Melee cavalry performance in 1.6.0

[Dabos37]

I am really curious about cavalry performance in 1.6.0 and if melee cavalry units feel better now. I have read something about cavalry AI getting improvements against other cav units which sounds interesting.

 

[Ananda_The_Destroyer]

Too early for me to tell, interested in what others report. Enemy Cav seems a little more likely to hit me when I attack them with my character, but so far it's hard to observe a better kill amount when they attack, but tbf I only charge them in busts and often in SW.

 

[Aurex]

I am really curious about cavalry performance in 1.6.0 and if melee cavalry units feel better now. I have read something about cavalry AI getting improvements against other cav units which sounds interesting.

I've tried out a little bit of custom battles (alright, a dozen or so, I was curious myself) and these aren't judgements on how it performed, just my feelings on how it reacts now compared to before in the same scenarions and maps. I didn't actually partake myself, I just ran around and watched without "aggroing" enemies and acting as a sergeant so I wouldn't interfere with the AI. Also, empty plain map to give cavalry an edge, which was the point.

I tried (before and after) a typical 250 vs 500 custom battle, Vlandia versus Khuzait, mainly to see how the heavy vlandian cavalry fares against horse archers, and another 250 vs 500 with Vlandia versus Sturgia and Battania, to see how the cavalry deals with massed infantry/archer support and how the enemy uses formations. After this I replayed the same battles but in a 500 vs 500 setup.
The setup was the same for 1.5.10 and for 1.6.0, I didn't use the new feature to customize the troops on the field as not to alter the balance.

Pre-1.6.0 results below.

Vlandia-Khuzait, 250 vs 500: Long-ish battle, eventual loss for Vlandia. AI performed borderline decently, often pushing the horse archers away and pursuing them with part of the troops while the rest engaged in sparse order. Superior number and HA meant clear defeat.

Vlandia-Khuzait, 500 vs 500: Rather short battle all things considered, khuzait melee cavalry defeated while their HA picked off the vlandians as they pursued. Khuzait melee cavalry suddenly routs after less than a minute of engagement, leaving all the vlandians against a handful of HA. Decisive Vlandian victory.

Vlandia-Sturgia, 250 vs 500: Rather long battle. Vlandia charges the front line (archers which kept staying there for no reason) while the sturgian infantry attacks with some delay. Eventually the infantry forms small clusters which are unassailable by the cavalry, Vlandia routs. Pyrrhic victory for Sturgia, given the losses and the numbers in the field.

Vlandia-Sturgia, 500 vs 500: Medium length battle. Sturgian infantry fails to protect archers yet again, archers rout, superior cavalry numbers eventually whittle down resistance in a fierce fight. Pyrrhic victory for Vlandia, since too many cavalrymen died while attempting to fight infantry as if they were also on foot, not using mobility at all and often attacking from the front whereas they could have attacked from behind or from the flanks.

Vlandia-Battania, 250 vs 500: Short battle. Cavalry annihilated by archers and massed infantry. Replayed it multiple time to verify it wasn't a fluke. Utter vlandian defeat, the AI charged as one - only to get bogged down in the usual melee fight, with the occasional lucky cavalryman able to do some slight tapping of shields before being brought down.

Vlandia-Battania, 500 vs 500: Long battle. Cavalry badly battered by archer fire. Cavalry eventually manages to rout the archers, a long melee skirmish ensues. Vlandia emerges as the victor, but with just 24 survivors. Pyrrhic vlandian victory. Same AI problems as the previous battles.

1.6.0 results below:

Vlandia-Khuzait, 250 vs 500: Short, brutal battle. Cavalry fails to pursue the HA at all but gets distracted by them, so it ends up slaughtered by the enemy melee cavalry as they reposition. Total vlandian defeat. The HA are shown to be far less deadly than before, with less accuracy and more willingness to engage in melee.

Vlandia-Khuzait, 500 vs 500: Very long battle. Vlandia doesn't pursue the HA yet again, but manages to defeat the enemy melee cavalry. Eventually the vlandian manage to rout the HA as well, so the battle ends in a victory for them. Only 11 survivors.

Vlandia-Sturgia, 250 vs 500: Short battle. Cavalry ignores the fact the enemy is in a melee circle and charges right into it. Routs shortly thereafter with absurdly high losses and only 53 enemy losses. Total vlandian defeat.

Vlandia-Sturgia, 500 vs 500: Short battle. Again the cavalry charges the melee circle and performs pathetically. Typical pattern of cavalrymen "charging" with no penetration whatsoever, and the few who got inside the circles were annihilated in short order. The stragglers tapped the shields of the infantry, nothing more. Terrifying loses to enemy fire in the first charge, too. Total vlandian defeat.

Vlandia-Battania, 250 vs 500: Short battle. Same pattern as with Sturgia: useless charge against massed infantry, cavalrymen destroyed as they stand still against the shieldwall (circled shieldwall, I suppose I could call it?) Almost no enemy losses, total vlandian defeat.

Vlandia-Battania, 500 vs 500: Medium length battle. Another futile charge, enormous losses inflicted by battanian infantry (the enemy archers were basically unable to act and shoot at all from inside the circle, but did melee nicely whenever a straggler managed to get into the circle formation.) 308 deaths and 191 routed vlandians, 32 dead and 11 routed battanians. Absolute disaster for Vlandia.

Conclusions: this wasn't an accurate test at all, I realize that. But it's how I killed some spare time earlier today. From what I've seen the cavalry performs worse against massed infantry than before, but that might be because now the infantry is far more quick to get in a circle formation. The HA are definitely less effective, but so are the melee cavalry. The pathfinding seems to have gotten worse, while being a little more reactive: the formations tend to respond quicker to threats in 1.6.0 but they do so in a scarcely effective way. There's no indication any significant progress has been made, aside from a little more reactivity.

P.S=It's getting less and less Mount and Blade and more and more Dismount and Shieldwall.

 

[ArturMorhghain]

I've tried out a little bit of custom battles (alright, a dozen or so, I was curious myself) and these aren't judgements on how it performed, just my feelings on how it reacts now compared to before in the same scenarions and maps. I didn't actually partake myself, I just ran around and watched without "aggroing" enemies and acting as a sergeant so I wouldn't interfere with the AI. Also, empty plain map to give cavalry an edge, which was the point.

I tried (before and after) a typical 250 vs 500 custom battle, Vlandia versus Khuzait, mainly to see how the heavy vlandian cavalry fares against horse archers, and another 250 vs 500 with Vlandia versus Sturgia and Battania, to see how the cavalry deals with massed infantry/archer support and how the enemy uses formations. After this I replayed the same battles but in a 500 vs 500 setup.
The setup was the same for 1.5.10 and for 1.6.0, I didn't use the new feature to customize the troops on the field as not to alter the balance.

Pre-1.6.0 results below.

Vlandia-Khuzait, 250 vs 500: Long-ish battle, eventual loss for Vlandia. AI performed borderline decently, often pushing the horse archers away and pursuing them with part of the troops while the rest engaged in sparse order. Superior number and HA meant clear defeat.

Vlandia-Khuzait, 500 vs 500: Rather short battle all things considered, khuzait melee cavalry defeated while their HA picked off the vlandians as they pursued. Khuzait melee cavalry suddenly routs after less than a minute of engagement, leaving all the vlandians against a handful of HA. Decisive Vlandian victory.

Vlandia-Sturgia, 250 vs 500: Rather long battle. Vlandia charges the front line (archers which kept staying there for no reason) while the sturgian infantry attacks with some delay. Eventually the infantry forms small clusters which are unassailable by the cavalry, Vlandia routs. Pyrrhic victory for Sturgia, given the losses and the numbers in the field.

Vlandia-Sturgia, 500 vs 500: Medium length battle. Sturgian infantry fails to protect archers yet again, archers rout, superior cavalry numbers eventually whittle down resistance in a fierce fight. Pyrrhic victory for Vlandia, since too many cavalrymen died while attempting to fight infantry as if they were also on foot, not using mobility at all and often attacking from the front whereas they could have attacked from behind or from the flanks.

Vlandia-Battania, 250 vs 500: Short battle. Cavalry annihilated by archers and massed infantry. Replayed it multiple time to verify it wasn't a fluke. Utter vlandian defeat, the AI charged as one - only to get bogged down in the usual melee fight, with the occasional lucky cavalryman able to do some slight tapping of shields before being brought down.

Vlandia-Battania, 500 vs 500: Long battle. Cavalry badly battered by archer fire. Cavalry eventually manages to rout the archers, a long melee skirmish ensues. Vlandia emerges as the victor, but with just 24 survivors. Pyrrhic vlandian victory. Same AI problems as the previous battles.

1.6.0 results below:

Vlandia-Khuzait, 250 vs 500: Short, brutal battle. Cavalry fails to pursue the HA at all but gets distracted by them, so it ends up slaughtered by the enemy melee cavalry as they reposition. Total vlandian defeat. The HA are shown to be far less deadly than before, with less accuracy and more willingness to engage in melee.

Vlandia-Khuzait, 500 vs 500: Very long battle. Vlandia doesn't pursue the HA yet again, but manages to defeat the enemy melee cavalry. Eventually the vlandian manage to rout the HA as well, so the battle ends in a victory for them. Only 11 survivors.

Vlandia-Sturgia, 250 vs 500: Short battle. Cavalry ignores the fact the enemy is in a melee circle and charges right into it. Routs shortly thereafter with absurdly high losses and only 53 enemy losses. Total vlandian defeat.

Vlandia-Sturgia, 500 vs 500: Short battle. Again the cavalry charges the melee circle and performs pathetically. Typical pattern of cavalrymen "charging" with no penetration whatsoever, and the few who got inside the circles were annihilated in short order. The stragglers tapped the shields of the infantry, nothing more. Terrifying loses to enemy fire in the first charge, too. Total vlandian defeat.

Vlandia-Battania, 250 vs 500: Short battle. Same pattern as with Sturgia: useless charge against massed infantry, cavalrymen destroyed as they stand still against the shieldwall (circled shieldwall, I suppose I could call it?) Almost no enemy losses, total vlandian defeat.

Vlandia-Battania, 500 vs 500: Medium length battle. Another futile charge, enormous losses inflicted by battanian infantry (the enemy archers were basically unable to act and shoot at all from inside the circle, but did melee nicely whenever a straggler managed to get into the circle formation.) 308 deaths and 191 routed vlandians, 32 dead and 11 routed battanians. Absolute disaster for Vlandia.

Conclusions: this wasn't an accurate test at all, I realize that. But it's how I killed some spare time earlier today. From what I've seen the cavalry performs worse against massed infantry than before, but that might be because now the infantry is far more quick to get in a circle formation. The HA are definitely less effective, but so are the melee cavalry. The pathfinding seems to have gotten worse, while being a little more reactive: the formations tend to respond quicker to threats in 1.6.0 but they do so in a scarcely effective way. There's no indication any significant progress has been made, aside from a little more reactivity.

P.S=It's getting less and less Mount and Blade and more and more Dismount and Shieldwall.

Aren't the damage values different for “custom battle” than they are for single player? I recall seeing this somewhere.

 

[Aurex]

Aren't the damage values different for “custom battle” than they are for single player? I recall seeing this somewhere.

I... genuinely have no idea about that. I think the AI is the same, though, and that's what I was curious to see, since I've seen the patch notes and some comments mentioning it had gotten better, especially for cavalry. Always felt cavalry was absolutely ineffective in Bannerlord.

 

[ArturMorhghain]

I... genuinely have no idea about that. I think the AI is the same, though, and that's what I was curious to see, since I've seen the patch notes and some comments mentioning it had gotten better, especially for cavalry. Always felt cavalry was absolutely ineffective in Bannerlord.

Yeah I’m sure AI would stay the same. And yeah I’ve felt the same in single player (although I’ve not yet gotten any cav in my most recent 1.6 run). In tournaments, anecdotally, their decision making feels a bit better? But that doesn’t mean much if melee cav are still ineffectual against archers up close.

 

[Aurex]

Yeah I’m sure AI would stay the same. And yeah I’ve felt the same in single player (although I’ve not yet gotten any cav in my most recent 1.6 run). In tournaments, anecdotally, their decision making feels a bit better? But that doesn’t mean much if melee cav are still ineffectual against archers up close.

I haven't tried anything outside of custom battles yet, sorry, didn't have enough time :/
From what I've seen, it's overall a little worse now than it was in 1.5.9 and 1.5.10. But it might be just me, or the particular battles I tried.

 

[ArturMorhghain]

I haven't tried anything outside of custom battles yet, sorry, didn't have enough time :/
From what I've seen, it's overall a little worse now than it was in 1.5.9 and 1.5.10. But it might be just me, or the particular battles I tried.

I’d believe your testing, definitely!

And I found the post detailing the weird quirk with custom battle damage not lining up to single player. Appropriately enough, in a thread about melee cavalry:

There is a thread discussed about this issue before: https://forums.taleworlds.com/index.php?threads/fix-cavalry-charge.442262/

As we know, developers changed reorganizing range of cavalry in Beta e1.5.10, but it affects few.

Every reason you can found in codes.

For example, the Mount Charge Damage:
Do you found that mount charge damage in Singleplayer SandBox is lower than Multiplayer and Custom Battle?

Let's see some codes before,
1. In SandBox.SandboxAgentStatCalculateModel.UpdateHorseStats(Agent agent, AgentDrivenProperties agentDrivenProperties):void

agentDrivenProperties.MountChargeDamage = (float) mountElement.GetModifiedMountCharge(in equipmentElement2) * 0.004f;

2. In TaleWorlds.MountAndBlade.CustomBattleAgentStatCalculateModel.InitializeAgentStats(Agent agent, Equipment spawnEquipment, AgentDrivenProperties agentDrivenProperties, AgentBuildData agentBuildData):void

agentDrivenProperties.MountChargeDamage = (float) mountElement.GetModifiedMountCharge(in equipmentElement2) * 0.01f;

3. In TaleWorlds.MountAndBlade.MultiplayerAgentStatCalculateModel.InitializeHorseAgentStats(Agent agent, Equipment spawnEquipment, AgentDrivenProperties agentDrivenProperties):void

agentDrivenProperties.MountChargeDamage = (float) mountElement.GetModifiedMountCharge(in equipmentElement2) * 0.01f;

These three lines of codes are about mount charge damage calculation, the first is about Singleplayer SandBox, the second is about Custom Battle and the third is about Multiplayer.

We can find the key point after multiplication sign, it's no need to know what "f" is about, you just need to focus on the number.
In Singleplayer, the factor number is 0.004; but in Multiplayer and Custom Battle, it's 0.01.
You can understand these numbers by following this:
Mount Charge Damage = Base Damage * Factor Number
So,
In Singleplayer, the mount charge damage equals base damage multiply 0.004 and in another two modes, it's base damage multiply 0.01.
For example, the mount charge damage in Multiplayer is 18, it will be 7.2 in Singleplayer, The reason for all this is the numerical difference.

And about another issue "cavalry attack miss and don't wanna attack" is also about some factor value setting like this above.
Cavalry cannot deal enough damage as it should.


The developers focus on the behaviors simulation and don't care the actual experience caused by value of number.
Edit: I don't mean that the developers should not focus on simulation, I just means that they can improve observably something by changing value of number like this above, it doesn't matter if simulation looks like a little foolish.

Thanks to @lzh for providing information.

 

[Antaeus]

So far in 1.6, I'm finding that cavalry still need to be managed. When managed correctly, heavy cavalry can be devastating, but they are still prone to fragmenting, even when in a tight formation. never F6 them, and I never deploy them without keeping an eye on them to manually regroup shortly after a charge.

I have sat with glee as I have successfully directed my heavy cataphracts to charge into the flank of an enemy shield wall, and been delighted at the wall of green coming up on screen, but in 1.6 you still have to run them through an enemy formation straight after contact to stop them fragmenting and being killed off in detail - they're still going to go chasing off after random targets rather than keeping close to their mates.

I'm yet to encounter more than a handful of horse archers, at once, and I tend to kill off fewer than 5-10 horse archers myself before a battle starts, so I can't comment on their performance against them.

The biggest issue I had with cavalry was the fragmentation problem - which leads to the requirement to micro them. But I don't mind this, I deploy my cavalry intentionally and when they're going to be most effective - which means they're standing back for large periods in a battle - so perhaps this is appropriately encouraging you into being a good general.

 

[CalarTheMerciless]

Cav units seem a little better at attacking infantry, but they could probably still use a little bit of help IMO as they are meant to be the elites and they hit less often than a peasant

 

[Ni]

Like others reported, I've found they tend to wander off and fight in small groups or individually instead of staying together. If the enemy has a lot of archers I usually tell the cavalry to follow me, then loop around the infantry, charge into the archers and tell the cavalry to charge. They will kill a few archers then run off in pursuit of infantry that are fleeing because of archer fire and no danger to anyone.

As far as enemy AI cavalry ... sad to say I often feel perfectly safe ignoring them and focusing on the main body of the enemy army. In Warband I had a rule when enemy melee cavalry was on the field "Never stop moving". I would never stop moving, I was constantly (almost frantically) panning the camera around to see if any were on me, because I knew if I stopped for an instant there was a good chance I'd get a couched lance to the face and be one shot. In over 400 hours of Bannerlord I have yet to be taken out by cavalry that I can remember. Maybe that's luck, but I have been completely stopped while ordering troops around and watched enemy cavalry ride past me while missing their attacks.

 

[mujadaddy]

As far as enemy AI cavalry ... sad to say I often feel perfectly safe ignoring them and focusing on the main body of the enemy army. In Warband I had a rule...

Same here: I've been couch-lanced probably 5 times total in Bannerlord, whereas in Warband it was almost every other fight!

One thing I've noticed in 1.5.10 is that I can find these giant blobs of HA or Cav and just ride into/alongside them and then go nuts swinging polearm into them, and they barely change their behavior. Horsemen NEVER deal with this problem, only sometimes do the HA's drift back to their lines where the infantry/archers start poking me. Only if I were to stop attacking would the melee cav even put up any effort.

It's... disappointing, and it doesn't really seem like anyone is calling it out.

 

[KingEroc1st]

The biggest issue I had with cavalry was the fragmentation problem - which leads to the requirement to micro them. But I don't mind this, I deploy my cavalry intentionally and when they're going to be most effective - which means they're standing back for large periods in a battle - so perhaps this is appropriately encouraging you into being a good general.

i learned to never charge my cavalry. turn them into long lines and i just move them across the enemy formation that i want them to charge. this way the formation is most cohesive. after they pass i just do a face direction commend so they turn around and i do another movement command to sweep across again.

 

[Bucktoothbandit]

I’ve been messing around with Vlandia and battania on sandbox since they always fight each other now instead of vlandia vs Skyrim. Interestingly they feel evenly matched. Both sides field a lot of Calvary. Majority of the AI victories come down to who wins the Calvary battle, but then again I am using Realsitic AI mod to avoid the circle formation death tactic. Battanian cav is pretty evenly matched against vlandia vanguard and scout, and vlandia nobles have a strong presence. Really depends on the terrain too.

Calvary actively goes for archers and I’ve been apart of huge cav vs cav battles where it’s usually around 100vs100 divided on both flanks. It’s been insanely fun and I recommend it. I play with limited mods that help fix AI and spears. Very vanilla with original troop trees. Hope this help sorry if me sound like mountain bandit me marry rock live in rock castle

 

[Apocal]

i learned to never charge my cavalry. turn them into long lines and i just move them across the enemy formation that i want them to charge. this way the formation is most cohesive. after they pass i just do a face direction commend so they turn around and i do another movement command to sweep across again.

 

[Pejot]

I can recommend Realistic Battle mod to You it fixes a lot of problems and Cavalry is deadly if properly used (not micromanagement like vanilla). They are not op cause mindless charging will kill them but if positioned properly You can just order them to charge and they will run through the enemy and turn around after making some distance. They also use couched lances and khuzait lancers and cataphracts took me down a lot when I didn't observe the surroundings properly.


As for the topic. Playing vanilla I didn't saw much difference in behaviour. They die easily cause I think inf is better at using polearms in 1.6.

 

[KingEroc1st]

100% accurate