Reports could mention median number of evaluations instead of mean

[Martin Janiczek]

I'm not very sure about this (not being all that good in statistics), but it seems to me like the mean number of evaluations might not give as correct picture of what to expect compared to the median. But, this is probably up for discussion.

Also related to #35 which might give yet more insight: the time-per-shrink is not same across challenges.

[Johannes Link]

Currently, there's no strict format for the report. If you just add it to your reports, someone else might pick the idea up.