I originally posted this over at the Total War Forums, in response to a query as to what features I would like to see in the upcoming Rome II: Total War PC game. I have been playing this series (mostly in the form of various mods) since the original Rome: Total War and therefore have a pretty developed sense of what I’d like in the next one. I’m also a long-time student of classical warfare, so most of my requests are for mechanics that improve the realism and tactical depth of the gameplay (I think they already have a relatively good handle on strategic depth).
After I finished the marathon forum post, I realized that it makes a halfway decent primer on the mechanics of classical combat, so I’m reposting a cleaned up version of it here for those who are interested.
I wanted to suggest some things that I haven’t seen come up yet and would like to lend my support to some of the ideas I have seen in this thread. I’ve listed these in an order that blends how important they are and how easy I think they would be to implement. The two biggies are morale and fatigue.
1) Morale is the Key
In battles fought primarily hand-to-hand, the winners often suffered negligible casualties; often fewer than 10%. The losers, however, often lost up to 30-50% of their army. This is largely because the great majority of casualties in all wars occur only after a unit breaks and flees. The effect is particularly magnified in hand-to-hand battle because a soldier must not only get close enough to the enemy to see into their eyes, but then plunge a blade into them; hear their gasps and screams of pain and see the injury they have caused. All the while they are close enough to be gutted themselves by the victim’s buddy. Even if they are perfectly willing to kill they probably value their survival over the death of an enemy soldier: their blocks and parries will be wholehearted, but their attacks tentative and probing. A good rule of thumb is that only one man in ten fights to kill the enemy. The other nine fight to survive.
The upshot is that units are not meat-grinders made of meat, which is kind of how they’ve been modeled in other Total War games. This is not a criticism; I have spent many an hour enjoying these games but I think that it would be innovative, fun, and not all that difficult to try a more realistic path. In game terms, morale rather than the number of casualties is what determines victory. Flanked units should run away long before they are chewed up to the last man. Another good rule of thumb is that units generally run before they are reduced to half strength. Depending on training and other factors, they might get close to that or even exceed it, but only in truly exceptional situations should they stay in the fight until they are down to 20% or so.
Even elite units should collapse like wet tissue paper when flank charged by cavalry, if only to seek a better position. And that is key, only particularly low-quality units should flee off the map the first time they get routed. Well trained troops will skedaddle a few dozen meters before their NCOs and officers get them back in order and aimed at the enemy. However, routs should be contagious. Having a routing cohort run back through your reserve might very well cause them to flee too if they are none too steadfast. In general, being near routing units should cause a morale penalty. The exact penalty can be set for gameplay purposes.
You can see how having some light cavalry or even just fresh infantry around to chase down your fleeing foes goes from “a nice thing to have” to “absolutely essential” now. It’s where most of your kills/captures will come from. Speaking of captures, having units surrender when surrounded would be a nice touch that I’ve seen other people suggest.
Another result of this whole not wanting to die instinct so common to humans is that troops tend to be reluctant to close with an enemy. Poorly trained, tired, or demoralized troops should close slowly and cautiously while well trained, fresh, or high-spirited troops can charge full tilt if ordered. Particularly aggressive troops who are poorly trained might rush the enemy just to discharge the incredible stress of combat.
Another interesting feature would be a whole-army morale figure that could be affected by factors like the personality of the general and the win/loss ratio the army has experienced in the past. It would manifest as a modifier to the morale level of all units in the army when battles begin. Ambushes might also do this. With this feature, if a crack army is put in the command of an incompetent general, it will fight at rather lower effectiveness.
2) Combat Fatigue and its Effects
Hand-to-hand combat is incredibly tiring. Even a person in perfect physical condition who is well fed and rested can keep it up for no more than 5-6 minutes at a time. An untrained lout or a man who last slept the day before yesterday, has been on half rations for a week, and marched 10 miles to get here can manage for no more than 2-3 minutes at a time before suffering exhaustion. The way this played out in real life is that once units did get up the psychological moxie to really close with the enemy, an engagement between two units would last for only a few minutes before each was forced to withdraw to some safe distance (researchers disagree on what this was… probably no less than 10 and no more than 50 meters) to rest for a few minutes and get some water.
In “barbarian” armies this likely played out in a distributed, uncontrolled way, the impulse to fall back being communicated from man to man by nonverbal and subconscious cues until the whole front rolled back for a few minutes. One of the things that made the Romans (and other organized armies) special is that they performed these small tactical retreats in a coordinated manner (the whole phalanx for Greeks, probably by maniples/cohorts for Romans). This would continue with one side gaining ground and the other losing ground (as well as stamina, men, and morale), one way then the other, until one side broke and fled. The existence of tactical subunits in the legions (centuries) meant that not only could really exhausted centuries be rotated back and replaced by fresh ones during the lulls, but that routs tended to be confined to one century and had a harder time spreading to the rest of the maniple/cohort, or the army. This implies that in-game, barbarian armies should be made up of fewer, larger units that therefore cannot be as easily or deftly maneuvered and give the player somewhat fewer options with regards to controlling the spread of routs.
Among other things, you can see how battles could go on for many hours without all of the combatants being annihilated; because most of their time was spent exhaustedly staring at each other and hurling insults or pila (or both). Combat was a series of crises, rather than one big one. The lack of this mechanic is why Total War battles have such an absolute feel to them, with victory usually being determined by a single factor or maneuver. Humorously, the absence of this frictional (in the Clausewitzian sense) element causes the battles to approximate a single decisive moment of pure force vs. force; the logical phantasm of Total War that Clausewitz warned against!
In game terms we obviously don’t want battles to actually last 2-10 hours. This dynamic, though, is absolutely essential to the way armies fought and is one of the reasons why even though the Roman legionary might not have had higher “stats” than his opponents, he usually managed to beat them. To a large extent having an organized means of distributing combat fatigue through the army is the essence of the Roman way of war. It’s why even though their grand-tactics were simple, often even boneheaded (see: Cannae, etc.), they tended to punch rather above their weight.
Mechanically, this can be tackled in a few different ways. The simplest is not to model the actual phenomenon at all and just give certain armies an arbitrary stamina boost (Roman, then Diadochi, then etc.). If you do want to model it then you can shorten the lulls to perhaps only 30 seconds so that the battles aren’t too slowed down. This links into my point about morale being important, because you don’t have to increase the lethality of combat to compensate. Not that many people died up until the rout. Units should pull back from and rejoin combat automatically to reduce the amount of micromanagement needed. Units pull back slowly and defensively; shields raised and on the alert. If harried by a fresher enemy who sticks to them, they should fall back before the enemy while minimizing contact. This should tire the pursuer more than the pursued and will equalize things after a minute or so unless there is a great disparity in fatigue or morale, in which case the pursued might break and run. In general, what I want to see more than anything is:
a) Fatigue must become much more important in combat. It should apply a penalty to fighting that causes units to end in a truly exhausted state where they can hardly raise their shield/draw their bow, and in which they are incredibly vulnerable to an attack by fresh troops. This will make the presence of a reserve vital (as it was/is in real life) and makes for a much more tactically interesting game than “outflank=win.”
b) Increasing fatigue should apply an increasing morale penalty that makes tired units more likely to flee. Since fatigue becomes so important in this case, it should probably be graphically represented in some way. When units become tired enough to go into “defense-only mode” and fall back, fatigue stabilizes, while resting out of contact causes fatigue to decrease. The exact ratio of increase in combat/decrease while resting can be set for best gameplay.
Have fun modeling the curses and obscene gestures during lulls.
3) Bigger, More Varied Maps
I would love to see bigger battle maps with the opponents starting further apart to allow for some grand-tactical maneuver prior to the fight. Seizing a hill, taking up a defensive position along a stream bed, or fooling your opponent with a feinted outflanking maneuver should be possible and have an important effect. As it is, deployment zones tend to start armies so close to each other that larger battles give you little choice of maneuver after your initial setup. The Blue and the Grey, an Empire Total War mod does a good job of this with its historically-based battle maps.
4) Archery and Skirmishing Changes
Previous Total War games have had trouble getting archers just right. They tended to either be useless annoyances or game-winners and you had your pick of the two. This is because the way archery has been modeled misses a vital subtlety: archery at long range and at short range are two totally different phenomena. At long range, even powerful bows are harassment weapons that demoralize an opponent more than they cause casualties. The archers do their best to arc shots such that they land in the general vicinity of the enemy, and those arrows that do hit someone don’t strike with much force. Unless the enemy is completely unarmored this kind of shower-shooting is unlikely to fell many of them. At short range, within fifty meters or so, the archers are performing in an aimed direct-fire role and bows become very deadly indeed.
Unless the enemy is heavily armored or has a large shield the archers will pincushion them with arrows that strike at very nearly the speed they left the bow. This is part of the reason archers can be deployed ahead of hand-to-hand troops or directly on their flanks. Most units trying to charge archers will slow down to a crawl as they attempt to ward off the arrows and will eventually be forced to fall back to harassment range. Heavily armored troops may still become so demoralized that by the time they reach the archers, they could be routed by them in hand-to-hand combat.
Gameplay-wise, this ties in with the morale mechanic I suggested above. At long range, shower-shooting causes minor morale drops in the target unless it is unarmored or poorly trained. Units that try to close the distance with the archers will begin to lose much more morale when they enter close-range. This slows their advance and keeps them in the field of fire longer. Eventually their morale drops to a certain point and they are forced to fall back to long range. Well trained troops in high spirits may make it to the archers, so a “skirmish-mode” should still be included that lets the archers retreat from attackers. Now it’s easy to see why horse-archers were so terrifying. A man on a horse can close to direct-fire range of his enemy, pepper them arrows, and retreat to safe distance with great rapidity. Foot-archers with their larger bows and greater stability are the historical remedy.
The pilum was probably only thrown at a maximum range of 20 meters from the enemy. Giving the iron-shafted pilum its historical effect of sticking in shields to render them useless would be a nice touch, if there ends up being time to model it. Slingers should actually have about the same range and effect as archers.
Cavalry is a weapon of psychological shock much more than physical shock. This is especially true in the classical period, prior to adoption of the stirrup and high-backed saddle. One often hears claims that “a medieval knight could transmit X0,000 lbs of force through the tip of his lance!” but this ignores the Newtonian fact that doing so would tear his arm clean off and/or toss him twenty feet backwards. Heavy cavalry doesn’t charge into the heart of a formation of close-packed infantry because horses are not automatons and have a finely developed instinct for self-preservation (augmenting the often underdeveloped one of their rider).
It is true that they can be trained to hit a man or two with their body, but even at a gallop they lack the momentum (and will) to bowl over five or ten of them like the charge of the Rohirrim in the Lord of the Rings movies. Horses are, however, utterly terrifying. I know this flies in the face of all the eight-year-old girls in the world, but it is true. If anyone disbelieves me I would ask them to go to a barn and walk an adult horse around for a bit. You quickly realize that if it wanted to, the horse could smash you to jam with very little effort. They have more muscle in their neck than many of us have in our entire bodies. Now imagine 500 of them running at you, all ridden by angry men with sharp things. But, like human soldiers, warhorses care much more for their own survival than for killing the enemy; the one-in-ten exist among them too, however, and there are many cases of warhorses doing rather terrifying things to enemy soldiers. Like picking a man up by his stomach, carrying him to a snowbank, and tearing out his intestines with her teeth. It would be interesting to see that in Rome II.
My point is that cavalry charges should result in massive morale drops before they hit a unit. The precise level of the morale penalty should be proportional to… well, how scary the cavalry unit is. Cataphracts in their gleaming scale-armor should cause wet spots to suddenly appear under enemy infantry, whereas light skirmishing cavalry should cause only mild unease. The closer the cavalry gets and the faster it’s moving the more of a morale hit the charged unit should take. This could be done with a function that uses the rate-of-change of the cavalry unit’s distance from the infantry unit to calculate morale drop. Morale drops from multiple cavalry units should be cumulative but experience significant diminishing returns.
When the cavalry does hit steadfast opposing infantry, it should pile up near the front ranks without bowling through the whole formation. All but the heaviest and best trained cavalry will actually stop just short and wheel away from the infantry. Then they can either stay put and try to poke them with spears or run away and try another charge. The latter has always been more common. If the enemy infantry do run, then the cavalry can mingle into their broken formation and cut men down at leisure.
Cavalry units charging each other historically resulted in the units passing through one another with some casualties on either side. Horses don’t stop on a dime, after all, and they also tend to avoid hitting each other head on.
Neither are horses perpetual motion machines; they cannot gallop for very long before becoming exhausted. Most cavalry charges canter or trot to within a hundred yards of the enemy and only then break into a gallop. Sometimes not even then; many successful charges have been delivered at the trot. Remember: it’s not about physical momentum, it’s about psychological momentum.
That’s about all I have to say. Honestly, I would be happy with just 1 and 2 being even partially implemented. I know that oftentimes, advocacy for increased historical accuracy in Total War games is perceived as obsessive-compulsive “button-counting” but the things I have laid out above are the basic physical and psychological mechanics of combat in the classical period, so I believe them important to a title about the rise of the Roman Empire. I also think that they would make for much more interesting gameplay without eliminating the wide appeal of the game; many of the changes are “under the hood.” If the developers include realistic morale and fatigue systems then I for one would care much less about the presence of war-dogs and battle-gladiators – which can always be modded out of the game anyway.