Comment by hobofan

But "rerunning reproducability" is mostly a neccessary requirement for independent reproducability. If you can't even run the original calculations against the original data again how can you be sure that you are not comparing apples to oranges?