[Tom]

PvP issue

Monk I think you're making the issue more complicated than it needs to be in relation to MLG Columbus. (If this gets too esoteric, we're going to leave everyone else behind, so I'll try and make it as simple as possible.)

In the brackets, you appear to have counted total PvPs, rather than PvP's per round. Obviously, some players advance to further rounds. Of the 56 Protoss players, many played multiple PvPs (that is, the same Protoss player may go on to play several PvPs - and this is actually what occurred in many cases). So by comparing the 56 Protoss players with 32 PvPs, you are comparing apples with oranges. You actually need to calculate PvPs as a proportion of all match ups if you're going to look at it this way.

Second, the reason why there ended up being a whole lot of PvPs in the losers bracket is because so many Protoss players got knocked down there.

Third, even accounting for the "more protoss players, hence more opportunity for mirror match ups" issue, it is still the case that Protoss went from being by far the most overrepresented race in qualifiers, to by far the most underrepresented race amongst those who qualified. There is substance to your point, but I tend to think it is not as important as you make out - you would expect slightly more attrition due to a higher number of PvP mirror match ups, but this cannot by itself explain the fact that the proportion of Protoss players halved whilst the proportion of Zergs and Terrans both increased by around 25%.

Other comments

In relation to your other comment that (a) basing this discussion off tournaments isn't a solid way to do it and (b) that MLG Columbus players arent "high level"):

(a) I also have looked at a wide range of GMs League data in my OP
(b) MLG Columbus is mostly NA Grandmasters with some Korean pro gamers.