Depends entirely on the time frame, and the unit. How remote is this battle? If the fighting is a day's march from a castle, and you're attacked by a dragon, anything that can't hide is going to die. The problem with a short hardcoded limit is that it presupposes you can just waltz to safety. Not always the case. You're backed up on a peninsula, where do you withdraw to? The ocean? Do you have time to take your armor off? Then you drown. See the problem? No retreats or withdraws, winner take all.. that's how melee combat generally works out. It's a video game, but I see no reason to make it so unrealistically easy. That goes double for just cutting the battle off and giving it to the biggest army. Why even bother with tactical battles if I've only got a few minutes to use those tactics?

Maybe you could retreat if you're much faster than your enemy.. maybe there could be a spell, like a smokescreen.. but apart from that? You're not going to just walk away. Why would they let you?

I agree at least partially about the fliers, but afaik there won't be many (or any) at release anyhow. It'd only be a problem if the creature could lob stuff at you, cause otherwise it's got to get close to hurt you, and then you stab it. I'd like to see lone flying units be able to harass any force it can't kill. Steal supplies or scare guys or something.

For the things you want input, here is my 2 cents.

- Make the turns allowed before morale drops based on current armies in play. It its just your sov vs 1 spider, this should end within 5 turns easily. If its 20 vs 100 it may take a while. I think battles from "The king bounty" or the "Heroes of might and magic" and almost right ont he money of being not too long, but not too short.

The other part of a tactical battle could be, while the battle is going on, the world is still moving (its not a direct freeze in time and places are still moving and growing)

As far as the maps, I am leaning towards the random generated ones simply because you said they would reflect where you were at the time. Would be silly to be on a mountain and the battle you have takes place out in a grassy feild somewhere. If it was possible to have say 100+ maps for each TERRAIN then I would easily go for the pre-made ones.

Thresholds: I don't really like this idea because auto-resolve and tactical combat in every game have different mathematics. And I doubt elemental will be any different because even if it resolves it using a tactical combat behind the scenes so it's the same "math" there is the always the fact that no matter how great the AI chances are they do combat much worst then the player.

Also what about if you set the limit to 10 and you running around with 2 high level heroes and encounter a group of say 8 enemy units. You might actually want to fight that battle so see your heroes kick some @$$. But your can't cause of the threshold. Also again the different mathematics what if had you be able to control them it would of been an easy victory but because of the auto-resolves math you lost by a lot? Or what if the AI waste spells in combat like it often seem to do in MoM. Now your guy is down mana you wanted to save for something else.

The different mathematics is why I think auto-resolve vs tactical should be something chosen at the start of the game and go through to all battles. Though a third option that lets the player simply choose at the start of each battle could be fine to as a way to simply speed things along, though I know some will abuse it given the different resolve techniques. Like I recall one game where if you auto resolved a siege attack you could take the city even though if you tried in normal tactical view you could not because you had no siege units to breach the walls and no range units to hit enemies inside.

The thing with "choice" option though is it works fine for SP. But in MP the problem becomes the player will always choose the method that favors them most. And then who gets the say on the choice, the attacker, the defender, a vote in which case who is tie breaker, and so on? Maybe it's tactical unless they both vote auto resolve, or you could go the other way to in that it's auto resolve unless they both want tactical.

Personally I'm in favor of the pick at start and applies to all combat. Or the choose for each combat in which case it should have a default to in case of ties during MP matches. Thus there are 4 options total.

• Tactical
• Auto-Resolve
• Choose before tie defaults to Tactical
• Choose before tie defaults to Auto-Resolve

The defaults to is so if one player picks tactical and the other picks auto-resolve they know ahead of time which way the combat goes. Thus players who prefer mostly tactical can choose to skip a small battle if they are both inclined to do so. Or player who prefer a quick combat focus can have auto unless they both decide there is some epic battle they both really wanna play out between their huge armies late game. And of course the AI player doesn't get a vote when it's Player vs AI and player chooses.

Controlling the length of a tactical battle: I think there is actually a few ways you can handle this. It kinda ties into the way other games such have Civ 4 effect the game length. And that is effect how long it takes to do stuff in the game. Obviously this will have some potential balance issues as in an Epic game of Civ 4 units mattered a lot more since it took them longer to build while their movement speed and combat strength was the same meaning a unit could explore more and had the potential to do more damage given healing the a unit in a Normal speed game.

When it comes to tactical combat I think there is a few things that could be done.

1. Map Sizes - With larger tactical maps the players would have more room to move around in thus giving them more time to flank and position.
2. Adjustable Damage - To make the battle last longer all units do less damage. Where as to make it go faster units deal more damage.
3. Adjustable Combat speed - For epic games units have few actions each turn while fast paced ones they have more actions.
4. Combat Timer - More for Multiplayer but limit how long a player can have on their turn so they don't stall. 
5. Adjustable Health - The other side of Adjustable Damage which might be harder to implement as some units only do 1-2 damage so so they would be do fractions. A health modifier in tactical combat could do it the other way then round back down on strategic map.
6. Adjustable Terrain effects - Tiles that reduce movement have a higher movement cost on long game setting making getting around that much slower.
7. Adjustable Attrition Rate - How fast Moral/Supplies drops from being in combat.

These options will of course end up effecting balance somewhat. But any time you change around stuff to effect the pacing it is going to effect balance some where along the line.

But more then simply making combat longer or shorter based on what the player wants I think it also needs to remain interesting. Especially for longer combat option setups. I mean you mean two hour battle but I don't think anyone wants a battle where both sides just sit there hitting attack over and over. Terrain bonus/penalties, flanking options, special unit abilites, and so on I think could really help to keep combat interesting in long combat games.

Good prisoner ideas.

And yes, I believe that in order to sound a successful retreat, organizing your units for retreat is key ... however I don't think the "retreating zone" needs to be a hard cap .. but rather that faster units can retreat from farther away, etc.

And also of course bloodless retreats if no combat has been engaged for X turns, and enemy units are X tiles distant from each other.

That quote is over a year old. Back then battles were supposed to be in real-time and feature thousands of soldiers. A lot has changed since then.

Both - have some pre made maps with key strategic features etc, and select randomly from those, or a randomly generated map - perhaps locking them once a battle has been fought on a particular hex.  Even with hundreds it will eventually get to the point of oh its this one again.

Well most RTS combat is nothing but zerging and selecting a bunch of units and throwing them at the zergers. RTS is pretty damn boring. The only RTS game that actully felt epic was the total war series.

Give me TBS over RTS any day

This sould be the tactical interface

I didn't read throughout all the different posts, but I think the easiest way to stay true with the stats that we have in the current build and still get several options to align is as posted already some pages ago:

* Armor and cavalary will modify the total amount of AP

* Weapon choice will be related to the amount of AP a strike, shoot, casting will cost (e.g. high level staff or book could for example reduce the casting cost by -0.5 AP)

* Spells should cost 1/10 of the mana it costs to cast rounded up to the next integer

* special types of skills should have an own AP cost.

My two cents

Elements of Tactical Combat

In no particular order these are the things that matter:

• Combat Speed. Your combat speed determines how many “moves” / attacks you get during a particular turn.  In the begging of Lord of the Rings, what makes Sauron such a bad ass is that he can attack so many units at once. He has, in game turns, an incredible combat speed.

I like this, so long as it stays strictly controlled who gets these extra attacks - A non-heroic swordguy might have maybe two at maximum level, but something outnumbered usually (a dragon or a hero) should get several.

I truly hope that "combat speed" and "movement speed" have nothing whatsoever to do with each other; Movement speed is important enough on its own, giving it as a freebie with increased attacks is madness.

• Morale.  Unit morale matters but for fun purposes, we try to keep it straight forward. Units have High Morale (25% combat bonus), Normal Morale (no bonus), Low Morale (25% combat penalty), and Panic (you don’t control them). It provides a whole new avenue for us to play in.

I can see this working well in a way similar to the X-Com system, though I think there's a lot more to "panic" than is mentioned here - How are they panicked? Are they moving towards the edge of the map to retreat, moving towards friendly units, immobile in fear, or going into a berserk charge against the enemy?

Hopefully a chance for all those things.

• Terrain. This is where the tile based part mattered for us (and for the AI). Some terrain, obstacles, and tiles simpls y provide better offensive and defensive bonuses, Controlling them matters.

"Better offensive and defensive bonuses"? There's so much more potential for terrain than this it seems that it's not being exploited. Walking in snow? Get cold and wet, weakening you versus ice elemental attacks, strengthening you against fire. Fight on a ford across the river? % chance a non-heroic unit will fall in and take damage proportional to armour-class.

This doesn't even touch on fortifications which I sincerely hope won't be along the same lines as Master of Magic's "Whoops, forgot to close the gate" city walls and nothing else. Gates you have to batter down (and can be upgraded to passively "defend" themselves with boiling oil). Walls that certain units might be able to damage, or certain siege weapons might try and bypass but otherwise only infantry will be able to approach, using ladders to get up on top of the walls to try and reach the gate house in a desparate attempt or the seige fails.... Deadly magical fire shooting towers...

As has been stated in the design document, city-spam is to be discouraged; individual cities are valuable - more so as they age; and generally cities are important. This means that city combat should be the most difficult for an attacker as well as some of the most interesting.

• WINNER. TAKE. ALL. This is the part where we want to hear your opinions. We do ask that you keep an open mind on what we ultimately go with.  My opinion is that the attacking player has the onus to finish the battle in N turns. After N turns, the attacker morale starts to go lower and lower at which point the defender can come out and make mince meat out of them.  The question is, what should determine what N is?  Or should we allow retreating? Should we allow draws?  I’m against retreats or withdraws because it’s one of those things that allows the game to drag on. It’s a strong personal preference of mine that two men enter, one man leaves. (Your heroes will tend to escape though).

My gut opinion: Don't limit combat on the assumption that combat is boring. Make combat fun and people won't mind multiple combats.

Gut opinion 2: Avoid instant-win conditions where possible. A ranged flying unit versus a melee unit is an instant lose scenario in MoM. A fast moving archer unit is either an auto-win or an auto-draw where the fast archer loses nothing and the enemy loses through attrition. Encourage secondary weapons (not dualwielding, which is arguably a single weapon choice, but a specific type of weapon choice) for a ranged alternative, and always, but always allow retreat in the face of impossible odds - but in the form of moving to an escape zone (such as the map-edge).

This all boils down to what you're fighting for, and whether a combat is just a straight forward issue of team A versus team B, Installation Defence, City-Raiding....

Out in the open, two squad skirmish should be fairly simple - both attack, one dies. Easy. Winner gains a tiny hint of a smile at their glorious triumph in the Battle of "log crossing small stream". Unless one unit gives way intentionally by retreating, this should rarely end in anything but annihilation.

With an installation, not only are there things like reinforcements to consider, it's quite possible for an attacker to win with zero kills and 100% casualties, so long as they destroyed the installation. This would provide a huge morale penalty on the defenders, most likely causing a rout.

For attacking a city, as mentioned above, first  you gotta get inside. When you're in, then you can either take out enemy troops, or, if you're not interested in holding the city, start attacking buildings and people (an attacking monster would always do this), destroying valuable buildings (and replacing them with rubble that must be removed before the city can replace it) and generally looting the place until their morale drops and they flee with their booty.

Ultimately, in any confrontation, a fast and loose turn limit is good, with an attacking force losing morale each turn after.... say the tenth, until all units begin to auto-rout, but this not a hard limit - killing an enemy gives morale bonuses, objective completion gives morale bonuses.... It might be the defenders who retreat, or the attackers may retreat very quickly when attacking a fortified stronghold where they just can't do anything to the defenders.

Combat would be slightly skewed in favour of the defender (and to enforce an eventual conclusion regardless, after a hundred turns or so, both sides would start losing morale (on top of the attacker's automatic penalty). One way or another, somebody's going to flee pretty quickly after turn 100 regardless of the number of morale boosting spells the players are using.

There are other aspects to "encourage" short combats, but I'll cover these in Richness below.

So... I keep saying that units retreat, how to stop them from just attacking over and over again?

The best way would be to handle things in a way similar to the old "Ogre Battle" game. For those not familiar with it, a unit whose leader died would retreat to their home base, where they picked up a new leader unit and could rejoin the battle.

So, when a unit retreated from combat, they are cut off from the chain of command, their unit icon is greyed out, and they move to their nearest "base". This could be their nearest town, a nearby Leader hero, or a "camp" tile improvement. Heroes never suffer from this behaviour, and a hero joining any greyed out stack could reactivate them, as could certain spells, but by default, a defeated army will retreat, not retry. Thusly it wouldn't be in your interest to attack without preparation - those units may rout, which would take them out of operation for a few turns.

Naturally, defenders that rout against raiders who then fail/decline to take the town when presented the "capture or pillage" option when they win (please try to keep "razing" to an acceptable minimum - you don't "raze" a vast city of millions, you plunder it over and over for a few days - the process of razing takes time, especially when cities are supposedly unique and important). Plunderers on the other hand retreat to their war camps, giving a plundered city some time to recover for a new attack (unless they're severely beseiged). It should never be viable for a single, weak unit to capture a town - it would almost instantly revert, killing that unit. A realistic degree of military presence should be mandatory to hold a town, and would need to stay there, and even then, la Résistance efforts would still probably kill some units though (difficulty based on global reknown, alignment, race and helpful spells).

As an interesting side note - this would be perfect behaviour for wandering monsters and bandits, who would attack, ravage a little, then return to their nest/camp before making another attack - and a valid way of dealing with a dangerous wandering monster would be to destroy its home, where it would be forced to the next nearest habitat possibly on the other side of the map near your neighbour.

Combat stays rare without forcing a player to invariably lose units because they have no alternative.

• Combined Arms. Archers have range. Mounted Warriors have great combat speed. Foot soldiers tend to have better weapons and defenses.  It means putting together your army matters a lot. It also is important to us that players understand precisely why they won or lost a battle.

I'm vocal enough about wanting at least as much from the Elemental combat system as I got from the MoM combat system - Unit quantity, role, special properties.... So I'll skip that part, and go onto another thing based on it - unit role.

It's fine to say "Oh, he's an archer, he arches things", and "he's a mountie, he has a funny hat and moves around fast", but this is just a vanilla unit, and what if I wanted to specialise one into a role?

I'll use a well-defined set of unit roles for an example, since they were explained clearly enough for me there:

Striker - This unit is about putting out great damage, primarily single units. Attack High, Defence Wet Tissue.
Defender - This unit is about keeping enemies targetting, and close to, them, or at least away from others. Attack Limitted in scope, Defence Great.
Leader - This unit helps their allies, curing or buffing them and generally making them perform better.
Controller - This unit is about battlefield control and areas of effect, dealing out wide-ranging damage, terrain modifiers and other sneaky activities.
Both leaders and controllers have middling attack and defence stats, in favour of more tactical special abilities.

A Ranged Striker should not be no more similar to a Ranged Controller than it is to a Melee Striker.

For an example, I'll cover Strikers in some more detail, then loosely define the rest to keep things brief.

Ranged Striker would have abilities like "poison arrows", "snipe" and "rapidshot" that mean its damage is high, its accuracy is high, and its number of shots is high. RS: Dealing lots of damage to distant enemies.

Mounted Striker would be a lancer, with abilities like "charge", "brutality" and "follow-through" that mean it gets into combat easily (and gets damage bonuses for charging), deals great damage when it gets there, and if it takes out a unit, it can spend any remaining move, possibly attacking another unit if it can. MS: Moves into combat quickly for glorious, high-risk charges.

Infantry Strikers are fast moving stormtroopers - "agility", "sideswipe" and "rend" that mean it moves past units easily, can attack units as it passes and keep moving, and rends lightly armoured units for massive damage. IS: Moves through enemy ranks to attack vulnerable enemy units.

RD: Ranged Defenders use special ammo and suppression fire enemies to limit their enemies' ability to attack and advance.
MD: Mounted knights with incredible armour get into the thick of the enemy quickly and deal enough damage to be a viable target, whilst being tough enough to take it.
ID: Slow moving heavy footman who autoattacks units trying to pass it by whilst being very resistant to ranged fire.

RL: Ranged Leaders Use long range attacks to create vulnerabilities in their targets making them easy for other units to take down.
ML: Heroic Charger Whenever they charge into battle, they inspire all nearby units with morale bonuses and such.
IL: Stoic bodyguard who gives bonuses to all adjacent units and shares its defence stat with them.

RC: Rogues who can place invisible booby-traps on tiles as they move, injuring the next few units who move across the square whilst firing on them from afar.
MC: Elephant riders who can use their mounts to devastate nearby terrain tiles.
IC: Stealthy assassin units who have a limitted amount of ranged poison grenades, covering tiles with spreading toxic smog which covers more area, but deals less damage, turn by turn.

Notably excluded: War Machines, Monsters, Summons and Mages.
This is a tiny glance into the principle, but I'd love to see armies that interact in this way, not just being as many of the fastest, highest-strength, highest-defence units as I can get. If I win or lose a battle, I want it to be because their army had better cohesion and teamwork, not "more ranged guys", because everyone knows Range Beats Melee.

• Thresholds. Players can set the tactical battle threshold in the menu. That is, they can say it requires 10 units on each side before it’ll actually go into tactical battle.  At any point, players can have a tactical battle auto-resolve.

I would never auto-resolve a combat unless I had overwhelming superiority, such as three heroes, two dragons, a siege catapult and a fifty-foot tall battlemech. As such, any threshhold would be irrelevant to me unless it could be set to "no contest" wins, where your forces have a projected expenditure of resources amounting to zero.

Controlling the length of a tactical battle.  We believe that users should have a lot of control over how in depth they want their battles to be. Should a tactical battle finish in less than a minute or should they last 2 hours? How do we make it so that players can control this?

That depends on the battle. If it's an epic battle for an enemy capitol, fully upgraded in every sense with a massive defending army, I'd be happy with a longer battle than "You meet a level 1 bandi- *splut*". In all honesty however, I'd prefer the system mentioned above - where an attack is limitted in time, and requires a little set-up to create a war-camp nearby where the units can retreat to and recover, spreading a long-term siege over several turns rather than overdosing on one endless combat. Siege warfare takes weeks, nor hours, and reinforcements could arrive at a well defended location before losing it only if the battle wasn't arbitrarily decided over a single turn (which is what, a day?). 

Randomization vs. Richness. I won’t lie to you, we have a trade off in front of us and it’s a big one.  We can randomly generate the battlefields in tactical combat OR we can have it pick from a series of pre-made tactical battle maps.  The randomly generated ones won’t be as interesting but they’ll more accurately reflect the local terrain.  I’m preferring the pre-made ones because we can add some spectacular strategic when we’re crafting them and have hundreds to pull from.

Both.

Seriously. Again, check out the X-Com series. Maps are random, but they're randomised segments, not purely randomised. Each map could/should consist of a certain terrain type - partially randomised with the open spaces and the uninteresting, partially randomised with component segments that form centrepieces for the map.

Example:

Farmland fight near a city. The default random terrain is wheat (minor move penalty, minor defence bonus), a few fences (can't cross them as they're waist height), maybe a well and some irrigation channels/ditches. Few to zero trees and some movement slowing hedgerows. A few wandering neutral aligned peasants are wandering around and have zero impact on the battle (self-defense only, not a victory condition, maybe a small penalty for killing them to the land's production for a few turns).

On one part of the map is a farmhouse with windows and maybe an expellable peasant family inside. Archers could use this for a defensive position, shooting out normally, but with great advantages versus those shooting in.

There may be more than one, there may be a few. Each "square" might be pulled from universal pre-gen terrain and so there might be an entire map littered with farmhouses, or an entire map of default wheat - random and rich at the same time.

Fighting in a cave. Most parts of the area are inaccessible - either cave wall tiles or littered with rubble and stone pillars. Damaging the walls or pillars may cause a cave-in, damaging random squares throughout the surrounding arnea.

Interesting points include a bat's nest - moving too close triggers the roost to fly away into the surrounding areas of the map, causing huge accuracy penalties to everyone within, until they leave the area. Underground lake - a shallow lake with something lurking within that might eat a passing unit every now and then.

Mountain: Combat takes place in very confined mountain paths, each with the small chance of instant death with every injury.

Interesting points: Solid ground where you'll fight for the opportunity not to fall to your death etc.

Fighting in a city: Not random at all. Each building has its own pre-defined square. Defensive city design is its own reward.

Forest:  Random trees everywhere ruin archery for non-woodsman archers, interfere with movement and generally make life difficult for large units.

Interest: An open glade with flowers.... These flowers may be pleasant and morale boosting, be currently snacked on by a unicorn, or be sleep blossoms, randomly selected at battle start.

This is generally the combination of both worlds - the strategy offered by a structure is only possible with a pre-gen, but the randomisation of which pre-gens to include (with fairly easy rules on how to make each square "accessible" to surrounding squares). Random generation is the chain which links the rich points of interest together.

Monsters: Random generation part deux, and the thing that will encourage shorter battles I mentioned - As a randomly generated pre-generated square, a monster tile. The monster depends on the area (farmlands have peasants for example, a volcano might have elementals or fire giants etc), and has its own pregenerated lair. This monster may be far, far more powerful than the players on either side could handle together, or something they might be able to deal with with teamwork (followed by then fighting over the spoils).

One wandering monster generation could be "Roc". Giant monster bird that will, every other turn, fly over to a unit, hit it with its two attacks (and hits powerfully enough to probably insta-kill both targets), then returns to its nest - to feed its young chick. It's a victory conditional, meaning it must be killed before the tactical battle is won for either side. Beat the Roc, get a baby Roc resource and possibly some treasure.

So now the battle is a race, either to kill the other side and retreat, get out of combat entirely and avoid mountain tiles for awhile, or try, somehow, to kill the Roc, with the other team either helping or attacking you the whole time.

Best of all, if not cleared out, this monster lair can then become a feature for the area, creating a quest for heroes to clear out the monster, or leave it in place, whence it will start showing up in every battle on the square.

Other monsters may not be victory conditional at all, but still perfectly able to punish nearby units for taking too long with the area. A passing herd of Dire Rhinos might not have a lair, but still provide a significant encouragement to finish battle before they get mad, and these could be added to any map square as a potential hazard.

Just my ideas on getting things random (interesting), and rich (fun).

I would like to know, what is the result of discussion, which method of battle map generation has been choosen in the game: random or premade maps?

Full Disclosure:  I'm not in the Beta, so I am basing my comments on GalCiv2 and reading other posts.

My 2 cents:  Allow retreats, but only if there is an available spot on the strategic map.  A surrounded/besieged army has to fight it out.  This will be fun on the strategic level as you try to box in armies, and attempt to surround the other player with weaker forces while you hit them with you main force.

Retreated units are scattered on the strategic map as much as possible.  Furthermore, the units are stunned for x turns based on their current morale. 

A stunned unit cannot move or do anything else.  It can be apart of an army, but does not engage in tactical fights, and is instantly destroyed if the army loses a fight.

For the people that like hit a run tactics, that should be a strategic ability (played on the larger map), not something that needs to be resovled on the tactical level.  I really don't think it would be fun to play a series of battles where cavalry archers run in, fire a shot, and then run off the map.  Rather, it would be better if a "Raid" option was avaible, and both forces fire x number of ranged shots at each other, and damage is done accordingly.  On the strategic level, this raiding can be countered by beefing up archer power or with abilities such as ambush, etc.

Any thoughts?