Im doing a medieval empire building rts with abstracted battles. Armies meet and battles are auto-simulated (similar to auto-resolve in total war or battles in europa universalis).

These unit types exists:

1. light infantry

2. heavy infantry

3. light cavalry

4. heavy cavalry

5. archers

They contribute to the army strength depending on the terrain type (attack settlement, defend settlement, open terrain, though terrain). Archers looses a lot of strength if too big portion of the army (needs to be backed up by melee soldiers).

Problem is cavalry. They seem a bit redundant at the moment. They are weaker in settlement defence and better in open terrain than infantry. Should they maybe also mess up enemy archers?

Or just stronger than infantry, but loose strength if to big portion of the army? I dont want the player to juggle army composition percentages too much...

Or combine the 4 first unit types into just 2 : troops and armoured troops.

Plz brainstorm and/or comment :)

I think the cavalry definitely should mess up archers if they don't have enough protection from light infantry with spears or terrain

Spearguys is what counters cavalry.

Cavalry counters archers.

And archers shoot the spear guys from afar.

Then the heavy infantry comes in chopping and cleans up the mess :) (and is usually better protected from arrows then light infantry)

All infantry is vulnerable from cavalary charges from the flanks or behind. If that's possible depends on terrain and relative composition I guess....

Just some random unstructured thoughts :)

I think Total War way is quite straight forward and logical. (If we assume light infantry is spearmen and heavy infantry is swordmen)

Light cavalry is good against archers and terrible against light infantry but have speed to counterbalance lack of armor.

Heavy cavalry is good against heavy infantry and archers but mediocre against light infantry

well there is no positioning or movement, armies just auto-simulate a result :)

(you can follow the battle in realtime, the numbers of men remaining in each unit cathegory tick down)

light/heavy inf is not "spear" and "non-spear". Its more related to tech and amount of armour. The game is very high-level (abstracted). The rock-paper-scissors triangle of spear-cav-archer doesnt quite work.

If cav indeed should be anti-archer i need a good, intuitiv mechanic for it. Maybe each company of light cav could half the strength of a company of enemy archers? So flooding your army with light cav isnt worth it since opponent will not flood his armies with archers?

And make heavy cav ("knights") less anti-archer and just generally very strong and expensive... (maybe half the anti-archer effect of light cav?)

>> Plz brainstorm and/or comment :)

when you struggle like this it usually means your model is flawed - usually too simple.

there are a number of factors you do not mention such as army size, training, moral, or random chance.

for two armies of the same size and composition, victory comes down to morale, training, and a roll of the dice.

you will need a system where the results are believable for any combo of size and composition for the two forces.

you might try just assigning a value to every type of unit and adding them all up. but i'm not so sure that would work.

you might try classifying armies based on composition, then have a lookup table with mods for one composition type vs another. so an army that was mostly knights would dominate again an army of mostly archers (assuming no mud and chokepoints like at Agincourt).

https://en.wikipedia.org/wiki/Battle_of_Agincourt

you may actually have to do some rather in-depth modeling behind the scenes to make those numbers count down in a believable manner.

and make sure the user can skip the countdown and go straight to the results.

and if you are not a wargamer familiar with medieval army combat, have a wargamer who is knowledgeable about medieval army combat test the game for you to make sure its believable. you may even want to ask them for advise in coming up with a believable combat model.

by and large the purpose of cavalry in combat (as opposed to scouting) has remain unchained since the time of Alexander the Great and his Companion cavalry - fast shock troops to be applied at a place and moment of weakness to turn the tide - and mopping up operations against a broken foe. this could even be seen in large scale tank offensives in WW1, when the "heavy cavalry" shock troops rolled along at a whopping 4 mph.

Yeah, I was just talking in general how you'd expect it to work, how to simulate it is a different problem :)

Even if the display is just numbers ticking down, maybe there could be some "placement" and "positioning" under the hood, but in a very abstract manner?

Something like units from the different sides "pair up", with a priority on what type of unit they want to fight.

So light cav have a buff against archer units, so they first start to attack an archer unit from the other side.

On the other side, his spear infantry prioritize defending against light cav, so they start generating attacks on the cav, which generates attacks on the archers.

archers are blocked from generating attacks on anyone as long as engaged with cav, but can defend (poorly)

All this can just be buffs/debuffs on attack, and a choice on who the attack is on.

If one army outnumbers the other, there will be more the one unit from one side generating attacks on the other, making them more likely too win.

But the real winner in actual war is local superiority, even a small army can beat a big one if they fight them in the right way. This could be partly simulated in how easy a unit can "switch" between enemy units, maybe just a time delay simulating that they have to move across the field.

All of it could be just numbers behind the scene, but I think you need to simulate this in some way for the confrontation to feel realistic. (unless you just reduce it to Civ style combat)

Btw, I definitely think you should have spear troops. Spear guys have been vital for warfare since ancient times, right up until modern times and gunfire squads. (which also had bayonetts, so they could be a bit more like spear guys) You need guys with big pointy sticks to keep the enemy at a distance, and particularly to defend against cavalry charges.

Even if you are doing a straight up simulation of the battles to 'auto-complete' them with no user input, that doesn't mean you can't still have the battle be 'played out' with relative positioning.

You could also plan the system in such a way that you could produce little videos of the battle, showing the different companies/regiments moving about on the battle field with arrows and such like a historical documentary. The little rectangles with Xs or Slashes through them sliding around with morale and health bars shifting so that the player can see what happened.

From that you could even expand a system to generate "Dispatches". "Such and such Company of the So and So's valiantly turned the right flank, and opened a gap for The Other So and So's to charge through and force a general retreat".

And then from that you can generate special battle honours. Now your units of Cavalry aren't just "yet another unit of horsemen", but that one unit is the unit that turned the cannon at that one battle in the last war, a pivotal moment in the war, and now have impressive buffs and shiny medals. And I can click on them to see WHY they have the medals.

That level of complexity won't work and there will not be distinct companies as in total war. Several battles can occur at the same time as the player manages his cities. The genre is more similar to crusader kings or europa universalis.

Im thinking to cover several of the foremensioned mechanics in a "flanking" rating (%) for the entire army (there is already a morale rating for the entire army which falls with losses, fatigue, etc).

Your flanking rating

... is defined by proportion of light cav (x3), heavy cav (x2) and light inf (x1).

... is increased by enemy archers

... is reduced by enemy flanking rating or (severely) when attacking settlements.

The flanking % increases the casaulties the opponents army suffer each combat day.

This would make players want to include some cav to not be too exposed to enemy flanking and archers (but they will not be cost-effective to fill the entire army; diminishing returns on the flanking %)

This makes cav naturally weaker when sieging fortifications, and gives light inf some "mobility" compared to heavy inf (who will now be even more focused on brute force).

This could work. Comments? Very abstracted I know but I need something along these lines:)

You do realize that EU and CK simulate battles in a scaled back version of what I was talking about, right? The point of the simulation is to make the outcomes more natural rather than just an overly simple dice roll, and be able to design a system that the player can look at and get more feed back about why something worked or didn't work.

Also plays into those random off chance wins better than a straight up dice roll. You didn't win or lose just because the dice were good/bad, you won/loss because "X happened" during the battle simulation.

Is movement visualised in CK? I remembered only the status page:

https://www.google.se/search?q=battle+crusader+kings+2&espv=2&biw=1920&bih=955&source=lnms&tbm=isch&sa=X&ved=0ahUKEwj57f-CiPvMAhXKDiwKHSiAB94Q_AUIBigB#imgrc=oWToFpTG4gs-vM%3A

You mean the 3 different flanks and how they behave? I think its not needed for my game.

Anyhow, simulating battles fully with movement etc will be too much work for too little payback. But temporary "events" such as

found good defensive hill

skillful charge

commander slain

etc that modifies the armies strength for x days of simulation (or even just a section of the army such as all cavalry get a boost from the "charge" event) can of course be added pretty easily and is easy to convey to the player through the interface...