[alphager]

We are as close as you can get, but we estimate that we're about 3.5% away from true optimum on the hardest problem we try to optimize.