[mrbonner]

Now, this would be a very high bias, low variance ML problem to tackle :-)