Don't you think longbows come too early?

[Levgre]

The problem is metal defenders - specifically enemy axes. Even axes are weaker there than bows, and so are swords.

Well, axes versus axes you break even with the melee bonuses.... so 5 v 5, without considering things that apply to the longbows also.

Generally I think breaking even is more than satisfactory if you have dealt collateral damage.
But perhaps there is a case for replacing all swordsmen and replacing them with longbowmen(surely,with little penalty,you can get the beaker benefit of skipping IW, if you have copper and feudalism. With no copper either and just feud, skipping IW is more circumstantial).



The generally purposes for swords are:

Kill archers (better than axes against them)
Kill catapults (they will be fully healthy, and axes lose to them more often)
Then mounted units, i suppose...

But against AI swords are basically a specialist unit, only better in that space when all the axes are weak enough that the axes won't defend against a sword instead of the archers. Or in cases where the enemy has no metal.

Longbows can do all the swords can, and then defend your cities better later against elephants and the like, while you reconstruct your new invasion force with elephants, macemen, etc.

However I don't think they easily replace the role of axemen, who are 70% cheaper, meaning you have more catapults for the same cost army.

 

[Feyaria]

I just played a game with Genghis Khan a few days ago.
Settings: Emperor, marathon, standard size, Great Plains
conquest victory in 360 AD, built 54 keshiks (39 were killed), razed 25 cities (kept the last 10 cities i conquerd)

At least on marathon and standard size its possible to win before longbows, maybe its possible to win earlier (maybe even in the BCs).

 

[bestbrian]

I just played a game with Genghis Khan a few days ago.
Settings: Emperor, marathon, standard size, Great Plains
conquest victory in 360 AD, built 54 keshiks (39 were killed), razed 25 cities (kept the last 10 cities i conquerd)

At least on marathon and standard size its possible to win before longbows, maybe its possible to win earlier (maybe even in the BCs).

Just for fun, play Mehmet and beeline Gunpowder - very early AD Jans don't seem to have too much trouble with LBs.

And Drill Promo units are awsome, but don't get too carried away with them, or those Knights you run into are going to be a very rude surprise. Like Zara found out with his gigantic stack of Oromos (80+) against my meager band of 20 Blitz promoted Conqs a few games back.

 

[PieceOfMind]

1. Medics matter more on faster speeds, where healing in effect takes longer. That said, medic III units essentially grant large stacks permanent healing (you just keep damaged people in and they heal on bombard or post attack turns). Sacrificing 1 general to give your stacks more speed/impetus longevity isn't that big of a deal, and IMO gives higher returns than post-war production of more experienced troops. Although again, this probably means more on normal than marathon, and epic is of course somewhere in between. Note: this has tremendous synergy with drill (since drill units with high str take less damage), and if you are doing this with infantry the turnaround is insane (drill III-IV infantry really shine). In a recent emperor game I went the drill line (while not even protective!) for an intercontinental invasion around the late 1600's - infantry/cannon using frigate galleon. I...capitulated 4 AIs, and not one had rifles before I blasted them all to pieces so it was just a city to city march fest. The medic helped heal the minor damage.

2. Good as longbows are, where are the mass cuirassers or rifles that late in an epic game ? Or if you are Unconquered Sun, where are the infantry ? My point is that longbows and cats are an early-mid ADs force at latest, and they hit hard then. Military in any era is served by striking ASAP once the advantage is secured. Whip, hammer, do what it takes to assemble the decisive force quickly! Drill III bows in 400 AD are better than Drill IV in 1200.

3. I love later-game GG's as military academies. Those ARE strong on slower speeds, perhaps stronger. How important is 2 extra XP, really? I'm usually happy with good cities cranking 2 promo units, and my HE city will frequently get west point (sometimes WP goes with ironworks instead depends on the game). Point is that you only need to settle 1 GG to make a triple-promo unit straight out of the gate...and guess what? The super medic unlocks WP instantly! A HE/WP/MA city with 40+ base hammers and later forge/factory/power is where it's at - those things can crank out some strong nonsense - my favorite being CR III artillery.

1.Absolutely agree the supermedic is more useful on faster speeds. Regarding the synery with Drill though, I don't really agree. Although Drill units have a greater probability of escaping combat unharmed, it is more due to the higher variance from coming out of combat with the average number of HP. For example, a Drill IV unit may be like
100HP 20% odds
60HP 60% odds
20HP 20% odds

Non-drill
100HP 5% odds
65HP 80% odds
30HP 15% odds

The point being that drill units when they do get injured can sometimes take more injuries than other types of promotions. It's really hard to analyse actually and I'm still not 100% on the conclusions, but I don't think having a higher chance of escaping unharmed always implies taking less damage on average when they do get damaged and so needing less healing.

2. I can always expect you, TMIT, to raise this point. Forget for a moment about what the year is. If I want a challenging game I may keep boosting the difficulty so that I'm not walking over longbows with my infantry. I can appreciate how it might take greater skill to get a huge tech lead at high difficulties, but for most players it doesn't change the fact that longbows are strong when they come into play, and they can be beelined relatively easily off the Monarchy techpath.

In this game I am playing above my comfortable difficulty, and this is a huge tectonics map with Better AI, with a larger than normal number of AIs, leading to probably slower teching for me at least, and probably faster teching for the AI due to more trading. I think these settings deviate far enough from your usual settings that it would make it hard for you to give a completely unbiased comment.

If you would like to play the game yourself I can give you the starting save and show me how you do, but untiil then I don't think it's fair to make broad assumptions about what tech position I should be at.

And for your information, with longbows I do have the decisive military advantage relative to Tokugawa and Monty to the south. Drill IV longbows in 1200AD are better than being dead in 400AD by the way.

3. I agree academies are going to be a better use for GGs and will usually build one with every GG after the granting tech. But I'm not going for hundreds of turns with my general sitting twiddling his thumbs. IMHO GGs earned before maces are without thinking going to be used as a leader or settled, never held for late game academies. By the late game the production modifiers start to crank up anyway, so the relative boost given by an academy can be smaller.

Longbowmen are actually more expensive than the metal units

axemen = 35
swordsmen = 40
catapult = 50
longbowmen = 50


So generally for attacking, unless you have a lot of free production atm, it is better to use the axemen and swordsmen still. Throw in some shock longbows for stack defense against axemen, or to hold cities once you capture them.

Problem is I had 0 metal. And longbows for their cost are far superior for attacking cities than archers are. Moreover, as I capture cities, the longbows do a supreme job of holding it, even with as little as 1 or 2 (or more depending on the enemy troops roaming around). With axes or swords etc. I'd need a better mix of units to repel counter attacks.

 

[azzaman333]

Only problem with Drill Longbows is those bloody Horse Archers. 4-7 Free Strikes rendered completely useless.

 

[PieceOfMind]

Only problem with Drill Longbows is those bloody Horse Archers. 4-7 Free Strikes rendered completely useless.

Trust me, with 5XP, Drill 2 Formation longbows do a decent job with the horse archers. Still, try to avoid flat ground with no fortify bonuses, but sticking to hills, forests or cities and horse archers can barely touch these guys. But obviously if you have metal you build spears.

In my above game example, I had no metal, and to prevent getting massacred by horse archers I went in and kept his horse source pillaged.

The real problem for Protective longbows is knights. But by the time knights are around I'd hope to have pikemen in my stacks.

 

[TheMeInTeam]

You know how I am with shadows! Although last time I played betterAI it lagged a lot . At least that isn't marathon (I'm gradually losing my will to play marathon, not because of any assertion that it's easier (IMO the only thing marathon makes easier in every game is a high score, not necessarily winning), but rather just my patience).

Edit:

In this game I am playing above my comfortable difficulty, and this is a huge tectonics map with Better AI, with a larger than normal number of AIs, leading to probably slower teching for me at least, and probably faster teching for the AI due to more trading. I think these settings deviate far enough from your usual settings that it would make it hard for you to give a completely unbiased comment.

Show me a completely unbiased human being, and I'll show you a LIAR . That said, I can see some things on the screen shot that are somewhat of giveaways. I've played huge maps before, several times and on different speeds/difficulties. I'm sure you're aware of the rule changes like more cities for national wonders (8), higher tech cost, and reduced distance maintenance more than I am, but that doesn't mean I'm not aware! Unless you did some punishing overexpansion or have been in breakneck wars for quite some time, 50 beakers in the ADs isn't ideal. Dirk once told me that 150 bpt was extremely low in that time frame and that 300 would be more typical/minimum . For a person who lived on warmongering and was trying to jump monarch---->emp at the time and finally in need of empire management, that was pretty disconcerting (but he was a deity player even then...you don't need 300 beakers @ 1200 AD to win emperor!!! Although it helps ).

So although there may be some bias with better AI and a larger map, methinks a heavier hitting longbow/cat earlier or more advance troops now would be in order. I see a mass buddhism bloc with a friendly pacal (which, unless the rules of betterAI are vastly different, implies pacal has an unlimited trade cap with you). The danger with such neighbors is that monty/toku will come after you due to their peaceweights, but again betterAI might impact that (i'd be able to tell a lot better by playing the map). Regardless, in my experience having more AIs on the map has never slowed the player down! It usually prevents runaway AIs or at least helps to do so! I would prefer monty/toku there, to just 1 big monty, for example. I'd also be slightly concerned early on about one of them sissyfoot stomping pacal. An early bribe might go a long way before those guys get too chummy, although the ideal situation is that the player trade-whores with some AIs and not others, creating a huge tech gap that the player can use to crush the backwards people - and conveniently the ones left out of the trade whore loop are usually less liked, too!

 

[PieceOfMind]

TMIT I appreciate the advice but it doesn't change my point. I'm not trying to argue longbows come too early. I'm just saying they can be used very effectively in attack. I'm not an elite player and regularly fall behind in tech. Obviously if I had the longbows earlier it would be even easier. The fact that I have an easy time walking all over Toku now suggests that if my play were to improve regarding tech-strength it would only further my point - not weaken it.

There is also the fact that my playstyle is such that I don't actively go for gamey type strategies. I don't like rushing AIs, for example. My style of play is not about taking advantage of every little detail about AIs that I know about from code etc. I get the impression much of the high level players do so well because they know exactly how the AIs work. What would you do if the AI was completely re-written? I like my strategies to be relatively AI-independent, meaning they would work however the AI changed. I'm not saying these strategies are better - it just makes the game more enjoyable in my opinion.

Having said that, I feel in the game above I'm actually in a reasonable position. I am not that far behind in tech, and at the point I'm at at the moment (a fair bit onward from that screenshot, about to take on Monty) I'm not behind in tech by much at all, and I'm ahead of Monty which is the important bit. In the screenshot my bpt will look abnormaly low because of unit upkeep and unit supply. Obviously by the time I take all of Japan's cities shortly after that, I will catch up again more easily. Obviously I am going for cities which will be of some value to me so having a low bpt at this particular moment is not really problematic. 10-15 turns later the situation will be very very different.

You could argue everything would be easier if I was able to tech better and just walk over everyone with superior units all the time. I don't deliberately try to do that every game though. I feel the more one relies on a tech advantage all the time, the less one appreciates the way combat has been balanced for civs that are closer to tech parity. I may appear to be contradicting myself because I'm attacking Japan with a slight tech edge, but that edge is very slight and I'm having very high success. For the hundred or so turns of war I probably lost about 5 catapults and 2 archers and maybe 1 longbow. I forget exactly how many.

The point of the example was not to show how awesome my overall gameplay is - just how effective even longbows can be in aggressive style combat.

 

[Bostock]

[M]edic III units essentially grant large stacks permanent healing (you just keep damaged people in and they heal on bombard or post attack turns).

I wonder if I am the only person who initially thought you were implying that siege heals while bombarding!

 

[TheMeInTeam]

TMIT I appreciate the advice but it doesn't change my point. I'm not trying to argue longbows come too early. I'm just saying they can be used very effectively in attack. I'm not an elite player and regularly fall behind in tech. Obviously if I had the longbows earlier it would be even easier. The fact that I have an easy time walking all over Toku now suggests that if my play were to improve regarding tech-strength it would only further my point - not weaken it.

I agree. It was kind of my point - those /bows would be utter wreckage if gotten in force early, and it's certainly possible. The reason I brought it up is that I've pulled (by both curiosity and in one instance necessity) super mass longbow/cat attacks in the timeframe I was describing. So you're right, if anything I agree with their versatility and usefulness in the right situation.

There is also the fact that my playstyle is such that I don't actively go for gamey type strategies. I don't like rushing AIs, for example.

I don't either. I'm always too tempted to give murphy's law a chance to screw me in early rushes, because they're expensive. You won't see many rushes in my summaries unless I am boxed in to hell or I spawn next to nappy/shaka.

What would you do if the AI was completely re-written?

I'd probably lose on immortal a bit more but not much else . Unless, of course, if it was written to specialize cities better, or to single out the human and not trade with it or attack it more frequently on purpose. But then anybody would do worse.

it just makes the game more enjoyable in my opinion.

This is the #1 priority in Civ. I just get carried away because improvement is a big part of the fun for me and sometimes I forget that isn't true for everyone !

You could argue everything would be easier if I was able to tech better and just walk over everyone with superior units all the time. I don't deliberately try to do that every game though.

I'm not exactly known for teching up a storm hehe. Most of my development before going emp/immortal came from just building more units, warmongering more, and learning to avoid strike. There's a couple things you can do in every game that will make life easier for you, such as a focused cottage bureaucracy site and running enough specialists to get GP's in a timely fashion. You can do those things independent of military production or techpath, and they'll certainly help. I...kind of prove the micro isn't too intensive once you learn it, but that's up to you.

I feel the more one relies on a tech advantage all the time, the less one appreciates the way combat has been balanced for civs that are closer to tech parity.

If only it were so easy for me . In games like LHC roosevelt and even that emperor "CE" walkthrough I had extensive periods of time where I was warring from the BACKWARDS position. I also did that in one of those longbow spam test runs I did - I fell pretty far behind while doing it but cities + diplo caught me up. After a while you learn both approaches to war, and sometimes switching from one to the other as the AI techs or trades for garbage!

The point of the example was not to show how awesome my overall gameplay is - just how effective even longbows can be in aggressive style combat.

Which is why I pointed out that if you hit with them earlier, they are EVEN MORE devastating .

I wonder if I am the only person who initially thought you were implying that siege heals while bombarding!

It definitely doesn't say that, but I misread things on occasion also .

 

[PieceOfMind]

I'd probably lose on immortal a bit more but not much else . Unless, of course, if it was written to specialize cities better, or to single out the human and not trade with it or attack it more frequently on purpose. But then anybody would do worse.

Exactly. I reckon Emperor is about the level where most players can rely solely on their own skills to win most, if not every, game. I haven't got every strategy completely honed yet so Emperor is a good challenge, but I feel that by the time I would be playing Immortal I would be getting way too far from immersion and relying on how AIs work.

My comment about the AI being re-written... Something like better AI city specialisation is exactly like what I mean.

In a way, I think it's reasonably doable to make an AI (hopefully in Better AI) that makes Immortal and especially Deity near impossible difficulties. But I reckon Emperor is around about the point where the AI would have to be super intelligent (like 20 years of 200 AI programmers work) before it could reliably beat a human most of the time. At Immortal and Deity I reckon the bonuses are enough to brute force the human player with only a bit more improvement on the AI side of things.

 

[TheMeInTeam]

I can XML mod AIs that tend to go runaway, and I'm not even a programmer. The in-game feature of "emphasize x" should be used in cities for AI specialization. I'm not sure how easy it would be to code but it would make the AI a LOT stronger.

Even without that though, making an XML AI that has a unit prob of 80, only attacks with a significant power advantage, favors the cottage improvement, and has a military flavor of "collateral" can make games nightmares. Even moreso if you make that AI ignore religion and favorite civics so that it has a close to 100% declaration chance once it's strong enough. I didn't think about playing with whether or not it tends to take vassals, but making it not do so would just make it obscene. If that thing spawns on another continent, at astro you're meeting a 50 city AI by itself .

Of course, anything monarch + could be made extremely difficult by having the AI single out humans.

 

[PieceOfMind]

I guess we've gone off topic, but that's really interesting, TMIT.
By the way, don't worry, no one is about to suggest making anything single out humans. One of the objectives in Better AI is to avoid human bias and remove any when found.

It makes me wonder then, how much can Better AI improve the AIs if ultimately they are going to be limited by their various personality values? If not many AIs tend to build enough siege or enough cottages then they're never going to reach the real potential.

Could you make an alternative AI mod (Runaway XML AI )? Tweak a few of the AIs so they are truly awesome opponents.

Especially your mention of AIs not declaring til they have a higher power ratio, got me thinking. AIs getting into wars because one or the other saw they were at ratio 0.9 or something is just going to overall slow all the AIs down. Any improved AI should have a greater reason for war IMO.

 

[Arlborn]

I can XML mod AIs that tend to go runaway, and I'm not even a programmer. The in-game feature of "emphasize x" should be used in cities for AI specialization. I'm not sure how easy it would be to code but it would make the AI a LOT stronger.

Even without that though, making an XML AI that has a unit prob of 80, only attacks with a significant power advantage, favors the cottage improvement, and has a military flavor of "collateral" can make games nightmares. Even moreso if you make that AI ignore religion and favorite civics so that it has a close to 100% declaration chance once it's strong enough. I didn't think about playing with whether or not it tends to take vassals, but making it not do so would just make it obscene. If that thing spawns on another continent, at astro you're meeting a 50 city AI by itself .

Of course, anything monarch + could be made extremely difficult by having the AI single out humans.

I gather you didn't really try that yourself? I might actually try it out of curiosity to see if the AI really gets stronger(then I may test it with auto-play). I am not sure I will find all the functions in the xml though.

 

[TheMeInTeam]

A long time ago I was creating AIs based on forumers. I did indeed make some fearsome things. Unfortunately I didn't do it in CUSTOM assets, and had to reinstall, so I lost that.

In test games, Washington with a buildunitprob of 80 and absolutely no diplo modifiers for religion (can declare at friendly, can't be bribed at all, high war odds at cautious/pleased/friendly) absolutely smacked the daylights out of the nearby AIs.

However, I can probably make them better now.

I am actually in the mood to do something like that. I will make 5-6 of the AIs godly opponents or opponents that are excessively opportunistic. Maybe fewer. They'd cause trouble if they fought each other unless one had a favorable start, but they'd essentially be "anti normal AI" AIs, capable of running through continents like nothing.

Spawning next to such an AI would probably be really bad too. Impossible to stop on immortal/deity just from unit spam. 50 units consisting of 20-30% siege and the rest assortments of garbage melee would be incredibly painful to stop in the very early AD's I'd imagine. The AI is bad with siege but stack attack still uses collateral first, so that is still a screw-job.

Edit:

I've got it. I read how the leaderheadXML works a bit more and I've come up with a way for the aggressor to essentially never take a vassal once it declares. I can make an AI that builds a large enough force, declares, and consolidates without taking useless vassals. Ooooh, this is going to be fun.

 

[UWHabs]

A long time ago I was creating AIs based on forumers. I did indeed make some fearsome things. Unfortunately I didn't do it in CUSTOM assets, and had to reinstall, so I lost that.

In test games, Washington with a buildunitprob of 80 and absolutely no diplo modifiers for religion (can declare at friendly, can't be bribed at all, high war odds at cautious/pleased/friendly) absolutely smacked the daylights out of the nearby AIs.

However, I can probably make them better now.

I am actually in the mood to do something like that. I will make 5-6 of the AIs godly opponents or opponents that are excessively opportunistic. Maybe fewer. They'd cause trouble if they fought each other unless one had a favorable start, but they'd essentially be "anti normal AI" AIs, capable of running through continents like nothing.

Spawning next to such an AI would probably be really bad too. Impossible to stop on immortal/deity just from unit spam. 50 units consisting of 20-30% siege and the rest assortments of garbage melee would be incredibly painful to stop in the very early AD's I'd imagine. The AI is bad with siege but stack attack still uses collateral first, so that is still a screw-job.

What if you put 2 of them next to each other? They'll just spam and spam units forever, since if they get equal starts, neither will dominate the other in power. That could be interesting to watch.

 

[Arlborn]

A long time ago I was creating AIs based on forumers. I did indeed make some fearsome things. Unfortunately I didn't do it in CUSTOM assets, and had to reinstall, so I lost that.

In test games, Washington with a buildunitprob of 80 and absolutely no diplo modifiers for religion (can declare at friendly, can't be bribed at all, high war odds at cautious/pleased/friendly) absolutely smacked the daylights out of the nearby AIs.

However, I can probably make them better now.

I am actually in the mood to do something like that. I will make 5-6 of the AIs godly opponents or opponents that are excessively opportunistic. Maybe fewer. They'd cause trouble if they fought each other unless one had a favorable start, but they'd essentially be "anti normal AI" AIs, capable of running through continents like nothing.

Spawning next to such an AI would probably be really bad too. Impossible to stop on immortal/deity just from unit spam. 50 units consisting of 20-30% siege and the rest assortments of garbage melee would be incredibly painful to stop in the very early AD's I'd imagine. The AI is bad with siege but stack attack still uses collateral first, so that is still a screw-job.

Edit:

I've got it. I read how the leaderheadXML works a bit more and I've come up with a way for the aggressor to essentially never take a vassal once it declares. I can make an AI that builds a large enough force, declares, and consolidates without taking useless vassals. Ooooh, this is going to be fun.

Did you change too many XML files? You can always put all your modified files in a zip and let us have a go at it as well!

 

[TheMeInTeam]

Did you change too many XML files? You can always put all your modified files in a zip and let us have a go at it as well!

No, I changed regular assets rather than custom when I did it way back when, which means no MP and no reverting to defaults. Maybe you don't always want to play with super runaway AIs when moving up in difficulty .

I'm thinking about creating a PYL map based on my new AIs, but I have to do them again first. Now that I know what every value means, AND I'm far more experienced with AI manipulation in general, I can probably make fun AIs - I have an idea for 4 themes so far: wonder spam, runaway murder warmonger, culture leech (it will vassal to warmonger AIs instantly and pursue culture like crazy. It will also over-prioritize liberalism just to annoy the player further), and erratic (ranges in its XML would be utterly huge, so it would behave extremely unpredictably and buddy up to different AIs each time).

We're a bit off topic though so I'll stop with this here after this post. Let me design the AIs and they we'll see something in the next PYL on S&T .

 

[Arlborn]

No, I changed regular assets rather than custom when I did it way back when, which means no MP and no reverting to defaults. Maybe you don't always want to play with super runaway AIs when moving up in difficulty .

I'm thinking about creating a PYL map based on my new AIs, but I have to do them again first. Now that I know what every value means, AND I'm far more experienced with AI manipulation in general, I can probably make fun AIs - I have an idea for 4 themes so far: wonder spam, runaway murder warmonger, culture leech (it will vassal to warmonger AIs instantly and pursue culture like crazy. It will also over-prioritize liberalism just to annoy the player further), and erratic (ranges in its XML would be utterly huge, so it would behave extremely unpredictably and buddy up to different AIs each time).

We're a bit off topic though so I'll stop with this here after this post. Let me design the AIs and they we'll see something in the next PYL on S&T .

You are right, way off topic hehe. But seriously, try to make it in custom assets this time! I don't think I'd play many games with it but I'd definitely like to try!

 

[champ82]

I think drill 2 or 3 leads to the bonus against mounted units. If you have a big stack promoting one or 2 of them to this might be a good idea. Maybe not too good against knights on flat non-woodlands, but it whoops horse archers.