I don't have any armies based on mass Ax (I note the forthcoming possibility for enormous numbers of Irr Ax(I) and Ps(O) in the Palaiolgians, but taking them means losing all the other stuff I have always regarded as obligatory, such as the Serbians). However, my Epirote Byzantines and my Later Mycenaeans always have a lot of Ps - the Trojans can have up to 50 Ps(I), (O) and (S), and they all have a role to play. En masse, even Ps(I) can hold against opponents in DGo, and are always a threat out wide or wherever the enemy lines can be stretched to form a gap. I find them very effective in holding back massed foot until I am ready for them, and it can be quite demoralising for the enemy player who realises that he has spent 10-12 Ps, and yet he hasn't really scratched the command (but don't, as I did in a practice game, then lose the general...) If nothing else, they can be a big reserve line behind the other fighting troops, as a cheap element to throw in to prevent the enemy turning on the flank of an internal gap.
Ps(O) will support other Ps, so a second rank behind Ps(S), or even Ps(I), gives you a really good chance against enemy Ps, LH(I/F) or as a delaying force against LH(O/S).
I like Ps(S) a lot - they can be surprisingly deadly and surprisingly resilient. However, they are brittle, as 2 is a low starting point, and in their own bound (when the don't get the (S) defensively) a low combat-dice roll will see them as dead as a Ps(O) would be.
For me, the art with massed Ps is to remember that you have a lot of numbers on your side, and try to use that to get outside small enemy groups. 4 enemy Cv, for example, can only be 4 wide, and cost as much as 14 Ps(O). It only takes 1 or 2 Ps(O) to get behind the Cv and the game gets very interesting indeed. But over-commit with the Ps and suddenly those 4 Cv can be killing 2ME per bound... It's often a situation where you toy and play with the enemy - threat and counter, dog and pounce - until one side or the other decides that the situation has to be brought to a head.
Tim Child