cavalry and morale mechanics make battles a cakewalk

[Achilles Lacedaemon]

Because of how easily exploitable morale mechanics and cavalry flanking is in this game and mod, battles are always a cakewalk. the ai cannot use cavalry and cannot flank with it, however, battles that do not involve any cavalry on both sides tend to be more challenging and engaging. One of the possible fixes to how easy it is to simply go "step one: have 3 units of cavalry, step two: use the 3 units to flank the enemy infantry, step three: make them rout, step four: win campaign" is to nerf all the cavalry in the game, or at the very least nerf the ability it has to flank and rout units, as well as how easy it is to engage and disengage.

[KAM 2150]

I am working on new experimental battle pack that will overhaul morale and cavalry among other. Now keep in mind that nerfing cav overall will make AI even weaker than now. On the other hand I constantly see AI cav atempting flank and rear manouvers. Defo cav needs to take a bit more casualties from disengaging than now, both for player an AI.

[Achilles Lacedaemon]

Not really, AI cav has a constant habit of charging straight into heavy infantry, not to mention how easy you can with 2 or 3 meele cavalry units to flank every unit of their armies, pursue archers and slingers, and overall demolish and rout them even when they have 1k plus men advantage.


perhaps lowering the cavalry speed would be a way? nerfing the player's ability to manipulate battle results with cavalry is def a problem though: most meele units get like 100 kills or less, whilst theres not a single battle a cavalry unit that i use to flank doesnt get 300-500 kills easily, not to mention the units that the cavalry just manages to rout without killing all soldiers whithin.


but as I said, battles with no cavalry in both sides are generally the most interesting and challenging ones in my experience, as the ai isnt utterly incompetent when it comes with foot units

[General_Rubenski]

I've noticed this too in my campaign as Rome. I've won almost every battle I've played due to my cavalry. With exceptions to siege battles or settlement battles where the AI doesnt try to sally out when they should just defend (issue i've been seeing. It says they're gonna defend but just rush out and get stomped by me lol)

I don't have a problem with it because ya I think cavalry should be able to rout units if attacked in their flanks or vs. light skirmishers. You just have to limit the amount of cav. in your army. I only have three. Two medium equites plus my general and that's just enough to let my melee units get a decent amount of kill and likewise for the cav. Mainly I use my cav to run done pesky missile units and wipe out routed units as well. I won't actually charge into the flanks or back of the AI when they're engaged with my melee units because 1.)most of them can handle and hold the line themselves 2.)I use (what I think are) OP sword units and velites to attack on the flanks and sandwich them in 3.) I hate getting my cavalry stuck in a unit to try to pull them out and get quite a number of them killed, especially if it's spears. Only if I face heavy units or units that I know will chew my line up, will I engage my cavalry and use repeated hammer and anvil strikes to whittle them down until they break.

Overall, cavalry gets tricky to modify because some people like it the way it is like me and others think its too OP or not strong enough. It will never please everyone. Best you can do is limit their effectiveness with ways you see fit. AI on the other hand? Well ya...AI will always be limited in their strategies unfortunately.

[Vardano]

Well, there are factions which are build all around cavalry, so nerfing them too much is not a solution for many aspects of the mod.

For me the problem of AI's use of cav are follows:

1. Timing. AI always rushes to flank from the start of the battle and Player is always there to counter it with javelins or spears. Can we teach AI to be more patient and hold its cav units before most of troops are engaged in a close fight?

2. Target to attack. Player always attack with cav the best units to attack (light units, not spears), AI will attack anything on its way, this is where it loses a lot of its troops.

I honestly think this can dramatically help AI with its use of a cav. If it also will have some awareness to disengage when it needs to, then it would be solid.

[KAM 2150]

Not really, AI cav has a constant habit of charging straight into heavy infantry, not to mention how easy you can with 2 or 3 meele cavalry units to flank every unit of their armies, pursue archers and slingers, and overall demolish and rout them even when they have 1k plus men advantage.


perhaps lowering the cavalry speed would be a way? nerfing the player's ability to manipulate battle results with cavalry is def a problem though: most meele units get like 100 kills or less, whilst theres not a single battle a cavalry unit that i use to flank doesnt get 300-500 kills easily, not to mention the units that the cavalry just manages to rout without killing all soldiers whithin.


but as I said, battles with no cavalry in both sides are generally the most interesting and challenging ones in my experience, as the ai isnt utterly incompetent when it comes with foot units

Does AI never flank for you? In all my test battles, only situation where AI cavalry charges forward is when majority of AI army is made from cav or only from cav, in which situation some units need to charge frontally to tie up player front. I just wonder what exact behaviour could cause your issue because I see a lot of activity from AI cav in my tests, it is constalty looking for flanks and rear, cycle charges etc.

Now in terms of cav kills, 300-500 is quite alright and I would consider it a bug if it would be otherwise. In other words, all units that take part in chasing down routing units will always have high amount of kills. This stirs from same misconception people had about pikes and hoplites that did little kills compared even to levies but they forgot to note that pike and hoplite units do not chase routing units. Normal units, when they attack routing units, trigger infinite charge bonus (thanks CA!) so they initially kill a lot of enemies due to their melee attack and melee damage being buffed by charge bonus. Same applies for cavalry, they auto trigger infinite charge so they will have the most kills of all units.

Lowering cavalry speed has own bugs, which are linked to infinite charge bonus for attacking moving units. Initially I wanted to have much slower down cavalry but it makes AI cav even weaker as now AI cav often pulls out of combat. Now this behaviour would be fine and all if AI would not try to pull own cavalry unit from 1v1 fight, even if it is winning. This means that as soon as AI cav unit tries to pull away, they use regular run speed (which if lowered, would make the issue even bigger) while unit that they were fighting gets infinite charge bonus until AI cav unit stops. With added charge bonus and AI cav not fighting back, your unit destroys it in seconds. This is why I have reduced cavalry charge speed so that now AI cav can pull away from combat. To get alltogether slower cav, I would also need to nerf cavalry charge speed, which would then be called by people immersion breaker and cause them to crap on DeI in turn.

Overall, cavalry gets tricky to modify because some people like it the way it is like me and others think its too OP or not strong enough. It will never please everyone. Best you can do is limit their effectiveness with ways you see fit. AI on the other hand? Well ya...AI will always be limited in their strategies unfortunately.

Fully agreed.

Well, there are factions which are build all around cavalry, so nerfing them too much is not a solution for many aspects of the mod.

For me the problem of AI's use of cav are follows:

1. Timing. AI always rushes to flank from the start of the battle and Player is always there to counter it with javelins or spears. Can we teach AI to be more patient and hold its cav units before most of troops are engaged in a close fight?

2. Target to attack. Player always attack with cav the best units to attack (light units, not spears), AI will attack anything on its way, this is where it loses a lot of its troops.

I honestly think this can dramatically help AI with its use of a cav. If it also will have some awareness to disengage when it needs to, then it would be solid.

1. That is issue all recent TW games have, it is very prominent in Three Kingdoms game. There in order to "fix it", CA just made cavalry run only a bit faster than infantry while their charge speed is between 2-3 times higher. The behaviour that caused is used by AI cavalry is that it always rushes own cavalry to player flanks BUT to fixed distance near players flank, ignoring player cavalry or second line. Due to them reducing cav speed in Three Kingdoms, cavalry seems to reach player line around the same time as AI infantry does BUT like I pointed up above, due to charge speed AI cav is just murdered by player cav as it is impossible to pull away from cav vs cav combat.

I have no idea why they just wont make AI wait for own infantry to engage player, which would work much better. Also even till this day, CA never added a command for AI to target player cav first.

2. AI has fixed priorities which I dont think can be influenced. I always see AI attacking player units on flanks (only first line) and player general, if AI has more cavalry units it also starts rear charging player units on flanks.



Now the good this is that I was finally able to pinpoint what caused AI units not to brace against cavalry and elephant charges. This will be included in next experimental battle pack, although sadly full fix for the issue would mean we are back to AI infantry doing frontal zombie rush.

[Vardano]

Kam, what you said about 3K is so sad really. They basically have to teach couple of commands to AI and they are not doing it for some reason. Very well modded Rome I AI cav does really intelligent stuff and here we go with what we have now.

I agree with guy in this thread who thinks that battles with no cav on both sides are harder for a Player.

[waidizss]

Man Kam why you keep saying 'in my testing'. You don't do tests bro just pump it out and wait for feedback

[Achilles Lacedaemon]

well I have some suggestions that wouldnt fix but at least would diminish the intensity of the problem:


1. lower some of the men count in cavalry units (an exception could be made to cavalry based factions, but it would take equity to determine when and when not to do this); it doesnt need to be much: in ultra settings, 130 men in a cavalry unit could be lowered to 100, 100 to 80, etc. a nerf of about 20-30 men in every cavalry unit could make it more fair (so far, cavalry can charge, engage and desingage without taking almost no casualties, it gets a free pass overall, but lowering the ammout of men would make the few casualties cavalry takes more significant).

2. increase cavalry's fatigue: if slowing cavalry down isnt an option, maybe making them tire quickly would be, it would make them perform slower after any initial engagement, which would contribute to prevent the reiterated flanking that one can do by moving the cavalry all around the map and flanking every unit one pleases and routing an entire army: basically right now the battlefield is the player's cavalry buffet.

3. make the unit spacing of cavalrymen tighter: this is a suggestion I'm dubious of but it could work: basically, making the cavalry unit cover less ground would make flanking attacks less all encompassing and devastating, it also could force the cavalry men to stay more time in meele (since they would be "more together") instead of what usually we see: half the unit in meele and the other half desingaging and already running miles away.

4. lowering cavalry's attack and increasing its meele defence: so that when the cavalry engages, it operates more like a foot unit rather than a battering ram that simply demolishes a unit, this would make it so that if cavalry wants to more effectively kill a unit, it would have to stay longer in the fight.

5. stopping units from routing when flanked by cavalry: many units can still hold on their own even if the cavalry flanks them, however the flanking and routing mechanics make it so that its easy to exploit situations where you can rout 60-70% of a good unit without it trying to stand its ground and deal some damage or resist.



Overall, as a side note, I'd also say that the whole of the morale mechanics should be reworked or outright removed, it never ever serves any purpose (outside of some very sad roleplaying some players do to compensate the lack of challenge and fun in the game) other than to harm the AI, make battles easier to the player or make it so the player can exploit routs to completely overwhelm the enemy. The few rare moments where the morale works against the player are in battles where the odds are already massively against the player from the start, or when the player fights with units so pitiful that upon meele they are bound to break.

I'd also point out that all of this isnt meant to make cavalry "as it should be", but merely to compensate for the AI's inability to exploit cavalry the same way the player can. If the AI could flank or clear the player's backlines of missile units with its own cavalry, then it would be devastating as the player is, but right now cavalry very rarely performs flanking, and when it does it sint done in a very intelligent way or in the most opportunistic manner (the ai gets stuck or outflanked, or it doesnt attack from the rear but mostly do frontal charges or, at best, side charges). And yes, the AI doesnt always do this but I've seen it more than once charging cavalry straight into heavy spearmen.


what to do with cavalry based factions is a discussion of its own, and I'd imagine many compromises could be made (like keeping their cav strong but weakening ther foot units even more, etc).


what the chap said above is true when it comes to taste, but i'm not aiming at reaching the most "immersive" or "historical" or superficially speaking "pleasing" result: even if everyone has a taste and playstyle that differenciates how they view cavalry should operate, it is undeniable to everyone that played this game for a substantial ammount that the AI is simply useless against the capacities the player has right now, and one of them is how the player can dictate battle results with something as simple as 3 cavalry units and non peasant level infantry (not difficult to deploy). The player can navigate the entire battle map with cavalry unopposed, it can selectively rear charge and rout almost every infantry, it can pressure the enemy backline and remove their skirmishers from the battle, amongst various other things (such as cavalry units being able to stand their ground and fight in meele with good decency).

one could even stretch and say flanking overall is a problem, but thats a fundamental mechanic of the game and it cant be changed nor should it, however its blatantly clear that flanking with cavalry has 1000x more the effective result than doing it with any infantry (perhaps theres an exception to be made to shock infantry that has high damage and speed, but i'm yet to see it). Even if some players like the current way cavalry is, I doubt they could argue in good faith and in knowledge that the current situation doesnt favor the player greatly and that the ai is incapable of using cav as the player can.



EDIT: this discussion is off-topic so lets not focus on it, but I'm sure KAM and the DeI team do their best, as one can logically infer from the results their mod achieved.

[nikossaiz]

I believe he is doing tests, partly because he wants to make observations and understand the way the engines behave, and to modded in his needs. That way he will upload one pack and not one every second day.

[KAM 2150]

I never add something that is not tested. How I originally made cav and morale was a bit different (less impactful) but due to large player feedback it was changed and gladly welcomed until recently. I would happily spend more time actually enjoying my lackluster free time than doing same test 10 or so times in a row.

1. lower some of the men count in cavalry units (an exception could be made to cavalry based factions, but it would take equity to determine when and when not to do this); it doesnt need to be much: in ultra settings, 130 men in a cavalry unit could be lowered to 100, 100 to 80, etc. a nerf of about 20-30 men in every cavalry unit could make it more fair (so far, cavalry can charge, engage and desingage without taking almost no casualties, it gets a free pass overall, but lowering the ammout of men would make the few casualties cavalry takes more significant).

2. increase cavalry's fatigue: if slowing cavalry down isnt an option, maybe making them tire quickly would be, it would make them perform slower after any initial engagement, which would contribute to prevent the reiterated flanking that one can do by moving the cavalry all around the map and flanking every unit one pleases and routing an entire army: basically right now the battlefield is the player's cavalry buffet.

3. make the unit spacing of cavalrymen tighter: this is a suggestion I'm dubious of but it could work: basically, making the cavalry unit cover less ground would make flanking attacks less all encompassing and devastating, it also could force the cavalry men to stay more time in meele (since they would be "more together") instead of what usually we see: half the unit in meele and the other half desingaging and already running miles away.

4. lowering cavalry's attack and increasing its meele defence: so that when the cavalry engages, it operates more like a foot unit rather than a battering ram that simply demolishes a unit, this would make it so that if cavalry wants to more effectively kill a unit, it would have to stay longer in the fight.

5. stopping units from routing when flanked by cavalry: many units can still hold on their own even if the cavalry flanks them, however the flanking and routing mechanics make it so that its easy to exploit situations where you can rout 60-70% of a good unit without it trying to stand its ground and deal some damage or resist.

1. That wont happen as it would not affect starting units on campaign map, editting which is not doable at the moment as entire campaign file would need to be rewritten.

2. I can bump a fit cavalry fatigue penalty for running, all other penalties are shared will all other unit types.

3. That would not do much as there are other values related to virtual (not graphical) model size. Most of them are already rather close but it could cause them to glitch if they are even closer due to models pushing eachother away.

4. Attack and defence work in a weird way that if unit has high defence, it means it will score more hits anyway. Damage would be better in this place.

5. That is doable but I expect there would be a backlash that flanking is useless and would would have similar type of thread, just the other way around.

[Achilles Lacedaemon]

I'm sorry but I think that someone complaining that their flanking "is useless" is much more of a pedantic issue of someone being annoyed that braindead playstyles arent rewarded... flanking will always have a tangible benefit, even if its just to have more of the enemy being killed because of more contact with your units.

Right now, this is the synopsis of this game:



the fact that fighting the ai is basically unrewarding, easy and a repetitive cakewalk is much more evident by the inumerable submods which attempt to increase game difficulty in all kinds of ways rather than submods that do the contrary.

all it takes is 1 cavalry, and no matter how elite their units are: they are going down.

Spoiler Alert, click show to read: 

in the battle above, 1 cavalry unit routed multiple seleucid hoplites and chalkaspides phalanxes (which routed with 248 out of 256 men). How absurdly exploitable morale is in this made a battle that should've been a loss an extremely easy victory, the enemy had like double my size during the start of the battle, and also they had elite units... which meant nothing. (btw, fyi the cavalry unit had a kill count of 800).

If you think this is acceptable I'm going to start looking for submods that remove some of the most cheesy aspects of morale mechanics, or try to do it myself.

[geala]

The problem with these "make AI and player the same" approaches is that it is not possible to do it with the restrictions of the game code and at the same time make unit to unit behaviors and effects at least slightly realistic. Moral played a big role on the battlefield, maybe the biggest role of all. Battles were usually won by the side which won the cavalry fight. Units attacked in the flank or back were really prone to break. And so on.

Removing this may make the combat more even between AI and player. In the end it would be a war of attrition to the few last men, the army with more or better units would win. from then on complain threads would start about how easy it is for the player to build better economy and better composed armies. Or how boring and long out drawn every battle is and how unhistorical unit behavior.

People who think battles are unfair and biased towards the player (which they are of course) should perhaps auto-resolve, which in my experience albeit having problems of it's own is more balanced towards AI and player. Because actually, if battles were mostly a grind event of infantry with stats of quality and quantity counting above all, with ineffective cavalry attacks and moral removed, why would you like to play it or watch it in campaign?

So, as much as I can understand the concerns and wishes for changed mechanics, I'm quite not sure it would be a better game then.

[charlieh]

I am inclined to agree with geala, and would not want to see the effectiveness of cavalry too far diminished. In my opinion, is it just as important to try and keep some semblance of realism and cavalry were critical in numerous ancient battles. Overall, I would propose continued small tweaks, not large changes.

[Antiochos VII Sidetes]

Please do NOT nerf cavalry. That is not the right approach. Cavalry taking more casualties when disengaging is why the TW Attila engine is so godawful and lame and arcadey. I think that would be a huge mistake -- it would just be better if the AI would use cavalry more effectively.

[Grabbin_Megroin]

I support the side of making cavalry excel at rear and flanking charges, but perhaps a bit less capable in stationary melee. I mean, that's how it worked realistically I think.

[Dumanthis]

One thing that boosts cavalry is the fixed depth of infantry formations, meaning that it is more difficult to cover flanks with infantry and thus making cavalry more able to just move to the sides and rear where they can do easily heavy damage to the morale.

I remember there has been versions of DeI that had infantry being almost like a concrete wall to charge with cavalry, making almost no affect at all. Not suprisingly, in this version cavalry was not in such a great role and light cavalry had almosts no value at all. IMO the current version has come a long way since the days of that and the days of 360 phanlanx spearmen defense boost

I think this is a more complex issue that in order to solve needs to concern the DeI balance and mechanics as a whole. I think that maybe spearmen and heavy infantry should be harder to kill with even heavier cavalry, making them more bulky blocks and more defended from charges. I would not remove the morale penalty from being surrounded or struck from the rear as it would make every unit like a spartan bodyguard (again ).

One thing that nerfed the cavalry a lot was the boost give to missile units, but at the moment I think missiles do a little bit too much damage to the regular infantry (maybe even up to 25% ~). Is it even possible to make infantry more defended against missiles than cavalry separately? Also javelins are great to kill cavalry, but they are prone to getting run down before they can even operate by cavalry.

Summa summarum: For now to nerf cavalry, in careful moderation, I would suggest:

1. allow infantry in more wider formations than before (but not to make ridiculous spaghetti lines, maybe 3 men in depth max),
2. increasing missile defense to infantry units (maybe 25%~, of course not for ranged and unarmored/shielded)
3. making spearmen and heavy infantry once again more "bulky" in general with heavier defence against cavalry

This would of course, have effects that might be undesirabe: Roman infatry would be more mega pain to kill, and phalanxes would become again far more valued units not only in defending choke points. There is not an easy way to effect a singe part of the mod battle balance without effecting all other parts of it at least idirectly.

[Achilles Lacedaemon]

1. we cant mod the AI

@Antiochos VII Sidetes
2. you only say its god awful and a huge mistake without basing that assertion on anything; I can (as I did above) provide all kinds of proof as to how cavalry flanking is so easy to exploit and perform, and so insanely powerful, that it can dictate the result of battles that should be difficult; IMO, flanking all kinds of enemy elite units by just having the brainpower to command cavalry to move around and attack from the rear, and then win every battle against any kind of army no matter its tier (requiring, at best, 3 meele cavalry units to win, despite the fact that sometimes only 1 is needed) is more than enough for me to conclude that this kind of cheese is god awful, it makes the player invincible.


so far all the arguments here acknowledge that the player can exploit this, they are either in favor of some kind of nerf or tweak, or against it due to "realism".


what I would point out against the latter is that this game is anything but realistic in its most basic manifestations; missile units can fire from gun ranges, without seeing the enemy directly, and have a precision better than if they had snipers, with a fire rate and coordination that certainly wouldnt be possible with 3rd century BC weaponry; no unit has any semblance of collision, making disciplined formations pointless and only disadvantegeous because it makes the space the unit covers smaller and so it makes it more flankeable, collision and pushing matches would be extremely tactically important in any kind of battle of this kind as we even know some of the greatest battles of history only were possible due to a general's using his troops capacity to fall back or push and exploiting some form of collision (see marathon, cannae, and many others); the player (and even the AI, for this matter) has complete omnipotent control of his armies and units: in TW, armies and units behave like hive-minds instead of orginzed humans, you can order units to permorm all kinds of manouvres in runnings speed that would be impossible or extremely difficult for most human groups, making our soldiers look more like a group of insects obeying a hive-mind (not to mention that everyone in the battle is omniscient because if something bad happens in your left flank, units that are not even realistically seeing that can immediately respond, by running around and reinforcing with such inhuman coordination and speed), and I could go on ad nauseum on small points that most people (except those who have a hollywoodian conception of realistic) would have to acknowledge and understand that they simply cant have a realistic experience in this game, or aim to have one, its not possible, and therefore its not a good priority to have when it comes to mechanics.

I hate to pull this one out here but "realistic" around here is often used in a completely subjective way: some people say its realistic for cavalry to completely rout the enemy by flanking and having a strong charge impact, others would say its unrealistic because horses wouldnt simply throw themselves en masse into a blob of meele combat, the animal having a behaviour of instinct and to at last a marginal level self-preservation. I'd respect more this point if it was done conveying what I think are the true purpose of this arguing: some people think its more "fun" to have super impact effective cavalry, others dont.

[meerkatology]

on VH money is a problem at least pre imperium 5. cav is generally pretty expensive to have lots of it for the player. they are also mostly useless in sieges. they also dont benefit from the general/army tradition buffs since u dont add cav ones unless ur a cav faction.

so for infantry factions - aka every one that isn't eastern/african, i rarely have more than 1-2 cav. becasue when u stack infantry bonuses and use them well they are more efficient than cav, i find.

i find that the AI generally will have something behidn their line until late in the battle, its not easy to just free flank their entire army with cav.

as far as phalanxs getting routed easily by flanking cav.... they get routed easily by flanking infantry as well. getting flanked IS their main weakness.

i dont think cav is op , even for the player. but with that said, i think morale/combat penalties from getting flanked is too much for nonphalanx units.
imo THAT is what needs to change, because routing/beating things with flanking infantry works just as good as cav... after all elite units shouldn't give too much of a crap about being flanked by peasants.

[Ptolemaios Soter]

Achillies, I don’t get your point about realism. You are of course correct no every way that you point out that we are in fact sitting at a computer playing a a game and not actually on the field of battle 2000 years ago. But the fact that this is a game and is subject to game limitations doesn’t mean that realism isn’t relevant. Realism isn’t an on/off switch. Rome could have machine guns as a factional unit. That would be less realistic than having highly accurate ballista able to strike quickly moving targets (aka what we have in game), which would in turn be less realistic than ballista that was only accurate against stationary targets (which is it possible given the game engine). So saying that we can’t have he last option means why not go with machine guns doesn’t make sense.

I play this game in large part for the historical immersion. I know many other players do as well. Getting it as realistic as possible within the game limits and without breaking balance/playability is a good goal - and one of the prime motives of this mod vs vanilla.

On the subject of cavalry specifically, I'm copying a post from a thread a while back:
---
The role of light cavalry was to harass the enemy flanks more to prevent outflanking maneuvers than to actually kill many troops. These cavalry would have no armor for rider or mount but would often have a shield, with the exception of eastern horse archers. Additional equipment such as spear, shield, and additional javelins might be strapped to the horse or kept back at camp and swapped mid battle. Light cavalry would often engage with other light cavalry, but rarely in melee with enemy heavy cavalry and almost never with infantry. These horses were not generally trained to charge, and so simply would not be able to do so against massed infantry. They were also used as scouts and chased down routers. Parthian horse archers specifically were used against Rome to try to disrupt the enemy formations to provide a target for the cataphracts.


Heavy cavalry really just meant anything trained/equipped/armored enough to engage enemy infantry. These riders often worse some armor and carried a lance in addition to spear or sword and shield. As above, this may have been strapped to the horse and/or left at camp/followers. Lances usually broke on the first charge. The literal weight of the cavalry is less relevant as it wasn’t so much a matter of physically bowling into the enemy infantry (which would have usually killed the horse by breaking its legs) as to cause the enemy infantry to break at the threat of the impending charge. Wedge shaped formations were designed to provide an easy way to maneuver to the side to abort a charge last minute if needed. The role of heavy cavalry was first to engage and defeat the enemy cavalry, and then to charge into the flanks of engaged infantry. Some eastern and steppe armies used shock cavalry with bows in addition to their lance to engage at a distance as well.

Cataphracts are a slight exception/caveat to the two classes because they were definitely noted differently in the ancient sources. Cataphract just means fully covered (with armor) and referred to the rider and horse. Often, in reality that meant only the front half of the horse. Cataphract armor was meant more for protection from arrows than in melee, though obviously it helped there as well. In contrast to how the game deals with armor penetration, it was more often simply yes or no. Arrows could usually not penetrate cataphract armor and so the cavalry were near invulnerable to missile fire. Interestingly, lead slinger bullets could, which is one of the main reasons Rome was able to defeat the Seleucids. Because of this, cataphracts generally didn’t use shields.
---

I think anyone arguing that (non-cataphract) cavalry should be able to charge dense infantry formations isn't wrong so much as they're just looking at the wrong timeframe. Horses were way smaller 2,300 years ago than they are today, and they were less specialized in breeding for war as well. By the high middle ages, cavalry could run over peasants and could often charge infantry formations as well. That simple wasn't the case in antiquity. Horses were primarily used to better position the rider in the battlefield (behind the enemy for example), rather than as an animal weapon itself. Even looking at those later examples though, the primary weapon of the horse was still psychological rather than physical. From the Winged Hussars of Poland to the Rohirrim of Lord of the Rings (sorry, I couldn't resist), a cavalry charge was dependent on the enemy (or orcs) breaking ranks in fear. Only then with their weapons turned or lowered could the cavalry bodily trample the opposing forces without suffering higher casualties from the impact.