Mostly the fact I am relentlessly unlucky. No matter how precisely I engineer a deck, I will consistently get ten turns in a row of nothing but land while my opponent brings out five 12/12 fliers and cast about fifteen buffs on all of them. Even if I run a card drawing deck, even if I play a deck designed by one of the pro tourney guys who won a million dollars, I still never get what I need when I need it. This isn't strictly speaking a fault of the game, but it also makes it not an enjoyable experience because I'm constantly struggling against game mechanics.

The flavor is amazing, the art is amazing, but there are too many moving parts to track and I am far too unlucky to make the cost worth the aggravation.

Needs more bears, IMO.

Fair enough. Most tribal decks have more of the appropriate creature than anything else, so:

Land:
20 Forests
4 Slippery Karst

Creatures DA BEARS:
4 Ashcoat Bears
4 Balduvian Bears
4 Grizzly Bears
4 Runeclaw Bear
4 Striped Bears
4 Werebear

Helper Cards:
1 Sylvan Library
1 Rowen
1 Primitive Etchings
1 Temple Bell
2 Words of Wilding
2 Bearscape
4 Door of Destinies

That's 24 bears, plus 2 copies of enchantments that make more bears out of your graveyard or extra draw, 4 card drawin' cards, and 4 artifacts that will make the bears get to maulingly large proportions.

I think unlucky is a feeling shared by most players Aaron. Mana screw is no fun either and I wish there were a fix for it.
MTG 2012 on console - ever notice the AI almost never gets mana screwed? On that platform I think the system should be designed so that you never miss your first 3 drops. After that, you're against the luck of the draw.

I have been accused by my current playgroup of having "special mojo" that makes their decks have more frequent mana problems, while mine putter along just fine. I think it's just an illusion of selection bias, we all remember the times we lost because we drew land 5 or 6 turns in a row while a couple 2/2's kicked us in the shin over and over. In reality, the Mana Screw in either direction doesn't happen all that often, and when it does in a casual game, the best response is to shuffle and ask if your opponent would mind trying that again, because you didn't get a fair chance with your draw.

Oh, absolutely. We even do that on XBL.

I don't play MtG but if I did I would want 25 bears.

This.

Caller of the claw would be a good sideboard card, for when opponents have board-sweeping cards.

The problem with Grizzly fate is it's really only worthwhile if you have Threshold. Bearscape burns cards out of your graveyard, which can take away Threshold.

Hibernation's End would do nothing the first time you paid it (no 1 casting cost Bears), get you a bear on the 2nd payment, then if you REALLY had mana to burn you could pay it with no effect on turn 3 so on turn 4 you can get a Striped Bear... Hrm, I don't think that works in this deck, despite the spell's name being a bear reference.

If we really want the Bear count to be 25, we can take out Rowen for Spectral Bears, that's a 3/3 bear for 1G, with the drawback that it takes 2 turns to untap if it attacks someone not playing Black.