Calavente: I'm going to respond to your post, because it is a very well thought out thing you wrote and it deserves a response, but I think we are just having some off-topic debating fun, as Black_Imperator is putting the mechanic in. So I'm no longer disagreeing with you in trying to argue against it, rather I'm just always up for some off-topic debating fun.
technically this happen a lot in history.
the French Foreign Legion, which is a kind of Mercenary Unit, but always bought by France, was often filled with ex-military of fallen regimes.... SA (German non-SS WW2 soldiers), Stazi...etc
"Military of fallen regimes" would fit my suggestion of if a civ gets defeated, letting their leftover units become merc, not individual defeated units
and IIRC, often, defeated armies could become mercenaries when they were fighting in far-off countries... due to lack of possibility to communicate with their Motherland, and difficulty to go back.
If you can transform your "IIRC" into a "here are a couple of examples" I would be more convinced. I looked up several articles on historical mercenaries, and didn't find a single mention of any major mercenary group originating this way. I know of tons of such groups that turned into breakaway warlords themselves, but that is much different.
Also, remember defeated army does not equal defeated unit in civ mechanics. A defeated army in civ mechanics might be you lose 3 out of 4 units, with 1 wounded one retreating, maintaining some unit integrity and command structure.
Actually, the best way to model your point here would be to add a chance for any unit that is disbanded to turn into a mercenary. That would cover both the situation you describe (units too far away, home civ abandons them), and the many instances that nations had extreme trouble with their units becoming troublesome after choosing to disband after a long war (see, Rome) and going bandit or hiring out to whoever will pay them. So I would agree that a similar mechanic for disbanded units (in addition to the lack of payment and civ defeated instances mentioned earlier) would have a strong basis.
and all those "small defeated units fleeing everywhere" often amounted to 50%+ of the original army size.. and they still needed to eat.
The problem is the lack of unit integrity. This is a problem in the binary way civ 4 represents unit combat--if a unit survives combat at 3% life, then it is allowed to completely replenish itself just by resting, which assumes that all 97% of remaining unit manpower had fled and could be recovered, had light wounds, or is easily replaced by conscripts (very odd for higher level units). But if a unit is drive to 0%, then the model is the entirety of the manpower was severely wounded or driven off in an unrecoverable way. It's not actually a terrible system, if a bit extreme in its averaging, but it rests on the assumption that breaking the unit down to 0% really breaks it. So in that context, assuming no nation could ever rally the command structure of a unit back, even though the nation probably has a large force nearby with ready commanders and the unit has preexisting loyalties, but instead a new command structure could magically appear and convince the unit to disperse just far enough to escape the rest of the nation's military but then come back together into a coherent unit and then abandon their families and home and travel halfway across the world...
but most of all... it is fun mechanics ... that 2-5% of defeated units can appear on the merc pool
If set at 2-5% I would disagree. I play with aggressive AI (as I imagine most who don't like continuous peace do) on deity where the AI builds loads of units, and gets them killed constantly. I also think a well designed random merc pool will provide just as much fun--remember, this is an invisible mechanic that occurs, from the player's perspective the only thing that matters is the composition of the merc pool, not how it got there, so there really isn't anything fun or not fun about the mechanic, it's just a question of what builds the most interesting merc pool. But I'm just speculating--I really have no idea how this plays out, and whatever Black_Imperator does set the chances to, I'll just provide feedback based on how it in practice influences the merc pool, and it's completely possible my speculation is way off.
Anyway, I'm happy to keep debating back and forth as long as you like, it's an interesting discussion.
EDIT: But also feel free not to respond. The mechanic is in, and you have made great points--and not everyone has my love for endless off-topic historical debate.
