draft better had a recent post on a 4-3-2-2 queue (MTGO). there were a tonne of great comments but nobody brought up an obvious oversight.
he picked a goblin assualt over a bone splinter (pack 2, pick 2). he did win the draft but as he admits he was very lucky. a 3/1 trample is not a high pick (it will probably table) whereas removal like bone splitters fits in perfectly with his unearth deck.
other comments that i thought were right on concerned mana fixing. trilands > fetch > obelixs.
i personally won’t even go for fetch or obelix very often. turn three obelix is just too slow. fetching a land on turn two is okay but on turn 3 or 4? those are really bad times. i typically find myself never sac’ing fetch lands because there are 7 and 8cc cards that will win games and if you’ve sac’d a land you essentially decrease your odds of hitting 7 or 8 pretty significantly. (i always run exactly 17 lands–unless i have a plethora of fixers in which case i’ll go to 16 but that’s REALLY rare).
i think the problem is that the tri-, fetch and obelixs lure too many players into going for four or five colors. two main and two splash is pretty doable if you have the right tri lands but i haven’t seen anyone win with a five color deck. best decks seem to be really focused within two colors (so part of one shard) and then touching on one or two of hte other shards. there are a plethora of great one-color+x-colorless spells and i think focusing on those as high picks is where it’s at. the alternative approach (five color) makes a player spend the first five turns fixing mana, while they get smashed in the face by a simple agro deck.