The bot started to understand that a fruit could sometimes be more valuable if used to create a tree, then converted into wood.
This was the early version of a plant-chop-drop idea:
fruit → plant tree → chop tree → drop wood
However, this system was still basic. The bot did not yet have a strong long-term plant-chop-drop engine. It was more of an opportunistic conversion than a fully optimized strategy.
Target selection
The bot still used simple heuristics to choose targets.
It looked for useful trees depending on the situation:
-
trees with fruits when it needed resources;
-
trees that could be chopped when wood was useful;
-
planting opportunities when carrying a fruit not needed for training;
-
shack-adjacent positions when carrying resources.
This made the bot more flexible than the earlier Wood versions, but the decisions were still mostly greedy.
Summary
In short, my Bronze strategy was a greedy economy bot with basic training and early fruit management.
Compared to the Wood versions, the bot started to treat fruits differently depending on their usefulness. Fruits needed for training were brought back to the shack, while less useful fruits could be planted and later converted into wood through chopping.
This was not yet a complete strategy, but it was the first version where the bot started moving from simple harvesting toward a more interesting economy based on training, planting, and wood conversion.
!
) was the “reckless chop” approach where you chop without gathering the wood, preventing or retarding opponent training. I stop the agression if the opponent started gathering wood, or have more troll that me.
It was very instructive for me so thank you for this opportunity!! But this contest, will it be accessible in the near future? personnaly I wanted to continue a second approche that i began the last 2 days.