Strategy Notes From Vel - The Early Game....

[hoyatables]

Getting leaders

Real quick, because I now have to do all the work today that I should have been doing all week instead of reading message boards voraciously:

Last night, started yet another game with the Americans, this time on an archipelago. I hate them, but I wanted to see how the dynamic worked and how long it would take to get past the ancient era with minimal contact with other Civs. Anyway, as I secured my continent and moved over to the next island, I found a lot of barbarians. Horsemen all over the place. At least three or four encampments. And my Swordsmen quickly because Elite...I imagine it would have only been a matter of time before a leader emerged...except: Can you get a leader from a barbarian win, or must it be through battles with another Civ?

Oh, and I take back my earlier admonishment of ancient warfare. Because I had iron and horses from the get go, I was able to build some Swordsmen and chariots to help conquer the Romans before they got legions. Unfortunately, I wasted a lot of Archers before I figured this out. It really does all hinge on getting those crucial units/techs/resources at the right time, and before anyone else - at least before your neighbor.

And finally, my starting island was RICH with resources. Tons of incense (probably 8 of them) and also gems (5 of them). Even had horses and iron. I'd say it was 'average sized' - large enough to support about 12 cities spaced moderately apart; nevertheless, I definitely had more luck here than in the past.

[Velociryx]

Hey bud! And that sounds like a truly OUTSTANDING starting place! Dayum! Trade goods for days!!!!

I'm only working half a day today, so this will prolly be my only real strat. post, but here's a refinement of something I've been working on.

Considering problems with corruption (and not counting a fledgling civ taken out in the ancient era), it's probably only a good idea to totally absorb one rival civ during the coruse of the game. More than that, and the corruption will eat you alive! Besides....having absorbed (maybe) one civ in the ancient era and another sometime later should put you in a position of clear dominance (not to mention any cities you gain via cultural conquest or skirmishes for resoruces).

So...if you're ready to take out that "one big civ" but you're not happy with the idea of slogging through the nasty terrain, try this....it's worked like a charm for me....

1) Contact said Civ and trade luxury items and cash for a Right of Passage agreement, giving you access to his roads/rails (already have your army en route by the way...and preferably just off the coast!)

2) Land your troops and use his own rail system to position your forces exactly where you want them.

3) When you're ready, launch the attack! With enough force, you can take the Civ out in a single turn, or at least grab all his key cities and cut the roads and rails to them to prevent a swift counterstrike.



-=Vel=-

[Gatamelata]

I played some more of my game as the Zulus on huge, 16 civs, where I started in India. India is a fantastic start location. It's not isolated, but you can drop a city right away and start exploiting a cow square. Just to your north is another city site benefiting from a cow. South is a grassland/wheat tile, and also northwest is another good city site with grassland/wheat, IIRC.

So I spread like crazy. By 500 BC, six civs had signed protection pacts against me, and I was fending off agression from all sides. I managed to send a swarm of 4 Impi out to harrass and discourage. After several turns of their tender mercies, I managed to net an Aztec city and peace from four of the six aggressors. Now I have only the Americans (N. Africa) and the French (Korea) after me, and I suspect that Paris, with its food-stuffed pyramids, will soon be merely another Zulu annex.

I made some early mistakes in this game. I completely spaced rush building until about 700 BC. Man, that hurt, especially with four fast-growing cities. I did expand like crazy, however, extending north of the Himalayas into the Gobi Desert (incense), west almost to the Tigris and Euphrates, and east to the Pacific. I also didn't realize the power of the Impi until pretty late. What can I say? I've been working insane hours lately, so by the time I get to play some Civ3, I am often playing on autopilot and not optimizing. I also put most of my workers on auto-improve this city. I've never automated ancient workers before, and I don't think I will again. They mined cow and wheat tiles instead of irrigating, which was not too cool.

I've been thinking quite a bit about Golden Ages, and what is the best time to get one. On one hand, it seems that an early Golden Age is best, because giving your early cities a productive boost means that they will reach their productive potential earlier in the game. It's a concept I have yet to apply a name to, and I haven't thought hard enough about it to articulate it well, but it's something like compound interest... the more shields you can get early on, the faster your city will mature.

Then again, a Golden Age that comes later in the game, when your cities are larger and are working more tiles, means a larger overall shield boost, so perhaps later is better. Personally, I am leaning more toward the early Golden Ages, mainly because I think the proportional boost is the same no matter the age. I can say that when my Impi triggered a Golden Age, I ripped out some serious early infrastructure (temple, barracks, worker) in most of my cities within that 20 year period, and it was a vital boost. I'd be very interested to read some thoughts from you guys about early vs. late Golden Ages.

[BelGarion]

Aztecs!

I haven't heard a lot of people post about the Aztecs so I just wanted to put in a quick plug for them here.

This relates to modern vs ancient warfare. What I've found is creating a large base of land and cities in the early game is better for my style of play than starting small and trying to expand later.

The special unit of the Aztecs is the Jaguar warriors and they ROCK for the first few hundred years. They are the same as regular warriors but they get two movement. This can make a big different in the very early game where they act as excellent explorers and fighting units. They'll also retreat in combat too so they last longer than regular units.

In my last game I was the Aztecs and I shared a continent with the Iroquois to the north and the Americans to the south. I build about 8 Jaguars and 2 cannons and took over the Americans while they were still producing spearman. I built the Forbidden Palace in their former capitol city and suddenly I went from being behind from focusing on war so quickly to being in the lead. With the Aztec religious advantage I pumped out temples and my culture picked up a few more Iroquois and English cities. From there I consolidated my base, made friends with everyone and built like hell. I picked up most of the industrial wonders and all the modern wonders. I never joined a mutual protection pack and every 10 turns I would call everyone up and offer them 10 gold as a gift. They never got above polite but I never had another war and my culture and production rocked!!!!

I could have won with the united nations around 1930's but decided to win the space race in 1960's.
In the latest game I started I'm playing the French and they start soooo slow!! Now I really appreciate the Jaguar warriors. It's good building a road every two turns but I'm crammed into a smaller space and my neighbors (English and German) are flat out mean. Oh well, time to adjust the playing style.

[Hastus]

Originally posted by Gatamelata


I've never automated ancient workers before, and I don't think I will again. They mined cow and wheat tiles instead of irrigating, which was not too cool.

Thats the way I usally do it myself. Under Despotism I usally just irrigate maybe plains,and build mines on grasslands,cows,wheat etc.. Cities that I can not reach with irrigation I just put mines up all around. And later on when the city reach a population of six I just arrange it so it got zero growth and maxium production.
Then when I discover enginering I replace some of the mines with forest. Putting up mines early on will help production alot, and if you ever wonder why the computer have that super production at the start its becouse he does the same thing.

[Rosacrux]

So...if you're ready to take out that "one big civ" but you're not happy with the idea of slogging through the nasty terrain, try this....it's worked like a charm for me....

1) Contact said Civ and trade luxury items and cash for a Right of Passage agreement, giving you access to his roads/rails (already have your army en route by the way...and preferably just off the coast!)

2) Land your troops and use his own rail system to position your forces exactly where you want them.

3) When you're ready, launch the attack! With enough force, you can take the Civ out in a single turn, or at least grab all his key cities and cut the roads and rails to them to prevent a swift counterstrike.

-=Vel=-

Great Great strategy Vel. Fantastic. I am using it to wipe out Ze Germans in my game... but I learned how to use it the hard way: The Persians have used it on me in the previous game

Actually, I needed their help to take out the Bleeding Romans and fend off some English aggresivenes, so I signed the RoP pact and soon I found my empire full of literally dozens of immortals, archers and pikemen (later on, just before war, a dozen or so of Cavalry units also showed up).

Can you say swarming? When we finished with the Romans (I got 3/4 of their cities and Persia the rest) I had more than 50 Persian units on my ground. In a while (since they were standing on tiles I wanted to improve further) I told them politely to get the heck out of there... and I found myself overwhelmed by a sudden attack! They took 7 of my cities (one-unit defended - damned, the Persian were "gracious" FCS!) in one turn and carried on doing the "roll-over" thing to me... I should have known better, right?

Next time, though, I did the same to Bismarck... and I enjoyed every minute of it

[Lohrax]

Combat AI guess

Howdy all,

I am a romantic Civver at heart (good term Vel). But being a Unix admin also makes me interested in the 'how' and 'why' in addition to the 'cool' and 'Overseer! Lash that slave until he dies!' aspects of Despotic rule... There's nothing more satisfying than sending a worker to a far-off city and force-building a barracks to build More Military UNITS to further Oppress my Petty, Insignificant Worms into extending My GLORIOUS RULE over others.... ah... uh... sorry... kind of lost it there.

I too have noticed the AI's love of engaging in seemingly eternally-reinforced battles against well-fortified mountaintops. Having seen the behavior first-hand myself and reading about it here, I lost some of my currently-insufficient amount of sleep pondering the possibilities. What I came up with fit the hypothesis, could be way off the mark, but I think it's a worthy place to start.

I wonder if it assigns different tiles a 'target' value based on what's contained in them? For example, a mountain w/ fortress and 4 fortified Musketeer (E/E/V/V). You, the player, have placed a very high tactical/strategic value on that tile - why shouldn't the AI? Your own value is based on what you have built/garrisoned there - why shouldn't the AI?

To start, let's talk about garrisons. Perhaps the AI does a quick calculation to determine the human player's garrison valuation of the tile; that calculation being based on the human player having 'U' units of defensive value 'D' ('D' being the sum total of all defensive modifiers cased by tile type, structure locations, etc.) located in tile 'M'. The aggregate defensive value is (U * D). Factor in hit points (using some coded variable) and you have a single number, let's call that 'Kenny'. The AI assigns this value as 'human player tactical unit value of the tile'.

Now, the targeting routine. How does the AI target where to attack? I would say you could break this down into two parts - human tactical value and AI tactical value. The AI looks at what improvements have been built in the tile (cities, fortresses, mines, etc.) and assigns that a value - call it 'HT' - 'human player tactical tile-based value'. The AI uses it's own complex formula taking surrounding terrain into consideration and expansion goals, force levels, etc. and assigns that to 'AT' - 'AI tactical tile-based value' (for example, a one-tile chokepoint with a mountain dividing two chunks of a large continent is going to have one h**l of a high AT value). So, the AI ends up with two basic variables:

HT - the human tactical value (extrapolated from the effort the human player expended developing the tile)
AT- the self-generated strategic importance of the tile

The targeting routine would really only care about HT and AT - what is the tile really worth. The targeting routine may/may not look at K - based on my own and other's observations either situation is as likely: a) yes, it looks at it and increases the HT of the tile based on a K-based modifier thereby making the tile a more attractive target and b) no, it doesn't look at it and therefore blindly passes the targeting information to the force estimation routine - in effect ordering the force estimation routine to take the target (but not necessarily at all costs). Based on my (and other's observations), I can't really hazard a guess as to the behavior - anybody out there want to venture some ideas?Including the K value and attempting to extrapolate the human player's willingness to defend the tile using a small database populated by a serious of HT+AT+Kenny measued over time would be cool though...

Next, the force estimation routine. It looks at Kenny and determines how much military it must send out to kill Kenny. It SHOULD look at Kenny and determine how many hit points it needs to inflict in one turn and send exactly that (or more, depending on a coded aggressiveness variable). Unfortunately, I think the force estimation routine is broken. It SHOULD build units and station them somewhere for a one-time assault as that minimizes the rest and recovery of hit points for Kenny - ensuring that Kenny dies in one turn. Humans concentrate forces at the point of attack. I've seen the AI do it and I've seen the AI not do it and try the everlasting trickle approach. This is what leads me to believe that something is broken/needs tweakage there...

The targeting routine has run and assigned target tile priorities. It has passed these tile priorities to the force estimation routine, which has calculated the force required to take the target. The production routine builds the units and the movement routine moves them in for the attack. So now the big question that I don't have an answer for and has been keeping me up at night:

Does the AI simply not know that multiple units are stacked and think that K is much smaller than it is? In my situation, the fortress was right on the border, and there were several AI units within visual range...

Thanks for your time, and I look forward to some good old-fashioned idea exchange!

-Lohrax

"I love the smell of perfume in the morning."

[mambo4]

Waiting for Settlers/growing slowly

I've been trying to follow Vel's advice and get into an early warrior,settler, warrior,settler, repeat bui8lding pattern. But the city I seems to always grow too slowly, and i end up waiting for size 3 to come so I can build my settler. Add to that the fact that at size 3 i get disorder... how can I maximize my production to insure that I make settlers more efficiently/quickly?

[Jack_www]

I dont know about any of you guys but when you talking about your invading forces they seem a bit too small to me. When I invade the AI go in with at least 20 moble units or more. In my current game I have 34 calvery units, and I use every single one of them when attacking. I like to blitz in and take as many cities as I can in the first turn of the invasion. In my game I am the Americans and the French are on one of my broaders. I attacked them and took 6 cities in one turn, did not even need to do a RoP agreement etheir. I always like to land an invasion force that is much larger than the one I am attacking. The key to my sucess in my millitary actions is to take as many cities as can in the shortest period of time, not 2 or 3 a turn(which I did orginally) but 6 or 7 a turn I find is much better. I also attack civs that are weeker than me, thus helps me become more powerfull.

In my game I am also been trading techs with the computer. As a result I am earning 1200 gold/turn and I have the tax rate set at 0%. This allows me to instanously upgrade my units when new one become avaible.

In the early part of my game I ran out of room fast to expand my empire, so I had to use my army to take cities form my weakest neighbor. I find this very efective in expanding your empire early in the game, I never able to keep up with the computer's expansion any other way.

[Jack_www]

I have been able to find a way to get cities form the computer through deplomicy. A had a couple of open spots in my empire, and the computer decided that it was going to plant a city right there. Some of you will try to wait for your culture to assimulate the city and others would start a war and take it, and others who dont feel like going to war will just let the city be.

Well I was able to find a way to get the city form the AI through peacefull means. To do this you must have something that the AI will want though. What you do is you contact the civ that the city belongs too. Then click on the acitve button so that you can get the active agrements that you have with the civ up. Then if you are able to click on the peace treaty. It seems that the AI is more willing to sell you cities when you are renogiating a peace treaty then when your are not. This way you can buy the city form him for a tech or two and a lump sum of gold. Although you have to becarefull, because when the AI is mad at you they will demand something form you when you are trying to do this to renew the peace treaty. I did this sucessfully a couple of times in my game and was even able to buy a city form the AI that had horses in it. Up to that time I had to trade with the Greeks to get horses, but now I have my own supply of horses, which really helps if you goto war alot.

[vmxa1]

Fisrt I love this thread, second I finally won.
Now I just had to say something to Vel about the partisan rebel. I for one would never complain if the rare rebel won a battle, it is that the rebel wins way to often to buy. Afgan lost a about a million people to russians, so the rebel did win here and there, but lost 9x% of the time. I seems to me the regular hoplite beats the vet riflemen and calv on even terrain most of the time (in my experience), even elite troops lost at times to lower level units.
I agree the AI has a very good tactical understanding and I applaud them.
Trade is not so good, AI is alway mad if you do not offer more than you get, even if you are number one ranked. Even after beating one civ twice in war, they want more than they give? They will offer maps and want maps/tech and xxx dollars? I do not even want their maps. Often I would get personal and declare war for that crap.
I hate the horses going in circles in some cities, plese stop them.

[Arrian]

Assorted Thoughts and Ramblings

Definitely a good thread - I like comparing my way vs. some of the others. Frankly, I've been playing practically NONSTOP since the game came out (man, I need to get more sleep) and yet I've only played 3 civs thus far - so people's ideas about some of those I haven't played and strategies I have yet to try are great. The game takes me much longer than CIV II - I played that game like I was on speed or something - I could play out an entire game on a small or maybe even medium map in one night.

Anyway, regarding the AI throwing units at musketeers in a fortress on a mountain... I haven't seen anything like that. Quite the contrary. The AI, in my games, seems to know if it has no chance at winning a given battle, and does not attempt it. Instead, it will try to flank me and capture my workers or break my roads. The worker capture thing I think I've mentioned before - Soren, if you're listening, it's a major AI weakness. In CIV II, losing a settler or engineer sucked, because it died and you had to build another. Now, you can just recapture them while you're mopping up. I end up with so many captured workers, I use every last one of "mine" to add pop to new cities. and have even disbanded large groups to help build a temple - not really efficient, I know, but I was just tired of hitting the spacebar.

Jack - I agree w/your numbers of mobile units. Units get hurt. It happens. You need a bunch in order to keep the initiative. Lose it, and the AI may hurt you.

I have found that building and maintaining a large military (compared to what I did in CIV II - one def. unit per city and a couple of the best offensive unit available at key points) is essential. Not only does having a half-way decent army gain you more respect from other civs, it allows you to ramp up for a war quicker. There are, in fact (as mentioned by Hoyatables), noticeable periods of time when your cities will have built all the improvements they need for a while. Build barracks and start pumping out units - so long as they upgrade. The exception is Cavalry, which does not, but it is useful for a while. If you neglect your military you will probably pay for it. The AI is capable of attacking in force now, so you have to be prepared.

Be not afraid - unit upkeep is $$ now, and so long as you can afford it, the bigger your army, the better. Once you build the intelligency agency, you get to plant spies and see your opponents' order of battle. That, IMHO, is the only really useful thing about spies (at least on Regent level) - but it's very, very nice.

Hoyatables - your "late republic" war description is remarkably similar to several situations I've been in, with one notable exception - I'm democratic by then. Other than that, it just what I've been doing... usually in order to get my grubby paws on coal or rubber. Once the AI gets infantry, Cavalry is no longer a powerful unit. You can still use it, but it requires much more artillery backup, and you may as well wait for tanks. Of course, the one time I actually did use cavalry vs. infantry was to get oil so that I could build the damn tanks. Arg, that got nasty... definitely had a WWI feel to it - artillery banging away, massive attacks to gain tiny bits of territory. I won and all, but it hurt.

Lohrax - love the signature.

-Arrian

[vmxa1]

Gat, that is some romatic take on the battles, but it sounds like AI cheating to me. I agree that they may get weapons, who determines that? An afgan witha stingers or greek hoplite with a machine gun, would still distinguishable to you and it is not in the game so don't give them bonuses. All I ask is we have a way to determine if we should or should not attack. It smacks of changing the rules in the middle of the game to me. If you publish a hoplite as 1 3 1 and rifleman as 3 6 1 and we know what the modifiers are, then you can not decide to say he has an usi once we attack. That a hoplite would kill a rifleman some time is fine, but if they have 10 battles, which do you want to be? If no bonus except the rifleman is a vet or an elite, how you like you chances now? In my game the hop won every time unless I was on a hillside. Do you really like the idea that I can not attack even a spearman with a calv unit with certainly unless I am on a hill? If so then you got it.

[Kriton]

I've learned many things by reading this thread. I never played Civ2, but I think I've picked up some good ideas. I've played a few huge maps w/16 civs on emperor. In one game I expanded as fast as possible under the given conditions and I still wasn't as fast as the AI. I even placed a town near the AI border and then backfilled. I kept pace in science along with everyone else, sometimes being ahead. This may have been a detriment because of the money I could've been saving had I been behind a couple of civs in tech because of the tech modifier.

So, I was the smallest of the civs in terms of cities. I eventually got into Democracy and built my FP and corruption was virtually gone. Had I had more cities and been on a smaller map, I assume I'd have more. I had been building railroads when I finally got Nationalim. My grasslands and hills were all mined and RRed. I then switched to mobilized and started building tanks. I was getting a tank every 2/3 turns from every city! When I had enough, I switched to communism and went to war. I took down a civ twice my size. Mobilizing is like a GA for war. It's great!

Now I have a few questions to ask about the game:

1. Is a unit's attack strength modified by the amount of movement it has left? (like Civ1 and CTP)

2. When is the best time for GA? (i.e. should I have settled everywhere first? should I have cities up to size 12?)

3. Does an elite have 'exerience points' which when enough accumulate you get a leader or does he have a % chance every time he wins a battle?

4. Does the AI value techs based on their research cost only?

I have more questions that I forgot. Maybe I should take notes while playing, huh? TIA all!

[Gatamelata]

vmxa1: I'm not suggesting that the hoplite's stats should change. I'm just explaining how it's possible in my mind that a lightly-armed force could conceivably win against modern armor. Tanks aren't all powerful, and can be extremely vulnerable to infantry in an ambush situation. It only takes a couple of guys with rather cheap antitank munitions to ruin a tanker's day.

That written, I think in general you can expect a tank to win against a spearman. But a rifleman against a hoplite? That's a 3:3 matchup, so the rifleman should lose half the time, statistically speaking. But ten times in a row? Did you attack with ten riflemen, or did you reload your game ten times and replay the attack? I am almost certain that the game uses what programmers call a 'seed' to generate combat results; I suspect they did this to prevent the frequent "save the game, attack, reload if you don't like the results" hack.

Kriton: My only claim to knowledge is having played quite a bit and a single reading of the manual, but here goes:
1: No, I don't think movement points affect combat, except in the case of a fast unit withdrawing from battle with a slower one.
2: I've been wondering the same thing. My suspicion is that the best time is early in your game, after you've stopped expanding but before you've started your infrastructure.
3: According to the manual, there is a flat chance per victory that an elite will generate a leader.
4: I believe so, yes.

Hope this helps!

Gatamelata