AI analysis: The first 100 turns
- Thread starter jdog5000
- Start date
I am a beginner in civ4 still. I think the thread is to make BBAI a better / harder opponent vs. an Human player.
Upgrade costs are not a huge factor for the time being, as the AI gets huge discounts on upgrading. A good power bar, but weak defenders might scare away an AI, but not a human player who will gladly exploit such an AI.
I hope i didn't miss something obvious.
The human player will often still often use the power bar to decide whether to go to war or continue to build up its army. Also the likely hood of getting AI allies in a war changes depending on the power bar.
Ok I do agree you have a point but Im just saying general high power units with out many abilities do have their place in defense.
Another unit that people underestimate are musketeers and the arguments for swordsmen are generally similar to the ones put forward for musketeers. They can be used against anything and the multipliers are bigger.
When I first started playing the game I always used to use the 10% strength upgrade, then I realised I could war against the AI better using more specialised upgrades.
However I then begin to play against my housemate, who is generally a weaker player but tended to stick to +10% and I begin to see the advantage. Against a clever human player it is often better to have units that can turn their hand to any situation rather than specialised units.
Hence why I think the AI building swordsmen as defensive units is not too bad a tatic and the AI can reuse as different unit types later.
PS at no point am I saying every unit should be a swordman but people will complaining if even if the odd defensive unit was a swordsman. This is the point I was arguing against.
Ok I do agree you have a point but Im just saying general high power units with out many abilities do have their place in defense.
Another unit that people underestimate are musketeers and the arguments for swordsmen are generally similar to the ones put forward for musketeers. They can be used against anything and the multipliers are bigger.
When I first started playing the game I always used to use the 10% strength upgrade, then I realised I could war against the AI better using more specialised upgrades.
However I then begin to play against my housemate, who is generally a weaker player but tended to stick to +10% and I begin to see the advantage. Against a clever human player it is often better to have units that can turn their hand to any situation rather than specialised units.
Hence why I think the AI building swordsmen as defensive units is not too bad a tatic and the AI can reuse as different unit types later.
PS at no point am I saying every unit should be a swordman but people will complaining if even if the odd defensive unit was a swordsman. This is the point I was arguing against.
MartinHarper
Warlord
- Joined
- Feb 16, 2009
- Messages
- 132
Note that both swords and axes have an iPower of six, so axes are better than swords for generating iPower. Of course, this shouldn't be a consideration for the AI.
1 Spearman with 25% fortification and +10% Combat I and +20% city defence being attacked by a chariot.
4 * (1 + .25 + .1 + .2 + 1) = 4 * (2.55) = 10.2
1 archer with 25% fortification, +20% city, +20% city defence being attacked by a chariot.
3 * (1 + .5 + .25 + .2 + .2) = 3 * 2.15 = 6.45
Using ^1.4 as a conservative power curve, we get 1 spearman is worth 1.9 archers defending against chariots.
Axes vs Axes:
5 * (1 + .25 + .1 + .2) = 5 * 1.55 = 7.75
Axes vs Archers:
3 * (1 + .5 + .25 + .2 + .2) = 6.45
1 Axe is worth 1.3 Archers defending against Axes.
Axes vs Spearmen:
.25 + .1 + .2 - 1 = -.45
4 / (1.45) = 2.76
One Spearman is worth 0.3 archers against axes.
Chariots vs Axes
.25 + .1 + .2 - 1 = -.45
5 / (1.45) = 3.45
One Axe is worth 0.4 archers against chariots.
So 10 spearmen (19 archers vs chariots) plus 15 axemen (6 archers vs chariots) is 25 archers vs chariots.
It is also 22.5 archers vs axes.
Conclusion: Spear + Axe is a less than efficient way to defend against an attack group that could consist of axes, or could consist of chariots, than just a stack of archers.
Sword vs Axe:
5 * (1 + .5 + .25 + .1 + .2 - .1) = 9.75
Sword vs Archer:
3 * (1 + .5 + .25 + .2 + .2 - .1) = 6.15
One Axe is worth 1.9 Archers vs Swordsmen.
If the unknown attack group is either a Swordsman or a Chariot based, one Spearman and one Axeman can dominate 2 Archers. Possibly not enough to pay for the production cost difference.
Finally, note that an Axeman+Spear defence group can use Axemen to attack the primary-Chariot attack group. As there are no units that get a defence bonus against Axemen, given level terrain the trade-off ends up being quite a lot better than letting the Axemen be attacked. If the axemen manage to penetrate the chariot-defenders, the Axemen are now at an advantage against the Chariots, instead of a disadvantage.
That final point makes a mixed Axeman+Spearman defence superior to an Archer defence even in the Axe/Chariot attack era. But it requires that the defender know how to counterattack with defending forces based on the attackers stack mix (you need to keep back enough axe defenders to soak the axes in the attack stack from reaping your spears, and you need to figure out a way to destroy the attacking stack in the field to prevent it from healing up after your attack).
This is probably too much analysis.
The catapult era is, naturally, different yet again. Catapults make hiding behind city walls (or in general, doing anything except mustering an attack force and destroying the approaching stack) by using your road-advantage, pointless.
4 * (1 + .25 + .1 + .2 + 1) = 4 * (2.55) = 10.2
1 archer with 25% fortification, +20% city, +20% city defence being attacked by a chariot.
3 * (1 + .5 + .25 + .2 + .2) = 3 * 2.15 = 6.45
Using ^1.4 as a conservative power curve, we get 1 spearman is worth 1.9 archers defending against chariots.
Axes vs Axes:
5 * (1 + .25 + .1 + .2) = 5 * 1.55 = 7.75
Axes vs Archers:
3 * (1 + .5 + .25 + .2 + .2) = 6.45
1 Axe is worth 1.3 Archers defending against Axes.
Axes vs Spearmen:
.25 + .1 + .2 - 1 = -.45
4 / (1.45) = 2.76
One Spearman is worth 0.3 archers against axes.
Chariots vs Axes
.25 + .1 + .2 - 1 = -.45
5 / (1.45) = 3.45
One Axe is worth 0.4 archers against chariots.
So 10 spearmen (19 archers vs chariots) plus 15 axemen (6 archers vs chariots) is 25 archers vs chariots.
It is also 22.5 archers vs axes.
Conclusion: Spear + Axe is a less than efficient way to defend against an attack group that could consist of axes, or could consist of chariots, than just a stack of archers.
Sword vs Axe:
5 * (1 + .5 + .25 + .1 + .2 - .1) = 9.75
Sword vs Archer:
3 * (1 + .5 + .25 + .2 + .2 - .1) = 6.15
One Axe is worth 1.9 Archers vs Swordsmen.
If the unknown attack group is either a Swordsman or a Chariot based, one Spearman and one Axeman can dominate 2 Archers. Possibly not enough to pay for the production cost difference.
Finally, note that an Axeman+Spear defence group can use Axemen to attack the primary-Chariot attack group. As there are no units that get a defence bonus against Axemen, given level terrain the trade-off ends up being quite a lot better than letting the Axemen be attacked. If the axemen manage to penetrate the chariot-defenders, the Axemen are now at an advantage against the Chariots, instead of a disadvantage.
That final point makes a mixed Axeman+Spearman defence superior to an Archer defence even in the Axe/Chariot attack era. But it requires that the defender know how to counterattack with defending forces based on the attackers stack mix (you need to keep back enough axe defenders to soak the axes in the attack stack from reaping your spears, and you need to figure out a way to destroy the attacking stack in the field to prevent it from healing up after your attack).
This is probably too much analysis.
The catapult era is, naturally, different yet again. Catapults make hiding behind city walls (or in general, doing anything except mustering an attack force and destroying the approaching stack) by using your road-advantage, pointless.
Now we're straying a bit off-topic and the above analysis seems to need a lot more tactical decisions than the AI has for the moment.
For now i think we need KISS (keep it simple stupid).
The AI needs to be able to fend of:
I am sure (but don't know) that the AI is not analyzing it's neighbors armys (e.g. does he have lots of axes, chariots,... ). So there can only be a basic distribution of useful defenders (like 1x archer, 1x spear, 1x axe basic example for you to tune up).
As long as there are no advanced tactical defenses i would say Swords should not be possible defenders.
For now i think we need KISS (keep it simple stupid).
The AI needs to be able to fend of:
- Axe Rushes
- Chariot rushes
- Horse Archer rushes
I am sure (but don't know) that the AI is not analyzing it's neighbors armys (e.g. does he have lots of axes, chariots,... ). So there can only be a basic distribution of useful defenders (like 1x archer, 1x spear, 1x axe basic example for you to tune up).
As long as there are no advanced tactical defenses i would say Swords should not be possible defenders.
MartinHarper
Warlord
- Joined
- Feb 16, 2009
- Messages
- 132
If we can't get the AI to only build defensive sword when appropriate, it would be better not to build defensive swords.
Fimbulvetr
Emperor
In the beginning (Vanilla), Swords had only UNITAI_ATTACK and UNITAI_ATTACK_CITY, in Warlods they got UNITAI_RESERVE and UNITAI_CITY_DEFENSE and didn't lose those flags since.
I agree, it's pretty obvious that UNITAI_CITY_DEFENSE is stupid for this unit, for Jaguar even more so, so why isn't it disabled already? It should be a really old issue if it's there since Warlords.
Anyway, I'll be sure to disable it, at least for my own games. And what does RESERVE do, to be removed as well?
I agree, it's pretty obvious that UNITAI_CITY_DEFENSE is stupid for this unit, for Jaguar even more so, so why isn't it disabled already? It should be a really old issue if it's there since Warlords.
Anyway, I'll be sure to disable it, at least for my own games. And what does RESERVE do, to be removed as well?
I think Reseve is for building units that aren't needed to defend the city right now, but can be moved around in response to threats that show up later. (city defenders are stubborn about moving)
I am not sure if a Reserve unit will be used to attack or not.
I am not sure if a Reserve unit will be used to attack or not.
r_rolo1
King of myself
IIRC Reserve units are basically floating defenders. Again IIRC they are also part of the SoD the AI do....
jdog5000
Revolutionary
Reserve units do a few things mainly inside of a civs borders, including escorting settlers, guarding cities and bonuses, manning forts, and the like. They will shift around to threatened cities and counterattack nearby enemy units only if they have good odds. They do not join offensive rushes and (aside from settlers) will never group with other units either.
Based on what jdog5000 just said, swordsmen sound like alright reserve units to me - but not so much city defenders. I'd use my swordsmen for that kind of thing, but with the caveat that they should become attackers when I need them.
When I'm defending, I typically do a lot of counterattacking against the enemy stack. So I prefer to use units that can attack as well as defend, such as axes and swords. I almost never use spears because they essentially can't be used to counterattack unless the enemy doesn't have any melee units of their own. It seems to be that it is possible to get by with fewer units if you are willing to be flexible with what each unit is used for; and of course it is often advantageous to save upgrades until you need them, so that your would-be city raider swordsman on their way towards the enemy cities can be diverted to become forest defenders if need be.
ok... I doubt I'm saying anything particularly useful, so I guess I should stop ranting and go to bed. Goodluck with the AI improvements.
When I'm defending, I typically do a lot of counterattacking against the enemy stack. So I prefer to use units that can attack as well as defend, such as axes and swords. I almost never use spears because they essentially can't be used to counterattack unless the enemy doesn't have any melee units of their own. It seems to be that it is possible to get by with fewer units if you are willing to be flexible with what each unit is used for; and of course it is often advantageous to save upgrades until you need them, so that your would-be city raider swordsman on their way towards the enemy cities can be diverted to become forest defenders if need be.
ok... I doubt I'm saying anything particularly useful, so I guess I should stop ranting and go to bed. Goodluck with the AI improvements.
I have a question about the AI strategy tags. Let's say swordsmen have AI_attack_city, AI_attack, and AI_reserve. I build a swordsman, and he initially gets the AI_reserve tag put on him.
Is it possible for that tag to ever change to the other two? Or is that swordsman an AI_reserve swordsman for all time into eternity, even if initially swordsmen are equally capable of being assigned as the other two?
It would be nice if tags could change. Like, if the AI is putting a stack together, and it needs an AI_attack_city swordsman in that stack more than it needs an AI_reserve swordsman camping on that iron mine.
Is it possible for that tag to ever change to the other two? Or is that swordsman an AI_reserve swordsman for all time into eternity, even if initially swordsmen are equally capable of being assigned as the other two?
It would be nice if tags could change. Like, if the AI is putting a stack together, and it needs an AI_attack_city swordsman in that stack more than it needs an AI_reserve swordsman camping on that iron mine.
jdog5000
Revolutionary
The AI player will change the UNITAI classifications of some units occasionally, but this is really fairly rare. For the vast majority of units, they are born one type and stay that way forever. There are two main examples of switching in BTS/BBAI right now: extra settler transports will switch to assault sea when the AI is preparing a naval invasion, and "floating defenders" like reserve will switch to more attacking types when the AI player adopts the CRUSH strategy (throw everything it's got at the enemy).
So under some situations the AI will do exactly what you're suggesting. However, it only does these UNITAI switches under particular macro circumstances.
So under some situations the AI will do exactly what you're suggesting. However, it only does these UNITAI switches under particular macro circumstances.
MartinHarper
Warlord
- Joined
- Feb 16, 2009
- Messages
- 132
Does this mean the hammers invested in the attack archer are wasted? Or do they get switched across to the construction of the city defense archer?
jdog5000
Revolutionary
They do get switched over, no problem there.
