To the judges, with respect to the evaluation of the quantitative winner:
Are you going to run the submitted learner once, and let the learner solve the tasks 1-4 one time?
Or do you think of running it e.g. 3 or 10 times, and take the average number of needed steps?
(I know that the rules are very clear, and don't mention running the code several times, but rather only once. Still wanted to know whether you are thinking about that issue)