I tested the list yesterday, playing three matches against big-mana grixis control, PO UBW control and PO combo with time vault. I won all six games, but that has more to do with my opponent drawing poorly than anything else. A few observations:
I only really comboed out twice. Once, food chain was merely a mana accelerator. In 3 games I just goblined the opponent to death. This is fine, of course: The aggro-plan in the deck is decent; especially against these three decks, that dont block lackey consistently.
The deck had a powerful plan 4/6 games. In the last 2, I just played little dudes, disrupted a bit and luck-sacked my wins. When the deck is doing something powerful, its because of one of these things happening:
-I played turn 1 chalice for one twice, off gemstone caverns and a mox.
-I played turn 1 lackey, which was not answered (4 caverns of souls being key).
-I played turn zero gemstone caverns.
Serum powder performed really well, I used it in 4 out of 6 games, either drawing it in the first opener or mulliganing into it. I never mulliganed below 6, and serum powder helps to not do this. I often had access to squee from exile, which helped me combo off twice. Once, I just played him from exile as a dude. Just having him there keeps the opponent nervous.
These are too few games to make many changes. Ill do this:
-1 siege-gang commander, +1 skirk prospector. In two games I explicitly wanted to recruiter for prospector, because I was stuck on two mana and unable to play food chain. This also makes sharpshooter an instant kill, so I can drop siege-gang, which is mostly an irrelevant card.
-1 matron, +1 recruiter. Matron was clunky and recruiter was less of a bad/risky card than I remembered. I recruitered for 2-5 goblins a bunch of times, typically tin-street hooligans and ringleaders. Not once did I goblin-screw myself.
Trinisphere might be bad, and both tarfire and maindeck earwig squad might be irrelevant, but I dont know yet.