"One other suggestion: allow defending armies to pin attacking armies for a turn if they can hold them off for x number of turns, with the battle continuing into the next turn.  This would allow decent-sized defender stacks to hold off stacks of doom for a while until reinforcements come in.  The ROTK game for the Saturn did this.

On the next turn, the attacker has the option of retreating or continuing the attack.  (any reinforcements are added on)  You will likely need to save positions of forces, which could be memory problematic though. (probably less then I'm thinking though)"

I liked this in ROTK, & this kind of setup might allow for dungeon crawling & such too?, so what it takes 15-30 turns or more for your hero's' to return from questing in multi-player, the other players wouldn't have to wait any additional time at all, it would be totally up to you if you wished to decommission your heroes for that long

"It's my firm belief that acting (attacking) MUST defeat defense (being passive), otherwise the incentive to attack is lost."

I think of Troy with the great walls, couldn't be taken.... Untill someone came up with a clever way to get around the 'Legendary' defences.  I think it depends on how much effort is spent into either attacking or defending that makes the difference.  If I spend the whole game building up my '1 big city' then I would like the defence of that city to be awsome due to my limitless amount of resources being spent to make it so.  If you prefer the offense side of things then more power to ya, but defending my major city with moats, bilistas, catapults, burning oil, rows of walls guiding them into an arrow massacre, spears stabbing enemies threw holes in the walls as they enter narrow hallways in the outer walls, this my cup of tea.  To me the best offense is an impenetrable defence, I care not to steamroll early, but to defense up & grab allies.  & I thought the point of defence was to complicate "make things boring" for the offender. I assure you, you would have to spend allot more time & resources taking my city then you would have ever wanted to. 

To me, making a game where you attack then you win is boring, my cousin one time tried to take me in a game of Bandit Kings of Ancient China, he was playing Welcome Rain & I Balding Eagle, the game offered some defence options but not enough to suit me, so when he attacked me with 5 times the men & better magic I was forced to use other tactics, I conviced him to 'duel' my leader with his and we would see who the greatest was, so he attacked me (only bringin his leader up to mine) at which time I jumped him with my other two units resulting in victory before he could react (This sorta goes against me in my next paragraph).  If attacking is difficult/boring, then I suggest changing your tactics up.

I just think that if two armies colide, one much larger than the other, but the smaller (containing a Sov./or some other important units) has several 'possible' reinforcement groups in the nearby area due to the hit & runs they were doin on Altar's contructions, even though the reinforcements are less than one turn away from the battle & the battle has tactics being used to prolong, the reinforcements shouldn't get there untill all allied units are dead?  I understand this might not be possible, but I think it would add alot to the game.  And/or I would be ok with the aow:sm's pullin units next to a battle into the battle as well, I love going into battles helping my allies men that just got attacked.

Depending on how far down the line the implementation is at this point, you could have a mechanic that the city is "in rebellion" for a number of turns equal to some equation based on population, it produces nothing.  Instead, it takes a certain number of occupying units (based, again, on the population of the invaded city) just to maintain order.  If that city is attacked again (by a liberating force), the occupying force would have a certain number of units in the army removed from their defense as "police units", trying to keep order in the city.

An example:

I invade with 12 soldiers (4 units of 3) a city with a population of, say, 250 souls.  By mathematica arcanum, that equates to 7 turns of rebellion (a population 5 city would have 1 turn, a population 1000 city would have 20 turns; I could probably come up with a power law equation, but it's Sunday night, and I'm tired).  I lose 2 soldiers in the capture of this city; my occupying force has 10 soldiers in it.  On the first turn after the city changes hands, I need 7 soldiers to maintain order.  Only 3 are left to defend the city against counterattack.  On the next turn, 4 soldiers would be available.  Next turn, 5, until all 7 turns are done, the city has essentially flipped over to my side fully, and I need no "police force" left to defend.  I can use all 10 soldiers to defend, and at that point, my soldiers start to "heal" in that city (until then, no production, no "natural" healing in the city; spells should still be allowed to heal troops).

What happens if you need 20 soldiers to hold that 1000 population city, and you only have 15 left to provide a police force?  Well, NO units are available to defend against a counterattack for 6 turns (until the counter drops to 14, and one unit becomes available for defense).  In addition, there's a 5% chance per turn per deficit number (5 in this case, for a 25% chance on the first turn) of the city going into revolt, flipping allegience, and killing your remaining occupying force.  That chance drops, as all of this mechanic, to 20% on the second turn, then 15%, and so on down to turn 5 after the capture, where there's no chance of a revolt, but there's also no one to defend.

What this mechanic solves:

Invasion and defense are uncoupled from occupation and control; it solves FB's issue with "4 swordsmen walk into a city, and decide to take over.  They then set up a legislative body, a judicial branch, and a police force..."  In effect, they DO do that, but that means no one's left to defend the city.  It also solves the issue of the AI not having sufficient force to blunt an attack because they're too diffuse; it could probably mount an effective counterattack, once it knows where the enemy is (which it would, since it's FOW is lifted around that city for the time of the rebellion).

It provides incentive to bring enough occupying forces to insure against a revolt.  The wise commander may bring two armies; a strongly armored and armed attacking force to remove the defenders, and a weaker, more poorly armored (but cheaper) occupying force to hold the city against revolt once the invasion is complete.  Of course, that then leaves a weaker force behind to defend against counterattack...

It does not favor attacker or defender; the "time to siege" mechanic is tied up in maintaining order in the streets, and there's a significant holdup in time until a city truly becomes yours.  If it's liberated before the last of the rebels is removed from the city, it immediately becomes productive again for the original owner.

Of course, it's way too late in development to implement this kind of mechanic, but it would provide a lot of what's been discussed here.

The problem is that TBS's tend to degenerate into steam rolling on general principle, and steamrolling isn't fun.  Sirlin calls this the "slippery slope" problem.   This is why comeback mechanics are necessary- you just want them to be intelligent.

If a game is going to be determined by the first steamroll, you might as well make that steamroll very difficult.

Wouldn't this problem be solved to some extent by having more ways to slow-down the march of a stack of doom?

You need a more robust mechanic than this.  What's to stop me from splitting my SOD into 3 mini-SOD and moving faster and reintegrating right before attacking? Some sort of siege mechanic and/or allowing tactical combat to last multiple strategic turns would better combat the SOD problem (provided they were implemented well of course). 

Well, we have magic in the game for some reason:

Armageddon

Kills half of all units (enemy and friendly) in a tactical battle.

You've got a big stack? Too bad, now it's only half as big.

Chain lightning

Does increasing damage the more target it hits.

Big stack? Big damage to you.

Bound to death

Target gets cursed. Each time a unit leaves an army which has this cursed unit in it, the unit leaving dies. If the cursed target dies, another unit in the army gets cursed. Only one curse allowed per empire.

You've got a big stack? Nice, too bad you can't split it up ever again.

I can think of tons of spells that make big stacks very very costly to use if they're not necessary and would help against stacks of doom.

I wouldn't mind steamrolling in the late game. It makes sense and is actually something one would expect in war. Magic is of course the best defense to something as mundane as that. This is where we need that tornado spell. There should be a number of late game methods of preventing steamrolling. Defense technology and spell research would be the primary solution I would guess. Once again I need to see the beta to really have any idea. 

In the early game logistics should prevent stacks of doom. Each town should have militia and defense bonuses, requiring a sizable force to conquer it. So you really won't be able to take more than undefended outposts until you spend some time researching the art of war. 

My biggest worry is that this will make warfare the most efficient use of research resources like we saw in WoM. I am hoping that magic and civilization get some good defensive and guerrilla tactics. In fact, some sort of flying monkey spell would not go amiss. 

Let's add some buildings for the civ guy that gives high attrition damage to attackers (%-based to simulate starvation and diseases which the attacker has to deal with). Add a new category of super-combatants that do increasing damage the more enemies there are. Problem is, balancing those is much harder then adding some spells that are able to combat the worst of a stack of doom. (And also spells make more sense then some mobs that just happen to do more damage the more enemies there are. ^^ )

Also magic is fun.

Yes, but most likely we won't have such a system in FE. We won't get a siege system (as said by Derek in some other thread), battles lasting longer then one turn won't go in (would need too much of a system change), etc.

But even then, take the Total War games. They have most of what you've just described. End game consists of a few (2-3) super armies conquering the entire map with ease. All the systems there won't slow you down once you got a critical mass of units (and the economy to support them).

The problem is that attacking in one place with massed forces against a spread out defense line always gives you victory if the enemy is not able to counterattack or reinforce their defenses in time. In games it is usually made worse by the attacker suffering almost no casualties when attacking with overwhelming numbers. My examples with the magic spells were to solve one half of the problem, by making sure the attacker suffers heavy casualties to reduces his numbers. Giving defenders time to reinforce won't go in as long as we won't get a siege system which, as has been decided, won't be in FE. (The other solution being teleports of course, which imo cheapens the whole strategic part of the game.) City militias with big defenders bonus could make super stacks even more necessary, to overcome the defenders bonus though. :/

Making counterattacks possible is simple, make the AI do it's own superstack and attack. But then you have blob combat, stacks of doom in a race who can conquer more faster.

Since there is no sort of terrain ownership or larger zones of control, encirclements are not possible (and would be a bit silly in a fantasy game). Also no supply lines to sever.

Limiting the numbers of units in a battle doesn't work too well either, since if your first attack loses, you just send another wave. The enemy will always have less units then you have.

As far as we know, we won't have a morale system, so it's always a battle to the death.

So yeah, I see not much possibilities to make the stack of doom attack not work with current system without a total rewrite. There are some (more or less) gimmicky solutions like spells to make superstacks vulnerable. But then, I'm not that imaginative anyway, so I wonder how it will be in FE in the end.

--------------------------------------------------------

Personally I'd love a Morale/Organisation/Exhaustion system à la Hearts of Iron. You'd have morale on a per unit basis. Now any action you take depletes some of that and it replenishes each turn by a bit (but it takes a few turn to go to full again) low morale makes a unit weaker or even prevents it from fighting and moving. Morale increases faster in your own territory. Movement in enemy or barren territory, fighting, losing units in battle, taking damage, etc. would all lower morale. In a battle with a large versus a small army then, the small army could win by inflicting enough morale reducing status on the enemy. Be it using magic, killing the enemies front line troops (army wide morale penalty for every dead unit), etc. Attacking a city would cost the attacker a bit of morale each turn and so on. The defender has it easier since morale is higher in its own territory and it thus could act faster again after either a win or a loss compared to an enemy. So, perhaps the defender would lose most of its battles, but eventually they could inflict enough morale damage to the enemy that enough enemy units are unable to fight.

Or the enemy makes a forced march towards a city and captures it. Keeping that city under control while regaining lost morale from all the distance covered and the battle itself would force it to stay there for some time before it can work effectively again. And during that time it would be vulnerable to counterattacks.

But yeah, that's not very likely to get implemented anyway. ^^

Edit: As for the topic: When the AI is able to use superstacks too, I've got no problems with it in the game. Only when the AI refuses to use them and thus can easily be steamrolled it bothers me. So, go frog go, make the AI evil.

I have to agree with Kantok that the right subsystems could and should greatly diminish blobbing. Really any game that fails to prevent blobbing is not really a strategy game at all. However we are unlikely to get a complete system change at this point, but even just a few tweaks would make a huge difference. It all depends on how big of a deal Derek thinks blobbing is I guess. Magic can't be a dedicated blob counter. Blobbing has to have inherent penalties, ones that get players to think and plan.

Some fairly easy to implement or already kinda implemented things that would discourage blobbing:

-Siege times. Castles were principally used as a delaying tactic historically, so reinforcements could gather. Ties down blobs so they can be maneuvered around.

-Require freshly conquered cities to be garrisoned. That way if players conquer too quickly they risk leaving a whole bunch of isolated units behind to be picked off. Until the city is fully assimilated it's own siege delays and militia would be disabled.

-If monsters are a persistent threat to caravans and cities players will have a reason to protect even peaceful borders, and not completely empty all their cities to build a blob. 

-I believe max army sizes are already in. This prevents blobs from reaching critical mass. ie the point were they have so many guys shoved in such a small space that you can't even get a shot in.

-Some kind of retreat mechanic for normal units, so fast units can harass slower ones.

Some other more complicated ideas that are unlikely to be implemented:

-Any sort of logistics system. Maybe non-champion units in enemy territory deplete a supply stat. When it is empty it causes health degeneration. The nearest city would automatically send out a special caravan that would replenish it routinely. Would force players to protect a supply route, and allow blobs to be cut off and forced to retreat.

-I myself obviously love the idea of multi-turn battles, but they are probably more of a mod or expansion thing. Wouldn't be hard to directly implement though, just add a turn limit to tactical battles. The problem is it would require so many little stat and AI changes because it would add so much depth and strategy.

While I had already suggested flying monkeys, you bring up an interesting point: what about the world creatures? Making defense too strong will cause expansion into new territory too easy. 

Historical Note: In the Dark Ages of European history there was no ongoing resistance by conquered towns. Why - The penalty was death and the prospect of the entire town being burnt to the ground. 

Razing of towns should be an option, and one the AI may need to be trained to do, though razing should take more then 1 turn.

1 turn per level perhaps, with it gaining some gold as well.

As for world creatures, I can see militia being less effective vs them, maybe due to "fear of monsters" or something.  Militia aren't regular soldiers.  They should have a penalty vs anything they're not used to, be it darklings or dragons.