Our standard 5/8" countersheet has room for 176 little squares, and when I'm cobbling together my counter manifest for a game, I generally contrive to get as near to that number as I can without going over (unless the game is going to use a half-sheet, in which case the magic number is 88). Such was the case when I was putting together Battles on the Ice, the third game in our Shields & Swords II series. I began with a rough idea of the types of units each side would have, in what proportion, and the total number of units that each side would field in each of the two scenarios. That gave me a unit count that fell short of the 176 I needed. Not only that, but it often fell short within the color-blocks of 8 and 16 counters that would help protect against unsightly miscuts.
The general situation in both battles however pitted elite, reliable units against undisciplined, poorly-armed rabble, and it occurred to me that I could model this uncertain troop quality by randomizing the units. At the beginning of the battle, units of a certain type would go into a cup to form a pool, and the player would blindly draw as many as were called for, putting the rest back in the box. It solved my countermix problem and added some fuzziness to the system that could aid in replayability. I liked the effect and decided to use it in the next S&S II game.
That next game is The Great Heathen Army, and it's actually an ideal fit for this mechanism - to the point where I can't imagine the game without it. The various armies that fought in these battles were mostly made up of farmers and peasants, barely held together by a handful of experienced men of war. That means that the forces are made up almost entirely of Levies, with only a handful of Veterans to impose some kind of cohesion. The Infantry-type Unit, which splits the difference between Veterans and Levies, is completely absent in this game. Light Horse Units make some sporadic appearances, but there are no Heavy Horse to speak of. Mostly it's lots and lots of Levies and a few Veterans.
The ratio of Veterans to Levies is much more pronounced here. That puts an emphasis on Wing Integrity, the mechanism by which Levy Units are automatically eliminated when they get separated from Veteran Units and their fellow Levies. It's something that's been part of the series all along, but often doesn't get highlighted because there are usually more than enough Veterans to go around. With some Wings having as few as one or two Veterans to maintain order over a dozen or more Levies, Wing Integrity becomes much more important.
The same is true for the quality of those Levies. In Battles on the Ice, the random Levies provide a bit of fuzziness, but here, pulling three or four C-D Units instead of three or four B-C Units results in a very different experience. I pushed this mechanism a little further by supplying archer units for each Wing. Now, if you've played the series before, you know that archers aren't a distinct type of Unit. Instead, the ability to conduct Fire combat is conferred by the presence of a Fire Factor, which is grafted on an existing Unit type. Levies with Fire Factors, Infantry with Fire Factors, even Light Horse with Fire Factors. The Units act the same as any other belonging to their overall Unit types, except that they can participate in a Fire Phase.
That is, of course, if the player possesses a Fire Command Marker. In some scenarios he does; in some, he doesn't. An "archer" Levy Unit can still conduct combat - they have other weapons available to them - but as their Combat Classes are on the low end of the spectrum, they're not going to be as effective on the offense. Their defensive capabilities are also diminished; because the unit illustration for the archer Levies lack a shield, these units cannot benefit from Shield Wall. It's an effective, quick way to reflect the kinds of equipment issues that plagued feudal levies of the period; I think it gives the whole thing a feeling of granularity without having to actually get granular and come up with a bunch of special rules or new unit types.