Description: Downhill search with varying tolerance. Early one is 0.01, latest is 0.00001 (increase by order of magnitude in subsequent experiment).
Results: Duration of optimisation seems just about identical and so are the results. Maybe the tolerance argument is meaningless in this context.