| Most times the dificulty can be adjusted. But it's also true that most times, the dificulty level is just a number that tells the program how deep it can go in a tree.
Data trees are tipical data structures used on IA. The programs calculates all the possibilities and their future consequences, storing them into a tree. Then, each node on the tree is evaluated using rules, and the winner node is the one that is chosen.
The dificulty level often tells the program how deep it can go while calculating nodes. The problem with this method is that it's not really inteligent, nor realistic. If you instruct your program to calculate only two future moves then it's idiotic. And if you configure it to foresee 30 turns then it's just impossible to beat, unless you are Gary Kasparov.
There must be chess programs that might use more inteligent algorythms that are not based on this principle. Though I can't tell you which one because I am not a chess master myself. |