I decided on four unit types - light and heavy infantry, and light and heavy armour. Heavy units are more likely to deal damage to enemies, and less likely to be damaged, but of course you have less of them to commit - the heavy armour in particular is extremely powerful, but you only have one - so you probably want it to go up against the enemy heavy armour. The enemy formation is randomly selected from a few options, so that you aren't sure if you face a force of mostly infantry with no heavy armour or an almost fully mechanized force. They will commit units steadily throughout the battle to different sectors.
At the end of each turn, every real unit in a sector makes an attack roll, causing a point of damage if successful. Then these damage points are randomly distributed across enemy units, based on how tough the units are (light infantry are more likely to take a damage point than heavy infantry for instance). If a unit takes six points of damage, they are removed from the game, and when they reach three points of damage their image changes to indicate this. You only get reports of exact damage on each unit if your commanders properly report it, but the images are always accurate, so even if you've had no reports from a sector in a few turns, you can tell if the units there are taking a lot of damage. And that's basically it - there are also some very basic rules for the AI to decide where to send their units, so that sometimes they will focus their heavy units in a certain sector, and other times they might start by focusing on one sector, before switching to another sector with their full attack after a few turns.
The second issue is one of balance and enemy behaviour. The enemy has more units currently, in theory because players could commit reserves early and destroy the enemy as they come in piecemeal, but that would require some very lucky guessing - and it's also only viable because the enemy doesn't react to your actions in any way. They need to have a better decision tree and far less randomness in how they send out units - early choices can perhaps be random, but soon they should be making choices based on where you have committed reserves and try to hit you where you're weak. The game also currently has no win or lose state - eventually both sides will have no more units to commit and each sector will only have units from one side, but that's hardly a satisfying end.
A big problem is that your only way of influencing the battle, the committing of reserves, is not very interesting or satisfying. All you do is shift units from sector to sector in the reserves before sending them in. The only units which really stand out in their utility are the light armour units, which you can deploy to the frontline of any sector at any time, while other units can only be committed to the sector they are currently occupying. There is very little reason to hold units back since the earlier you send them in the quicker they'll destroy the enemy, which is the only way to reduce damage taken by your units (this is also connected to the random behaviour of the enemy, where you can't really tell when they might have committed their best units, so you might as well go in early rather than waiting for something you have no idea will even happen). Some potential solutions to these problems below:
1) Units always deal one damage, and heavy units have a 50% chance to deal 2 damage instead. This keeps some randomness, but also means that large numbers of enemies will put down a noticeable amount of fire, making it more clear when a commander is greatly exaggerating how many enemies they are facing. Your commanders will estimate how much damage they are receiving with a certain degree of inaccuracy, so you can compare their damage reports to enemy activity they've reported, and see how they change over multiple turns.
2) The enemy will no longer randomly pick unit types to send, and instead will follow a script of unit deployments - for instance one light infantry and one light armour to a random sector for the first three turns, then pick the weakest sector and send one heavy infantry there for the next three turns, then send two heavy armour to the opposite sector. You will lose the battle if any sector has no friendly units in it at any time, and you will win if that hasn't happened by a certain turn number.
3) Add some abilities to the heavy infantry and heavy armour which take effect when you commit them. The heavy infantry reduce the current damage of all allies in the sector by two when you send them in, and the heavy armour will damage all enemies in the sector one point when they go in. This will hopefully lead players to want to hold back heavy infantry until other units have taken some damage first, and you want to send heavy armour in when there are a lot of enemies in a sector.
I'll hopefully try these changes out and see if they bring the game towards something that might actually be interesting to play. There are hints of how it could be interesting currently, such as when you see several heavy units appear one turn, then get downgraded to light units the next turn, before finally being confirmed or disappearing, and how you can get a sense of which commanders are more trustworthy, but it's lost in the fact you don't feel at all in control of things or like your decisions really matter.