All-act times is too much RNG. Even vota is too much RNG. I get a 4 sec variation on Ghom tests with SNS alone and that's fighting the same target, same skills, same map, same position, etc on one target. Boss kill times are not true eps either because aside from a number of reasons, it doesn't represent a typical farming environment, but it's the closest we got to a constant measuring stick.
As for the arcane torrent player getting a good boss kill score and farming carefully with no deaths, let them. If a GC SNS wants to do the same, let them. What will happen is another player will try it and see that they are not getting the same result. The original poster will get challenged and I can see/record how they are (or not) pulling it off. Thus, "Try not to be the wizard that put in 0 deaths, then get killed twice when challenged."
If you're arguing on better accuracy with overall run times because most people don't do bosses, then someone will say, most gg players don't play solo and the test should be based on group skills and gameplay, and then another person will argue that although gg gear, mobility, and group synergy is important, it pales in comparison to someone who just plays an extra hour and real life time management is how we should rate wizards. Someone will always be unhappy with how we measure something as elusive as edps. We use bosses for now with the acknowledgement that it's not a perfect test.
I also want to measure these tests based on normal gameplay as much as possible. So I don't want players to be forced to rush it because they are on the clock. Just play normally. If your favorite build can't pass SB, then that's perfectly fine. I'm not going to give you a prize if you can beat SB.
Sands is a good test. It was my #2 or #3 pick, the other being Oasis. But I picked vota because of high density elites. So builds that do well with elites can shine there, much like how builds that do well with trash can shine in FoM. If more people like sands over vota, I can change.