Lower bounding the time limit of a problem is a potentially expensive operation (as the time limit is unknown, we're forced to let programs run for a long time). Thus, we typically want the set of programs used to lower bound the time limit to be as small as possible (but not too small, to cause a too low time limit, of course).
In legacy, we only used accepted submissions to lower bound the time limit. I think this should be kept as the default for the newer version. Given that there is more configuration opportunities, a problem setter who for some reason want their RTE submission to set the time limit can now explicitly add that.
Lower bounding the time limit of a problem is a potentially expensive operation (as the time limit is unknown, we're forced to let programs run for a long time). Thus, we typically want the set of programs used to lower bound the time limit to be as small as possible (but not too small, to cause a too low time limit, of course).
In legacy, we only used
acceptedsubmissions to lower bound the time limit. I think this should be kept as the default for the newer version. Given that there is more configuration opportunities, a problem setter who for some reason want their RTE submission to set the time limit can now explicitly add that.