More balanced and AI-friendly promotion-system

[Lord Tirian]

I will open a thread (in the balance subforum?) and post the tree once finished. Hope I get it done tonight. At latest it will be ready tomorrow.

That'd be nice. In general, though, it would do us good to get more discussion in both balance subforums, it's been a bit quiet lately (mainly me and Funak with mitsho and Gazebo popping in every now and then).

 

[Bernd-das-Brot]

I have opened a new thread in the balance subforum and posted a detailed proposal how I think the promotions should work for a more challenging and fun game.

Here is the link: http://forums.civfanatics.com/showthread.php?p=13538212#post13538212
@Lord Tirian: I will look in the other balance threads as well and post some ideas.

 

[guto8797]

If I recall correctly, the reason CIV 5's AI is so darn bad is that it cant forsee future turns. For the AI, what matters is what is right now, what is going on right now, not the fact that moving that tank to kill a near dead obvious sacrifice soldier unit will put the tank in range of a city and 5 bombers, this is why the AI is so bad, it cant forsee or plan in the future

 

[Gazebo]

If I recall correctly, the reason CIV 5's AI is so darn bad is that it cant forsee future turns. For the AI, what matters is what is right now, what is going on right now, not the fact that moving that tank to kill a near dead obvious sacrifice soldier unit will put the tank in range of a city and 5 bombers, this is why the AI is so bad, it cant forsee or plan in the future

It isn't quite so cut-and-dry, but yes, the primary reason that the AI suffers in combat is that is unable to 'view' the outcomes of movements more than a turn ahead, a reality that is due to the AI's inability to abstract situations (unlike the human mind). Because of this, minimization of situation-specific promotions can only help the AI.
G

 

[guto8797]

It isn't quite so cut-and-dry, but yes, the primary reason that the AI suffers in combat is that is unable to 'view' the outcomes of movements more than a turn ahead, a reality that is due to the AI's inability to abstract situations (unlike the human mind). Because of this, minimization of situation-specific promotions can only help the AI.
G

Honestly, that's the only thing we can do as of right now (on "we" read: "you and the awesome CPP team"). Any further optimizations would require: A: drilling a tunnel into the core of the game; or B: create superhuman AI's, restrain them with the Laws of Robotics and then hire a Detective to stop the inevitable revolution.

 

[Deluxe]

Honestly, that's the only thing we can do as of right now (on "we" read: "you and the awesome CPP team"). Any further optimizations would require: A: drilling a tunnel into the core of the game; or B: create superhuman AI's, restrain them with the Laws of Robotics and then hire a Detective to stop the inevitable revolution.

Well, I understand that for a mod like CPP this it too much to ask, but the tactical AI in Civ 5 is plain bad. Some strategy games manage way better results than that. I hoped for some deep changes in BE, but apparently that's not their main focus (sadly, because with a better AI I would really consider Civ5 as a perfect game.)

 

[Deluxe]

On a player yes, but wouldn't then the AI try to move a "desert archer" towards the desert which means onto the other side of the city it wants to take, through enemy infested forests...

Yeah, that would not change anything for the AI, but this could add some flavor for the players.

Btw, why not have a simple promotion system for the AI (simple strenght bonus that the AI will always take) and a more interesting one for human players ?

Honestly, that's the only thing we can do as of right now (on "we" read: "you and the awesome CPP team"). Any further optimizations would require: A: drilling a tunnel into the core of the game; or B: create superhuman AI's, restrain them with the Laws of Robotics and then hire a Detective to stop the inevitable revolution.

Well, I understand that for a mod like CPP this it too much to ask, but the tactical AI in Civ 5 is plain bad. Some strategy games manage way better results than that. I hoped for some deep changes in BE, but apparently that's not their main focus (sadly, because with a better AI I would really consider Civ5 as a perfect game.)

 

[guto8797]

Well, I understand that for a mod like CPP this it too much to ask, but the tactical AI in Civ 5 is plain bad. Some strategy games manage way better results than that. I hoped for some deep changes in BE, but apparently that's not their main focus (sadly, because with a better AI I would really consider Civ5 as a perfect game.)

I still think firaxis must have CIV6 in the works. No way IMO they would take so long to copy paste CIV 5 and reskin it, I think BE is just a cash grab while they make CIV6.

 

[Deluxe]

I still think firaxis must have CIV6 in the works. No way IMO they would take so long to copy paste CIV 5 and reskin it, I think BE is just a cash grab while they make CIV6.

Yes, I know for a fact that they are working on it (and from what I've heard, changing the whole engine to be compatible with tablets and mobiles). I talked with a guy working on TESO and who had an interview at firaxis a few months ago to work on this engine.

Let's hope they focus more on the AI this time

 

[MortalD]

What i did back in the day ( and i believe it was adapted to cep at some point ) was to remove all flat terrain def penalties. Basically moded flat terrain to 0% def bonus ( aside from marsh ).

To lazy to figure out why my old little mod dosnt work now but wjats ipmportant is that this little change boosted combat AI a LOT.

Atm combat is trivial. If ai expand they end up with gazylion unhappines, have no def buldings in cities and very poor army. ( 1 shoot from threbush take all hit point from medium city Oo )

But thats diff topic...

 

[TrystanCT]

AI seem to really struggle with taking cities because its whole tactic is to lemming charge the walls. If we want to make this really AI friendly a dramatic change could be to remove the city attack. That should make AI invasions much more effective if combined with the promotion change.

It seems there are three ways to improve the AI performance in the game. Modify the base AI code, play around with the personalities and flavors to make them more hard coded towards certain effective play styles and modifying the base game to be more AI friendly.

 

[Bernd-das-Brot]

AI seem to really struggle with taking cities because its whole tactic is to lemming charge the walls. If we want to make this really AI friendly a dramatic change could be to remove the city attack. That should make AI invasions much more effective if combined with the promotion change.

It seems there are three ways to improve the AI performance in the game. Modify the base AI code, play around with the personalities and flavors to make them more hard coded towards certain effective play styles and modifying the base game to be more AI friendly.

I would not miss the city attack if it would be removed. But with the toned down City strength it is not that much of a factor, especially compared to vanilla.

If I recall correctly my capital without walls (and no hill) had CS 7 in early game, other cities without garrison even CS 3 or 4.
In early modern times Lissabon with three defense buildings and 28 pop had a defense of ~45. In vanilla such a city had a CS of 100+, one-shooting nearly everything.

So the power of the city attack is okay, I think. But again I would not speak against abandoning it. Coastal cities with ranged garrison and a ranged ship in harbour means still three ranged attacks you/the AI can not avoid/directly strike back.

An alternative could be to remove city attack but give a strength boost to ranged garrisons attacking from within city walls. And a bigger boost from a melee garrison to city defense CS to make both interesting.

 

[mitsho]

Concentrated fire from up to 4 ranged attacks (city, ship, archer, newly built archer/ship) from one tile plus all the others from around is a bigger problem than the ranged strength per se. If we can reduce the numbers here, it'd be fine.

Ideally, I would like a system were ranged attacks give diminishing returns with the logic that at first, the enemy soldiers stand close to each other so it's easy to hit a lot of people. When the lines begin to thin, it gets harder to hit the same amount of soldiers with a volley. This might not be the most realistic system for warfare of all ages, but it's an intuitive explanation for diminishing ranged strengths. And it makes melee units much more important, no? (not sure if it'd add balance though )

 

[Lord Tirian]

I would not miss the city attack if it would be removed. But with the toned down City strength it is not that much of a factor, especially compared to vanilla.

*twitch* Personally, I'd miss city attacks a lot! They're one of my favourite changes in Civ5 because they a) free up the early game (building a garrison unit for every city just to avoid a random barbarian gobbling it up), b) make sieges a lot more interesting and c) make cities "special" in 1UPT as you get an extra ranged attack that you can't get otherwise.

Ideally, I would like a system were ranged attacks give diminishing returns with the logic that at first, the enemy soldiers stand close to each other so it's easy to hit a lot of people. When the lines begin to thin, it gets harder to hit the same amount of soldiers with a volley. This might not be the most realistic system for warfare of all ages, but it's an intuitive explanation for diminishing ranged strengths. And it makes melee units much more important, no? (not sure if it'd add balance though )

I'd like that, it sounds neat. Doesn't even have to be super complex or be a huge change. If units take (100 - current HP)/2 percent less damage, it'd be fine (i.e. -12.5% at 75% HP. -25% damage at 50% HP, -37.5% at 25% HP), it wouldn't change too much but distinctly favour melee unit attacks for the "kills". That would certainly add another dimension to melee units.