It looks like Monte-Carlo algorithm provides at least the same level of difficulty estimate quality as a basic and quick empiric formula for the cost of a few more seconds for map generation. However, it is still too poor to be able to use it as a basic method.
Maybe, a better implementation should be invented? That'd include specific map peculiarities and will be able to improve difficulty estimate. Or another approach should be used?
eugeneloza
commented
8 years ago
It looks like Monte-Carlo algorithm provides at least the same level of difficulty estimate quality as a basic and quick empiric formula for the cost of a few more seconds for map generation. However, it is still too poor to be able to use it as a basic method. Maybe, a better implementation should be invented? That'd include specific map peculiarities and will be able to improve difficulty estimate. Or another approach should be used?