Dies to Removal

What do you guys think about it? How could it be optimized?

[deck=Todd Anderson's Dark Rabble]Creatures (12)
4 Ashcloud Phoenix
4 Goblin Rabblemaster
4 Brain Maggot

Planeswalkers (3)
3 Sarkhan, the Dragonspeaker

Lands (25)
3 Mountain
8 Swamp
4 Bloodfell Caves
4 Bloodstained Mire
4 Temple of Malice
2 Urborg, Tomb of Yawgmoth

Spells (20)
3 Bile Blight
4 Hero's Downfall
4 Lightning Strike
1 Despise
4 Sign in Blood
4 Thoughtseize

Sideboard
4 Bloodsoaked Champion
1 Magma Spray
2 Pharika's Cure
1 Silence the Believers
2 Chandra, Pyromaster
2 Liliana Vess
3 Anger of the Gods[/deck]

Why Phoenix over Stormbreath Dragon? They Both deal 4 dmg no sooner than in 5th turn. I understand that Phoenix is sort of CA but does this deck really want to play CA-game?
Hmm on second thought I get it that Stormbreath is in the same slot as Sarkhan. But still not sure if i'm convinced.

It reminds of old Machine Head from Invasion standard. Alittle bit at least.

I think he's going for the archetypal grindy Rakdos deck, so the phoenix is slightly better fit (you cast it, they kill it, they have to kill it again or you flip it later on).

Deck gameplan appears to be to kill/hand hate your opponent's threats and win on pseudo-card advantage from phoenix / rabble or real card advantage from sarkhan/sign in blood.

Feels like it might need more lifegain though, as you will be taking damage from the fetches, sign in bloods and thoughtseizes, so could get run over by hyper aggro or if you stumble on removal/mana.

The Caves has gain 1, and Pharika's Cure is in SB for aggro matchups. Given how creature-heavy everything is at the moment, perhaps flop the numbers on Despise/Thoughtseize to reduce your self-induced hits?