Take one look at the Dreamhack seedings and you know that whoever did those must've been either drunk or severely misinformed of the state of Dota2 right now.
For those of you who might not know, here is how you properly seed groupstages: First, you separate all available team in skill tiers: 4 top seeds, 4 high seeds, 4 mid seeds, 4 low seeds. Then you proceed to draw one random team from every tier and throw them into a group. 2GD actually did this live on stream for the Quake Live tournament, so we know for certain that this is a method of seeding Dreamhack uses.
So let's divide these teams, based on recent performance, and compare them to the current seedings. The tier distribution is somewhat arguable, but I'm using the GGnet ranking as well as recent results (e.g. win rate of last 10 games) as a basis.
Top tier: Na'vi - CLG - EG - mTw
High tier: Mouz - Darer - Quantic - TTB (maybe AL)
Mid tier: Keita - Infused - VP - AL (maybe TTB)
Low tier: All BYOC teams
Especially VP, AL and TTB's positions are somewhat arguable, but GGnet tells us that TTB is doing amazing recently while AL has lost quite a lot, hence how I seed them.
Now let's look at the group divisions Dreamhack made:
Group 1: High - High - High - Mid
Group 2: Mid - Mid - High - Mid
Group 3: Top - Top - Low - Low
Group 4: Top - Top - Low - Low
Arguably, the seeding might turn out so that both group 1 and 2 have 2 High and 2 Mid teams.
This is not an even distribution. This is not good seeding. This right here is 2 very tight-knit groups where literally anything can happen, and 2 groups that are completely decided from the getgo. It means that all the games played in group 3 and 4 are meaningless because we already know which 4 teams will advance there. If you really care about seeing into the brackets, then Na'vi vs CLG and EG vs mTw might be interesting, but those results aren't significant at all.