Ah I'm not testing yet, just theorymashing.
Right... lets look at Muster the Guard, and acceptable returns.
We know that losing the command-snipe is a big thing for any warlord, so I think what we want is a return of 4R on use.
One number we can't do anything about is that Muster the Guard is only going to be in opening hand (with mulligan rule) 65% of the time, though also we only need to worry about double-Muster collision 6% of the time.
If we want a 4R saving, we need to deploy at least 4 units. Essentially that means a hand that includes at least 4 units of 3R cost or less (discounted to 2R or less, obviously).
Sample size is 7/49, as we already know we have a Muster in hand as a prerequisite.
37 AM units of cost 3R or less will give us a 95% chance of hit there. A more conventional 32 units leaves us at only 82%.
As we're multiplying this by the chance of getting Muster in hand, 82% is too low. Basically something like a 52% chance of getting a worthwhile play out of Muster.
Off the 37 unit deck, its still not much better: a 62% chance of getting a worthwhile play out of Muster.
The problem here is the chance of getting Muster up front can't be pushed past 65%, and the nature of the game means that any Muster NOT drawn on turn 1 is likely to just be a shield card. Sure, there might be a later game play, but we're unlikely to ever have a chance to make a 4R saving in the later game.
Still, I think if we go for a deck with 37 AM units of cost 3 or lower, we're looking at a deck that over half the time will essentially be deploying 12R worth of units on turn 1.
Thats a pretty strong advantage!
Winding back to reality, the problem with finding 37 AM units of cost 3 or lower is that there's only 13 choices at present, one of them is Unique, two of them cost 0, and one of them (Augur) is useless unless we have 10 Supports as well. Happily, the sig squad accounts for 4 already, so we basically need 33 picks, and thats not doable without taking Augur, taking a 0 cost, or taking more than 1 of the unique.
Good picks include:
3x Cadian Mortar Squad (Core Set)
2x Captain Markis (Core Set)
3x Interrogator Acolyte (The Howl of Blackmane)
3x Iron Guard Recruits (The Scourge)
3x Ratling Deadeye (Core Set)
3x Sanctioned Psyker (Core Set)
3x Stalwart Ogryn (Core Set)
3x Steel Legion Chimera (Zogwort's Curse)
3x Tallarn Raiders (Zogwort's Curse)
We could splash in some 4-cost, or splash in some allies, but the maths starts going wrong if we do that.
Because we're seven picks short, I'd say the Muster deck isn't quite good to go yet.
However, as soon as we see 2 more AM cards that cost 3 or less and would fit this deck, I'd be willing to give this deck a punt.
The remaining questions in that circumstance, of course, would be:
What are the 9 non-unit non-signature picks for that deck?
How do we play the deck on those 1 in 3 games we can't open Muster?
Can we play a 32-AM unit deck that only gets a strong Muster 50% of the time instead, and have a deck that performs better in the 50% off-Muster games?
Do we really need a 4R saving for Muster to be worth playing? (I say yes, others may disagree, and if we're happy with a 3R saving, the deck just got a lot easier to build)
These are tough questions, but right now I'm just laying some theoretical groundwork...
Right now, I believe the Muster deck will still be outperformed by more conventional 2-faction decks that go for a less single-minded and toolboxy approach, but I do like thinking about it!