Robot Arena – Failure

This week I was working on a roguelike with a working title of Robot Arena. It was inspired by the programming and racing mechanics of the RoboRally board game. You program your robot with 5 instructions and then watch while they are executed. You are racing against 3 competitors through a dungeon. The dungeon is filled with conveyor belts, lasers, spikes, smashers and monsters.

I wanted to encourage the type of play where you are always thinking several moves ahead even without perfect information. I also wanted to set up some tension between running at the front of the group and facing more enemies or hanging back and passing at the right time. There were other choices about risk reward like should I take a shortcut through the laser even though I take damage.

I spent a fair amount of time on it this week which was documented on a blog. Ultimately I didn’t make it far enough to get all the features finished and tested. It still needs some work to call it a complete game. I felt I was making good progress but I struggled quite a bit with the racing AI. I was also trying out a component entity system which I did enjoy. I learned a lot, had a great time and am looking forward to doing it again next year!

I have a working prototype which can be downloaded here. Here is a screenshot of how it looked at the end.

End of Day 7 - conveyors - Upload

Congratulations to all of the challengers this year. I have been checking out some of the games and you really outdid yourselves!

3 thoughts on “Robot Arena – Failure”

    1. Great name! I was struggling to come up with one. I was excited to try out your game. I made it to level 4. I like the way you put everything together. The shuffling cards and inserting them into the board are all nice touches. The AI seems pretty good for the confusing board. I could easily see them getting lost with the turns and movement. Did you have any specific pattern or algorithm you used? I was struggling for a bit with mine. You definitely went for a tighter approach which I think is good for a 7DRL. Great job!

      I was sad about not finishing too. It is playable but it needs more balance. Right now its too easy. I recently changed the opponents to not be directly attack-able which I intended to make it harder to beat them since you had to bump them into things but that didn’t work out so I may change it back.

      Thanks for contacting me. It was fun to see your different translation of the same theme.

      1. So, the AI bots are playing more or less the same game as the player. They also have a deck of cards which they can select from while you make your selections.

        They make a decision by randomly selecting hands, simulating how the next turn will play out (naively assuming no other robots input orders), and rate each possible hand on the basis of 1) How much they managed to shoot the player, more is better 2) How close they got to the player, closer is better, and they pick the best of these random possibilities when the player is ready.

        This means they play rather like a mechanically well-versed, but strategically very naive human would in the same roll.

Leave a Reply