More than one Day battle thread is here.

50% for a 90 isn't too high.  That's about right for maintaining formation.  It's not the simplest thing to get a bunch of guys wearing armor and carrying pole arms to change direction while maintaining formation.

For that, it really depends on how many units in a stack.  Say if a stack has 1000 soliders, it is about right.  If a stack has only 10 griffins, 50% is too much.  This % depends on the unit & its quantity.  Conincidently, it will make bigger army harder retreat, which is nice.

Ireally like the option of moving units to the edge of the battle in order to retreat. No additional losses need to be added, if you can get the units to the "leave the battle" boundary, then they escaped.

I would REALLY like to see this also use surrounding terrain tiles. For instance, Player A has an army on a tile, and to the North is impassible mountains, to the West and East are plains, and to the south is the attacking army of Player B.

Player B moves onto the same tile as Player A's army. The tacitcal battle should have Player A's army in the northern portion and in formation towards the south. Vice versa for Player B's army. Player A CANNOT retreat north due to the impassable mountains - the only escape options are East, West, and South thru Player B's army.

If he splits and goes both East and West, then the army is now divided into two stacks. If he moves the entire force either East or West, the units that are closer to that will likely escape without a lot of loss. But the units that have to cross most of the battlefield perpindicular to the attacking force are going to take pretty heavy losses. And of course South thru the attacking army would be devastating.

This would also make it so you could corner an army with multiple armies. Fleeing in a direction with another army in it would take you from one battlefield to another. Technically the second army would not even have to belong to Player B, it could be Player C's army that Player B "pins" Player A's army with. Of course, if C and A are allies, then A could retreat to the tile occupied by C.

If you managed to chase an army onto a peninsula with no escape on three sides, it would force them to make a last stand as the only exit is on the other side of the attacking army. I think this is a good method of handling retreat because you can make not just the square that the fight is occuring on, but surrounding map tiles matter.

Like Climber said it really depends. Yeah maybe a squad of rookie pikemen will take some effort to turn around, but a squad of even rookie swordsmen, let alone veterans, would be able to turn around on a dime. It does not take much effort to turn 180º in place, even with armor. Having a giant pole with a blade on top makes things a little harder because if not everyone keeps them perfectly straight people will likely get their poles entangled (wow that sounds wrong). On the other hand, trained pikemen should be able to accomplish this as easily as swordsmen. Obviously it makes sense for cavalry and the like to require more effort to turn around, but then they also tend to be able to move farther and faster and so using 50% of their movement just to turn left could still be excessive.

It also matters what the formation is and what it is going to be. If you just want each individual to about face, it would be very simple in pretty much all cases except extreme ones. On the other hand, if you want to rotate an entire formation, requiring troops to actually move, it would take longer. But then this would mean turning 180º would be cheaper than turning 90º and that might feel weird in game.

I'm leaning towards no movement costs for turning. I like the way it is in Total War - you can rotate your troops and formations whenever and however you choose and the major factor to consider is "is the enemy going to hit my ranks while I'm making them turn?". That won't work so well in a turn-based combat, though...

No real comment on the specifics, but "Yes, please" for the general idea. I wanna see *something* that makes terrain more important in strategy than the simple path-blocking objects GC2 has.

That also depends on what a turn is.  If each turn is an attack round, a likely event, you're not talking about five minutes of walking or spinning in place.  You could easily be talking about 20 seconds.

Sure, but that still seems secondary the question of how the (psedo)turns in tac combat will relate to the (real)turns at the stratgic level. If a tac combat begins with forces in four adjacent tiles, are they all automatically involved regardless of who attacked whom? If not, how many tac turns will the contesting forces have before the other two forces get a chance to consider moving away?

Your post reads like you're talking about one thing, but you're replying to something else, so I'll make two replies.

Assuming you're talking about adjacent forces in tactical combat, which is the subject I was on:

Go read a warhammer rule book.

With modern computing power, it's not a chore to run them the real way in real time.  I'm hoping units will be made of actual units, and not just representations of a number of units.  In that case, they'd be attacking whoever they're physically attacking.  Without such a system, combat will be quite shallow by comparison.  You can't do much when the entire unit magically attacks regardless of their location relative to the enemy troops.

Assuming you're talking about adjacent forces on the strategic map.

First thing first, adjacent armies should never automatically fight each other just because they're present.  Without initiation by moving into the same square, it's a bloody mess.

Second, this was the only thing I liked about AoW in comparison with MoM.  Assuming hex, all adjacent armies to the defending target should be involved.  Assuming square, either all adjacent or all adjacent and perpendicular armies should be involved.  Hex is much cleaner for this matter.  Such an implementation is both fair, sensible, lays out nicely, and gives added strategic depth to force deployment..

This sounds like it would be a really accurate way of working out retreats with the most tactical depth possible, as basically you fight out the retreat on the tactical map rather than having some algorithm sort it out. As a result you should be able to make the neat tactical decisions of whether to sacrifice some units to save others, and if you can pull this off then that's how it is. This would take into account whatever battle mechanics are used in the game with regard to morale/flanking/unit speed etc to model the retreat as effectively as is possible (everything that's in battles in the game would be taken into account in this system with everything exactly as important as it is in a normal tactical battle... as you're fighting a normal tactical battle :D).

For people who might feel this is a little involved when they simply want a "Get-The-Hell-Outta-Here! Button" given that the devs have said previously that we can speed up or slow down the speed of tactical combat as much as we want you could probably just select all your units, point them to the edge of the map and crank the speed up to max which would be pretty similar to a formula randomly deciding if units die dependent on their speed/location relative to the enemy.

One aspect of retreats that I dont' think anyone's brought up.. which I think perhaps could bear some consideration is that it can allow some fairly cheap tactics with regards to spell casting. What I'm thinking of here is the situation where you send in light cavalry against a more powerful enemy who cannot move as quickly as them and thus cannot engage/stop them retreating. So you attack the enemy, dance around the map a bit not engaging while shooting off your load of combat spells for that battle (or however this will work, I'm assuming there'll be some limit of how much you can cast per battle and it won't just be your entire empire's mana pool) then when you're out of spells you retreat your unit of cavalry. You then attack again the next turn..burn off your spells.. and rinse and repeat until the more powerful enemy unit is dead. Doing this you take no risk as to losing your unit of cavalry and can beat considerable odds simply by persevering and burning off mana.

There are of course some assumptions in the above example, obviously if the enemy you're attacking can cast spells themself or if they have sufficient ranged attacks to kill your cavalry then this won't work... but if it's something like a great behemoth monster guarding a lair with massive amounts of hitpoints/conventional attack stats but low speed with no ranged capability/spells... this could take them out every time which hopefully it will be possible to avoid in some way :s .. although I'm not sure how.

Oh yeah, another thing I'm assuming in the above that you can even cast combat spells when your channeler isn't actually present at the battle.. we haven't had this confirmed at any point to my knowledge, but if we can't then the title "War of Magic" might seem a little misleading

I like a 'risky' retreat option. I think first you should have to disengage your units, and move a certain distance away from the assault point. Then you can attempt a retreat, but some units will almost certainly be lost as you leave the field of battle. It would be optimal to be able to designate a certain number of units as 'blockers', from which group the units killed in the retreat would be chosen from.

My optimal system would be where you must choose 2/3s of your units to be near the blocking line, and then each one of those has a 1/3 chance of getting wounded, a 1/3 chance of being killed, and a 1/3 chance of escaping scot free. This allows you to preserve important units, but still gives a potentially fairly heavy toll for retreat.

I feel that a retreat penalty is important, as it makes you weigh the consequences of going into a battle. Certain units should have a 'guerilla' ability that allows them to retreat unharmed no matter what, but no serious army is going to be made purely out of such units. This system would force calculated large battles, but still allow for small scale guerilla raids. Which would nicely mirror real-world tactics.