![]() A great improvement indeed as Nick’s simulations demonstrated that this Hunt algorithm completes 50% of games within ~65 turns, as long as it is not defeated beforehand. Every human who has every played Battleships will do this intuitively. ![]() This improved algorithm includes an instruction to explore nearby spaces whenever a prior shot hit. Basically, it would lose from the average toddler, given that the toddler would survive the boredom of playing such a stupid AI.Ī first major improvement results in what is dubbed the Hunt algorithm. ![]() At best, the expertise level of this AI would be comparable to that of a distracted baby. Nick simulated 100 million games with this random apporach and computed that the algorithm would require 96 turns to win 50% of games, given that it would not be defeated before that time. The AI resulting from such an algorithm would just pick a random tile to shoot at each turn. The first, stupidest approach is to just take Random shots. The resulting levels of artificial intelligence (AI) seem to compare respectively to a distracted baby, two sensible adults, and a mathematical progidy. ![]() The story starts with this 2012 Datagenetics blog where Nick Berry constrasts four algorithms’ performance in the game of Battleships. I thought they might amuse you as well : ) Past days, I discovered this series of blogs on how to win the classic game of Battleships ( gameplay explanation) using different algorithmic approaches. ![]()
0 Comments
Leave a Reply. |
AuthorWrite something about yourself. No need to be fancy, just an overview. ArchivesCategories |