I can tell everyone now as long as the design aspects are right and high end units are expensive then people will almost always mix in weaker units with the heavy hitters. Those weaker units will be in the front lines keeping the more expensive power house units from taking damage. Here's a good example, Empire: Total War.

In Empire: Total War you have a vast variety of units you can command. Some of these are better then others and of course some of them are much more costly then others. It's completely necessary in that game to use a varied army structure. My front line of units consist of Line Infantry. These front line units are of a middle expensive range and always end up taking the most casualties in a battle. They take the brunt of the attack while my more expensive Dragoons circle around to attack from the rear or to flank my enemy. If I fielded a whole army of Dragoons without standard foot soldiers to back them up they'd get wiped out. Their unit count for a full unit isn't quite as high as that of my Line Infantry so I wouldn't have as many fighting men in a army comprised of just Dragoons. Also the Dragoons alone would not have the defensive capacity to stand up to any other unit they may encounter on the battlefield, such as units designed to counter calvary.

This demonstrates that with proper unit design concepts, cannon fodder line infantry are still needed even though there are much better units available to use. That's not even mentioning the cost of upkeep to keep that many Dragoons up and fighting. They have a considerably higher upkeep cost than line infantry.

You missed my point again. In most games, you are better off training only the best troops you can possibly train (within the different classes of units, of course), even if you could build, say, 2x or 3x as many troops of a little lower quality.

All in all, like has been said, I think most of us agree in general times on what should happen. We should be encouraged to train units of different types, and units of different caliber. Personally I'm against anything that forces us into that (in some rare occasions I might actually be better off fielding a homogenous army, monolithic or not!).

I'm also against the 'exceptional' people idea, because it seems pretty complex to me without really adding much to the game. If I train two regular Archer units and one is an 'exceptional' person and one is a 'lazy, worthless' person, I'm going to be confused as to why one is so much better than the other. It just doesn't seem worthwhile to me, and IMO as long as training costs and times, upkeep costs and combat effectiveness are well-balanced, it's purpose is highly diminished. If those things are not accomplished (which would be really disappointing) then perhaps it'd be time to consider it again...

Also, somewhat off-topic. I was just thinking about being able to reabsorb military units into your population while they are not out marching or patrolling somewhere, or defending some mountain pass or what-have-you, and doing so without losing their training & equipment. It's been brought up a few times and I think I just though of a pretty painless way for it to be implemented... Have it so that armies fortified inside your cities contribute a percentage of their population cost towards the city's population (this would have to be accounted for separately, though, as you wouldn't be able use that population for things like training new soldiers). Perhaps to prevent abuse, you would have to click some sort of 'demobilize' button on the army or on individual squads; to remobilize them would take some turns - as such you would not want to do this if you aren't pretty confident in the safety of the region.

Hopefully we've got that straightened out, and can stay on the point you argued against...

Other ways to limit the 'best' units (time to train, cost, special resources, etc.) tend to lose their effect as the game wears on, so something else may be required should this be a design goal.

This is not part of the 'rock-paper-scissors' balance (where pikes are always useful vs cavalry even in the end-game, etc.), it's something else -- another layer to the game.  It would allow very special unit types (the ancient red dragon, or Navy Seals, or the like) without the game devolving into those being the only units one ends up with.

And your being against this idea is fine by me 

I was responding to multiple people, and not just you. You are not the only person in this thread who is making comments that I think are worth responding to...

Ok, well that's an improvement over the initial impression I got. But it brings up another issue... With fully customizable units, what determines which unit templates require 'exceptional' population and which don't? Wouldn't this also lead to the ability to abuse it by creating units that fall just below the thresholds, resulting in nearly equivalent units that can be produced in much greater numbers than slightly superior ones? I suppose you could have 'population type' as something that goes into unit design, but that wouldn't prevent me from decking out an average person with incredible equipment and thus being able to mass-produce damn good units anyway... They might fall short of 'exceptional' people with the same equipment, but nonetheless the forced variety among troop quality would still be greatly diminished.

Also... ancient red dragon?! You intend to turn regular people into an Ancient Red Dragon?! I don't understand. We will not be training dragons (and if we were I doubt they would be made out of people...). If you are playing a human faction, you will only train human soldiers and likewise for Fallen. Magical creatures will be recruitable via unannounced means, but they will not be trainable. 

Edit: Forgot to reply to Demiansky's post!

Food consumption and overcrowding? Personally I'd still like to see troops require food upkeep. And demobilized troops wouldn't count towards city population for things like city size, etc.

Come now, use your imagination.  When downtrodden villagers sacrifice virgins to dragons you don't honestly think that they eat all of the virgins.  Surely, there must be some halfdragons running around in the general population that can enter a barracks and, upon leaving, promptly transform into a 2,000 year old... okay, yeah, I see your point.

okay, now my elder-dragon epic lair is going to be filled with not only a large Lizardman tribe, but also the religious guides to the lizard-tribe will be a sect of half-dragons that accepts their heritage and has inhereted several draconic attributes, like wings, hardened skin, and/or dragon's breath. Many dragons in fantasy can transform from a Dragon form to a human form at will, while Half-dragons are stuck with whatever the genetic recombination blesses them with.

Yes, I said that tends to happen. But I believe I also said (in different words) that it doesn't have to, and that crazy new mechanics aren't necessary to achieve that. Scarcity of resources, for one, does not make it into many games at all. Training time and costs, and upkeep, for the strongest units tends to be disproportionately low compared to the weaker units, too. Additionally, most games have really low caps on army sizes (like MoM, AoW - armies can only have 8 (?) units in them?). The above combination is the perfect recipe for Monolithic Armies being the superior strategy. Removing any of the above helps, removing all of the above I think would be the cure. Adding in combat mechanics that take into account things like flanking and morale and you're even better off. Just look at Total War - it removes only some of the above causes, and it takes a very very long time before monolithic armies become a problem at all (I have completed it many times before it became a problem at all).

Wow, that sounds terrible to me. No regular, trainable units should ever be able to compete with a hero, let alone any of the crazy powerful magical creatures in the game. I vote a very firm "no" to that idea...

I also do not like arbitrary limits on things. 5% - why 5%? Why not 7%? 3%? 42%? Why not make the feature work as intended in the first place so that you don't need some arbitrary hard cap? If such a thing is needed, then it means the feature was not implemented well to begin with.

Eh, this seems different to me. In Medieval:TW the number of elite units you can train per turn is limited, not the total number you can have at any given time. Same in HoMM (in this one the numbers can even be justified as racial reproduction rates...). They do not say "only 5% of your army can consist of level 7 creatures!" While anything can be chalked up to game balance, something like how many elite troops you can have being determined by some %age cap just feels wrong. It might work ok in terms of gameplay (but it can be made to work just as well or better with a little more creativity), but it'd ruin any feel of immersion. "Oh no! I've run out of exceptional people in my nation, what am I to do!" It doesn't seem fun to me, and it doesn't seem necessary - in this particular case there seem like much better ways of having limits arise organically.

If I can have at most 5% of my population trained as 'exceptional' troops, why would I ever not train 5% of my population as exception troops (not including the beginning of the game, where this problem tends not to be relevant)?  The only reason I can think of for why not is that perhaps they cost so much time or resources to train that I might be better off training an even smaller amount. But if that's the case, then a 5% limit is not needed, because the cost-benefit of said troops already provides incentive not to always train the maximum allowed, and has the advantage of not preventing you from training more as strategy might dictate. It allows additional creativity when coming up with a strategy.

Perhaps you aim to be a major economic force with 10% of your population trained as truly elite soldiers, and very little else in the way of military might. The economic force would be necessary for this strategy in order to support such a large number of exceptional units. On the other hand I might want an army consisting of 40% of my total population trained as mediocre soldiers. Why not? This set up might be much easier to maintain economically, even considering how much more of your population it ties up into troops. Either of these should be viable as long as the player is willing to make the necessary concessions and sacrifices, and of course each would have its pros and cons, and thus you have to consider which is more appropriate in a given situation. Saying "You cannot have more than 5% of these robs the player of interesting, meaningful choices, and breaks immersion.

double post

It's the same idea applied to a different part of the game.

The thing is that normally you can train up to 100% of your military in elites and usually that's what players make, because it's the best choice. Reducing the percentage doesn't disallow you to have a pure elite army, it just makes it smaller compared to the armies other people could have. It may continue to be a viable alternative in some situation.

And btw, if hiring units doesn't break inmersion for you (a totally artificial and alien mechanic), it's surprising hard limiting the number of elites (something that happened in the real world) is going to break any inmersion at all. What is inmersion breaking is watching an empire only raising hordes of paladins armed to the teeth, but not squires, and all the support that should be associated with them.

Yep this is exactly the issue that needs to be addressed. If i want to train 40% of my pop as mediocre soldiers what are the actual benefits and tradeoffs? Is there any benefit at all? If there are no benefits, then of course everyone will be going for the 10% elite troops, a.k.a. you end up with monolithic armies.

Remember when i convert 40% of my pop to soldiers i am taking a severe hit in terms of my economic development compared to someone that only converts 10% pop to elite troops.  Meaning there has to be some sort of incentive for me to go the 40% route to offset the fact that i am taking an economic hit. What would that incentive be? Just lower maintainence cost?

Next, when the elite army and mediocre horde army meet, will they be evenly matched? Or will the elites win all the time? Or vice versa? Again if one side always win then of course people will again always gravitate towards either all elites or horde only mediocre troops.

So really, what specific aspect/rule can make both playstyles unique yet equally viable?

I'll point out something that's not been talked about much: Counters.

One way of neutering elite units is to have counters for them, that force you to use meatshields.

For instance, if wizards can cast fireballs that ignore armors, are you going to send in a troop of expensive armored soldiers or a mix of unarmored meatshields to take out the fireballs with a few armored soldiers to deal with the enemy elite?

Yes, it's rock paper scissors again, except not along the usual axis (infantry/cavalry/ranged), but along an elite/non-elite axis.

So armor negating spells make expensive units less useful, buffs (natural armor which doesn't stack with worn armor, berserk, boosted morale or skills) make cheaper/weaker units more useful, can change drazstically the cost/value ratio of units.