I have a feeling this blog post assumes a 360 cube (with no duplicates) as the standard, the way it talks about a lack of odds in a cube. I did some calculations (picking probabilities for a draw without replacement) for my 450 cube. If I include two copies of a card, there's ~64% chance that both show up, ~32% that one shows up, and ~4% that none show up.
I really like the concept of a core cube with occasionals and putting carefully chosen duplicates into the occasionals. That would change the odds, and provide some levers to customize the odds.
But why duplicate or even triplicate cards at random from your cube? Even set cubes have the sense only to repeat commons and uncommons. You could end up in a draft with an oppressively large number of mass removal cards or counterspells, leading to overpowered control decks. Or conversely you might be providing too much support for aggro with a plethora of cheap creatures and burn. Of course much of the time it will all balance out, in which case why bother?
I think it can be managed to create an interesting level of variability within an acceptable range. I like the idea of sometimes having a little extra burn, wrath, counterspell, ETB, fatties, ramp, etc. The drafters can feel that out in the draft to see what the best deck is instead of assuming that it's all the same as usual.
Now, one could debate whether it's best to have duplicates of spells or different spells with similar roles, because the latter would signal to the drafters that something extra is available in an archetype. Some cards don't have that option since there's no real substitute. But even if it's extra copies, I think it's really cool if there might be a second copy of a card like
Blood Artist or maybe a glue card like
Tireless Tracker, or some extra high quality removal, or a second
Goblin Guide.
)
