Luke, fair enough. You do seem to be able to take suggestions.
At some point when your methods are improved, if you want your predictions to be evaluated statistically (Roger is the only one who does this) you have to issue your predictions in a way that can be evaluated.
Chris