@sam,
I get what you mean about failed predictions.
But as humans and making predictions about anything, especially the future, we generally in the fail category.
Yes, there are one-offs. But I would not bet the house on them.
Realistically, I think there are too many variables, to many changes, and an algorithm/model or the computing power to make any AGW predictions with any real degree of accuracy does not exist at this time.
And if there was, we would call it God.