It's a similar situation to developing a new medicine.

If early test results look promising, you may feel you're costing lives by witholding the news and in fact that may well be the case.

But until the tests are complete you can't be sure of what you're doing.

Predictions in the training phase should be kept private until sufficient successes have accumulated to ensure confidence in the method.

Publication on this board really isn't public in the fullest sense; who pays attention to us, after all?

Roger