My first bot for Halite.IO has played over 250 games!
I’m currently averaging the 250 rank in Halite, it’s doing better than expected for my first bot, I will continue to develop onto this version of my bot in hopes to get above the 200 mark.
My next step to improving the bot will be first of heuristics, the pathing to the boundaries of my owned areas is quite wasteful at the moment, as it doesn’t take into consideration of the strength of the cells that I am using to move from point A to point B, which decreases my maximum output.
Second step will be taking the newly destroyed areas into consideration when moving, at the moment the bot is recognising the empty cells as capture-able but due to it being on the border it will almost always try and take it, regardless of the enemy strength nearby.
A good example for both these issues are here.
What will I do after that?
Well, I will have to wait and see how my bot performs with its new algorithms and I will see what the most flawed behaviours are, off the top of my head I can see strength maps becoming handy (knowing where the low cell strengths are in the map) so I can get cheaper production.
Secondly I will probably need some pathfinding algorithm to move my cells, this will likely replace a lot of my movement algorithms and heuristics I will be using already.