How Bot ChrisH Works

[ChrisH]

[Parasprites]

It sounds like your spawn control isn't actually very fancy after all. My implementation might actually be better. Though unfortunately, you've got a lot better combat and food gathering.

[ChrisH]

[zaphod]

Its still miles and months ahead of my spawn control (which simply does not exist)! Thanks for sharing your ideas. I plan to use these great ideas to develop my own bot as an exercise during free time. I've really started admiring the fascinating subject of AI after this contest and the Stanford AI class!

[Parasprites]

[ChrisH]

Here is a good example of how my spawn control helped me we a game between otherwise very closely matched bots.

You should take note that each bot starts with three hills on this map. Each bot has two hills relatively close to each other and a third hill farther away and relatively close to two hills belonging to one of the other bots. My bot wins because it is the only bot to keep all three of its hills. The other three bots all lose their isolated hill. My bot is able to hold onto its isolated hill because it directs a higher percentage of its ants to that hill in order to match the higher number of ants coming from the more numerous enemy hills in the vicinity. Meanwhile the other two hills don't need as many ants because they can support each other and are not as threatened by the smaller number of enemy ants in the area.

Although this strategy doesn't always work, it paid off in this tightly contested match.

http://aichallenge.org/visualizer.php?g ... 20&user=10

[Parasprites]

One thing I've been wondering is why you leave your blockers on the hill all the time. In my bot, I just try to keep the blockers within 1 square of the hill. As long as they are adjacent, they can block the hill on demand.

[ChrisH]

[Hippo]

Yes, I like your teleporting ... I was confused why bot's around 130 place try static defense and I decided not to do it.
The only question was the one "defender" in a role of teleport controller. I have noticed you are probably the only one using this technique. Having more time, I would probably implement it as well ... but I would start using it much more later, when the value of single ant is not such big.

[Parasprites]

I managed to implement it in less than a day on the final day of the competition. Though as usual, it took a while to work the bugs out.