Domains at War: Rules I Got Wrong

During my first play-through of the Domains at War rules, I really did only a quick skim of the first couple sections of D@W: Battles. Even then, I missed a few things that now jump out on a second, closer read-through. I'm writing all these down, in hopes that I'll remember them for next time.

1. Units have individual morale modifiers based on type, just like in 80s-era D&D (or my own FF rules). For example, most cavalry gets an innate +1 or +2. These aren't mentioned on the tables that contain Rout or Morale modifiers, but they are mentioned in the text. This is pretty important for the sake of determining whether a given unit will rout, since even  +1 modifier on a 2d6 is a substantial effect.
2. Disordered units get a -2 to AC until they become ordered again during activation. This is very important when dealing with high-AC units like those blasted wyverns. Hit them once, and every additional attack against them is at a 15+ instead of a 17+.
3. Units in a division each move and attack one by one. When asked about this, I answered they all move first, and then attack all at once. Not so! You move a unit first, see if it hits with an attack, and then move the next unit, one at a time. This means that you don't even need to decide which unit to activate until you see what the last one did. Maybe not very realistic, but it does create some new strategic options. (Shoot a missile volley, if it disorganizes a target, converge on that weakened target to kill it!)
4. Units can only voluntarily retreat from the map if they reach the map edge. Since routing automatically "teleports" units off the map, I figured that voluntary retreating works the same way.
5. Units can't move through between hexes that are both adjacent to the same unit. So you can't move around to attack a flank, once you're already in front of a unit. This is true even if the hex isn't threatened (in front of) the enemy unit.
6. Crossbows (unlike other ranged weapons) can only be used by a unit that stands still, not by one that marches.
7. Units that ready to attack not only get a free attack at anything that attacks them or a neighbor, they can also force that attacking unit to switch to them! This creates another reason to use the "phalanx" units to merge polearm infantry.
8. Units cannot focus-fire their targets, if another target is available. That is, if at the end of movement you have the choice of attacking a unit that hasn't been attacked that round, you must attack it (instead of that badly damaged unit next to it). This is another way that strong units can protect weaker units by remaining adjacent.

Of the things above, the one that makes me a little uncomfortable is the business about moving to the map edge to retreat. It's odd that an involuntary rout is a much more efficient way to get off the battlefield than a voluntary decision to leave. It creates a perverse incentive to find clever ways to fail a morale check, in order to whisk away your valuable leaders and elite troops to safety. I think I would rule for consistency that a unit may elect to automatically fail a shock/morale throw in order to voluntarily rout.

I also posted to the designer forums, to ask about the consequences of running away from a battle. In a real campaign, there are plenty of times when this would be appealing to avoid being crushed in an imbalanced fight. It didn't seem right that this seemed to be so easy, just by stepping back off the map at the start of a battle.

The answer is that, in the Campaign system, a lost battle immediately triggers a loyalty check by all fleeing units, in addition to the usual pursuit rolls. This is like a morale check, but results in the permanent desertion of the unit. Note that as a special perk religious commanders (cleric, paladin, etc) get a small number of absolutely loyal followers who can run from any battle (or endure other calamities) without needing to make this check! They would make pretty good scout or guerrilla warfare units, I think.