Welecome to the Wolverton group's metal detection challenge. The point of this game is to test whether the intuition of scientists is better than a model produced using machine learning.
You will be competing against a model trained against 3000 randomly-selected compounds from the ICSD. See above for the rules governing which compounds were used. To create this model, 81 attributes were calculated for each compound and used to create decision rules with the rotation forest algorithm. These attributes include things like the average electronegativity of each element and the maximum difference between thier melting temperatures.
To make the statistics about this more interesting, please let us know a little bit about you. Also, feel free to provide a user id if you want us to keep track of your scores.