I am trying to avoid that aspect of Z game, because I dont like such type of games where one or few mistakes at the start can make the game pointless, and because it is very hard to balance the game. Actually holding one more territory (point) will give the player/AI great bonus and will win the game. I made the opposite - if the enemy is losing territory its units are becoming stronger and if it has more territory then the units are less powerful. Also many players prefer to build their base, upgrade, build units and defense until they are strong enough to attack and in my game this means total defeat usually because less territory means less resources, unit types, etc. But with what I made territory close to players base is always the easiest to capture except when the game is in middle stage and the AI put there a lot of units.

Imagine troops needed to be resupplied with something. The more territory someone expands the longer and harder it is to resupply troops which makes sure that the player with the most map control does not snowball to an ez victory. You could also use the same mechanic to let the attacking player to bleed out the defending player. Without supplies the defending player has to surrender, but the defending player has a chance to comeback because the attacking players supply lines are thin and far from his original base. You would need to test how long units take to go through their supplies, no one wants a long boring siege but a quick siege may be fun

Just avoid a final challenge that is triggered at a certain point. Eg: You must kill a hard boss that is spawned when you control 70 % of the territory. This will just make gameplay awkward as player will try to balance at 65% and build strength, and only trigger "final challence" when they feel prepared enough to not be able to loose.

I think the AI giving up when it's impossible for the player to not win is ok. In any case avoid the situation (i think warcraft 1? or dune 2) when you have to kill every last units to get the win. Searching the map for stragglers are very very boring.

Better to win when killing some main building (that cannot be hidden, or is revealed when only 10 buildings remain by that player). Or just win when only 10 buildings remain:)

When the game is "already over", do not force the player to continue. Actually "winning" may be needed to unlock the next map or similar so players feel forced to continue if they need to.

On 22-2-2018 at 10:48 AM, Pleistorm said:

I made the opposite - if the enemy is losing territory its units are becoming stronger and if it has more territory then the units are less powerful

So then it is not a RT-strategy game but a puzzle/strategy combination.

The puzzle-part is this - one-time (bottleneck) - when players figure out how much resources they need to save before trying to overwhelm their opponent.

Instead of rebalancing the game mid-play i would make a proper balance from start(for example +10 resources per time) that gets outperformed(after a while) by balance from gameplay(for example +1 resource per territory per time),  but this all depends on what kind of game you're making.