Props to tactical AI improvements

[crdvis16]

The changes to tactical AI in recent versions has made warfare MUCH more difficult and I'm really digging it. It has really shown me how lazy and undisciplined I previously was with war. Have others noticed this in war-heavy games?

I'm currently working on games with Assyria and I've had to re-start multiple times because I'm getting myself into untenable situations due to not executing war well enough. Mistakes that I make now seem to be very reliably and harshly punished by the AI whereas before I largely got away with them. For reference I play Immortal/standard/standard.

Examples:

In previous versions I could usually blitz a nearby AI city and knock it down before much of a counter attack arrived from the AI but that doesn't seem to work now. Attempting to blitz and surround a city just leads to units being exposed and over-extended and any unit in that position seems to get focus fired and killed off very very quickly. In the past I could play a little more aggressively with my units and sometimes gamble to kill an enemy unit even if I was putting my own units in peril- it would sometimes get punished but sometimes not. Now it is almost ALWAYS punished and even units that I thought were relatively safe often aren't as the AI seems much better at the order of attack and striking from out of LOS with mounted units.

Terrain also seems to be MUCH more important than previously. Before I felt like terrain could be a minor annoyance to taking down a city or making progress on an invasion. Now I have to very carefully plan which city I can even think of attacking and from which side based on the terrain. Failure to take terrain into account means you're probably doomed in an invasion where previously you might be able to get away with poor planning by exploiting a relatively weak tactical AI.

All in all it seems necessary to be much more patient and careful in warfare than before. What is working for me now is to take up specific tactical locations in the border near a city I want to take and setup some kill zones where I can reliably kill AI units that expose themselves. After damaging or killing units for 10 turns or so the AI might be weak enough for me to begin inching forward and eventually push into positions that allow me to siege the city down. I typically can't outright surround a city anymore, as doing so probably exposes my own units too much but I'm still able to take cities down this way with very minimal losses.

The biggest problem with the slow/careful approach is that warfare takes much more time. It takes longer to capture cities without heavy losses so there's more time for other strategic issues to come into play. The longer the war goes the more you start to suffer from war weariness and have hits to your happiness or supply cap (though your enemy is likely suffering more than you on this front). A longer war also means there's more time for other AIs to join in. I've had many wars in my recent games that were going pretty well but before I could achieve my goal (typically taking the enemy capital and forcing capitulation to eliminate them as a future threat) I had to cut the war short due to weariness or another AI declaring war and me being unable to fend off two fronts simultaneously. In these games, failure to finish off the first enemy means that they eventually declare on me later, typically at a very inopportune time and I'm stuck in endless multi-front wars that might not lead to me being conquered but keep me on the defensive and stop me from really capitalizing on a warmonger strategy.

I'm also noticing that the big picture strategy needs to be much more on point in order to be successful. Choosing who to invade and when becomes a game-winning or losing decision now whereas before it seemed like the choice wasn't quite as critical. Going to war against a civ who currently has its UU is largely a no-no. Choosing to invade and take cities from a neighbor that leads to your civ having multiple long borders with other civs is asking for trouble in holding the lands. Going on the offensive against a civ that is ahead of you in military tech is MUCH more problematic than before as it becomes really difficult to keep your units in safe positions when they are taking much more damage from enemies with higher CS.

Before, domination style games felt like grinds of inevitability. Now warfare feels risky and much less of a sure thing where a bad decision has serious consequences. It feels... suspenseful!

Is anyone else being similarly pushed by the tactical AI these days or is it just me?

 

[crdvis16]

Also, I forgot to ask:

When going Authority do Deity players go out of their way to demand tribute from CSs much past the early game? I find that I'm often too busy to move troops near CSs in order to demand once the mid game starts because there's almost always a war to be fought elsewhere, but perhaps I'm missing out on free yields that good players rely on?

 

[tu_79]

Can't say for certain. I was too stressed by playing Emperor, not finishing any game, while new releases are online, so I thought to go back to King to feel the late game again. What made me turn back, among other things, was loosing too many units in the first attempts.

 

[Kim Dong Un]

Can't say for certain. I was too stressed by playing Emperor, not finishing any game, while new releases are online, so I thought to go back to King to feel the late game again. What made me turn back, among other things, was loosing too many units in the first attempts.

Even though I think Emperor is what currently offers the best challenge to balance ratio, my mantra has always been to play on King with standard settings and Continents, until VP goes Gold. Sure I win 9/10 games, but I feel we're all basically play testers, and King is good middle ground (although we'll continue to need perspectives from all difficulties and settings of course). Plus I can get through more games, and not feel too bad if I have to abandon one due to rare chance of a crash or altering bug.

 

[swapoer]

I am finding it harder too. I am looking for someone who would write a guide about how to carry out an invasion.

 

[crdvis16]

I am finding it harder too. I am looking for someone who would write a guide about how to carry out an invasion.

Not a guide, but I can try to list some tips that I think would generally be agreed upon for offensive war:

Rule #1: never trade units with the AI. They generally have a larger army, more supply, and more production so if you're losing units to kill theirs you are only going to fall behind. The higher the difficulty the less losses you can probably afford. Personally, I lose a few units during the course of a war sometimes and that's fairly easy to absorb but if you are losing a unit every turn or even every other turn you're probably going to lose even if you're inflicting even bigger losses to the AI.

The best way to avoid losing units is to not get out of position or over extended which typically happens when you are being too aggressive killing an AI unit that is low health. It's better to be overly conservative and allow AI units to escape than it is to be too agressive, kill the unit, and then lose one of your own.

Rule #2: Terrain matters. Attacking a city that has a single choke point through mountains is probably going to be a terrible slough that is not worth the time. Attacking a city that has hills in front of it from a low-land direction is also not advisable. Rivers, jungle/forest, etc are all obstacles that can either impede your progress or be used to your advantage. When I am planning an attack on a city I figure out ahead of time how the terrain is going to limit my ability to siege it down. How many tiles are available for siege units to be placed in range of the city and be able to have line of sight to attack (note that you typically never want siege/ranged units directly next to an enemy city)? How many tiles do I need to occupy with melee units to protect the ranged/siege? Is there good, flat land on the flanks for my mounted units to attack his weaker ranged units and still escape afterward? Are there certain tiles that I need to stay out of because they are exposed to too many attacks from his ranged? I typically know exactly how I'm going to eventually be positioned around a city in order to siege it down once I've killed off enough of the defenders in the way.

Rule #3: Time your wars when you have an advantage in army strength. This can be achieved by beelining certain military techs/units, having access to strategic resources for the stronger units they allow, or attacking when you have access to your UU (and they preferably DON'T have their UU). If you're able to unlock Knights before your enemy (or maybe they don't have horses in order to build Knights of their own) then that might be a really good time to attack, for instance. There's a number of units throughout the game that act as a power spike and can greatly advantage you if you unlock them before an enemy has. For most warmonger civs, attacking when you first have access to your UU is probably a good time, but other power spikes include Knights, muskets/gatling, artillery, tanks, and bombers (others might have their own personal preference here).

Rule #4: Either be able to surround the city or bring enough siege to be able to damage through the city healing. If you can surround a city with troops (typically after overwhelming the defenders and being able to old off any reinforcements) then the city will no longer heal and will take extra damage. In that scenario it's usually pretty easy to take the city down even without siege specific units if necessary. If you are unable to surround the city then you need to be dishing out enough damage to overcome the city's substantial healing each turn in order to make progress on taking it in a timely fashion, and that typically requires a number of siege units in positions where they can safely attack. There's nothing worse than fighting your way to an enemy city only to find that you don't have enough ranged/siege units to whittle it down.

 

[Minh Le]

Horses are extremely useful in offensive war in higher difficulty because they can hit then run so they can kill very effectively while dont expose themselves to danger. Even if youre playing Greece with a line of Hoplite, remember to bring some horses.
AIs are very good with using Fog of war, they often hide their units so a wrong move can ruin your attack and you will be forced to retreat. So using a recon unit is neccessary.
In the current version, AIs take every opportunity to kill you, so before declaring war on someone make sure to secure the other borders.

 

[MorphBer]

@rule 4: what if the city has a lake or mointain tile adjected or is costal?

 

[amateurgamer88]

Not a guide, but I can try to list some tips that I think would generally be agreed upon for offensive war:

Just going to add my two cents from my experience with war. There are some good tips but I disagree with or have some things to add to.

Rule #1: never trade units with the AI. They generally have a larger army, more supply, and more production so if you're losing units to kill theirs you are only going to fall behind. The higher the difficulty the less losses you can probably afford. Personally, I lose a few units during the course of a war sometimes and that's fairly easy to absorb but if you are losing a unit every turn or even every other turn you're probably going to lose even if you're inflicting even bigger losses to the AI.

The best way to avoid losing units is to not get out of position or over extended which typically happens when you are being too aggressive killing an AI unit that is low health. It's better to be overly conservative and allow AI units to escape than it is to be too agressive, kill the unit, and then lose one of your own.

I don't agree with never trade units. There will be time where you want to use speed to overwhelm the AI and will sacrifice some units at the start of the war. A scenario I can think of is when your enemy is busy fighting another war. You want to take advantage of a weakened front so you timed it to attack when most of the AI's units are on the other front. If you approach too cautiously, you will run the risk of either the other war ends or the AI responds, both of which will make your life harder. If sacrificing some units allow you to take a core city of your opponent, you are going to be in a much better position as it weakens them and strengthens you. In addition, you are also switching from attacking to defending which we all agree is easier in most cases. For sure, you shouldn't trade units in most cases as you cannot rival the AI in unit production most of the time.

Rule #2: Terrain matters. Attacking a city that has a single choke point through mountains is probably going to be a terrible slough that is not worth the time. Attacking a city that has hills in front of it from a low-land direction is also not advisable. Rivers, jungle/forest, etc are all obstacles that can either impede your progress or be used to your advantage. When I am planning an attack on a city I figure out ahead of time how the terrain is going to limit my ability to siege it down. How many tiles are available for siege units to be placed in range of the city and be able to have line of sight to attack (note that you typically never want siege/ranged units directly next to an enemy city)? How many tiles do I need to occupy with melee units to protect the ranged/siege? Is there good, flat land on the flanks for my mounted units to attack his weaker ranged units and still escape afterward? Are there certain tiles that I need to stay out of because they are exposed to too many attacks from his ranged? I typically know exactly how I'm going to eventually be positioned around a city in order to siege it down once I've killed off enough of the defenders in the way.

Terrain is interesting as it poses different challenges depending on the era you're in. Early on, hills can play a huge role in defending while, later on when you get units like Field Guns, the Indirect Fire can turn your advantage into a disadvantage. Obstacles are also highly dependent on the civs. For instance, rivers are horrible for all civs in attacking except for Songhai. Mountains can protect you from all civs but the Incans. Overall, it's far more complicated than what's put here and will require a guide for this topic alone. For what's mentioned here, it's definitely a great idea to plan ahead on how you will attack. Regarding placing ranged/siege units, it might depend on how aggressive you are. Sometimes, the knowledge of your enemy getting a tech that unlocks a strong unit might make you be more aggressive. As for planning how to position around the city, that can be being quite optimistic as warfare can be very chaotic and your units won't always be in their ideal location. I'd rather be adaptive than have a plan that I have to stick with. I still remember one game where, as soon as my enemy can embark, the battlefield suddenly got a whole lot complicated.

Rule #3: Time your wars when you have an advantage in army strength. This can be achieved by beelining certain military techs/units, having access to strategic resources for the stronger units they allow, or attacking when you have access to your UU (and they preferably DON'T have their UU). If you're able to unlock Knights before your enemy (or maybe they don't have horses in order to build Knights of their own) then that might be a really good time to attack, for instance. There's a number of units throughout the game that act as a power spike and can greatly advantage you if you unlock them before an enemy has. For most warmonger civs, attacking when you first have access to your UU is probably a good time, but other power spikes include Knights, muskets/gatling, artillery, tanks, and bombers (others might have their own personal preference here).

In addition to beelining certain techs or getting UU, it's never a bad idea to attack an enemy that's fighting multiple enemy. What people don't realize is that diplomacy can be very powerful in these cases. Even if you can't convince someone else to attack, befriending the enemies of your enemy will help you more in the long wrong. Those sort of timing will work very well in your favor as you will encounter less resistance. Why not take advantage of the chaos and use that to snowball?

Rule #4: Either be able to surround the city or bring enough siege to be able to damage through the city healing. If you can surround a city with troops (typically after overwhelming the defenders and being able to old off any reinforcements) then the city will no longer heal and will take extra damage. In that scenario it's usually pretty easy to take the city down even without siege specific units if necessary. If you are unable to surround the city then you need to be dishing out enough damage to overcome the city's substantial healing each turn in order to make progress on taking it in a timely fashion, and that typically requires a number of siege units in positions where they can safely attack. There's nothing worse than fighting your way to an enemy city only to find that you don't have enough ranged/siege units to whittle it down.

Bringing enough siege units is the better approach since you usually have to be noticeably stronger than you opponent to surround a city. However, there will be times when your supply limit is low so you only have a handful of siege units as you need other units as well. In those cases, learning how to fight with smaller numbers is absolutely vital, especially on higher difficulties when your enemies will easily outnumber your forces from the very start.

 

[amateurgamer88]

@rule 4: what if the city has a lake or mointain tile adjected or is costal?

A mountain tile will mean you have one less tile to occupy to blockade the city. With coastal, you do need ships. Lake is the only scenario where you cannot blockade since embarked units don't count, though I have seen CS build naval units in lakes so...

 

[crdvis16]

A mountain tile will mean you have one less tile to occupy to blockade the city. With coastal, you do need ships. Lake is the only scenario where you cannot blockade since embarked units don't count, though I have seen CS build naval units in lakes so...

I remember reading somewhere that you can place units around the lake to still blockade the city, but I haven't tested this myself to know for sure.

I agree with all of your additional points. Someone should probably do a really in depth guide. It would probably end up being a pretty huge post if it really did the topic justice.

 

[amateurgamer88]

I remember reading somewhere that you can place units around the lake to still blockade the city, but I haven't tested this myself to know for sure.

I agree with all of your additional points. Someone should probably do a really in depth guide. It would probably end up being a pretty huge post if it really did the topic justice.

Interesting, I'd love to be proven wrong with the lake adjacent thing. As for an in depth guide, I think I can start a thread and start updating with the help of everyone. There's a lot to learn but combat seems to be something that isn't going to be changed much until gold. An in depth one will definitely require the help of all the veterans and it would be nice to flesh out some stuff before gold comes out.

 

[tu_79]

My personal spike power unit is the Cruiser.