So very cold (of the map generator)

[ShuShu]

Worth it...

...and possible, are two diferent animals. From the early days of Vel's strategy guide to now, it has been accepted as a truism that you must start your city as early as possible because there will be no time to catch up.

I believe your thread debunks that statement.

As for 'worth it', I think it depends on your strategy. I only play emperor and take a wonder building strategy. In that world it is worth the twenty turns to find the river with cattle and 8+ grasslands with resources.

If you are playing a strategy that builds warriors and settlers, you just need to get production of 5 for a pop of 3. (2 grasslands with resources). So its not worth it to do much exploring beyond that requirement.

With either stategy, it is worth it to explore until min requirements are found, and possible to do so and still win...

[Aeson]

Well here's the map, forgot to include it in the last post. Down to 89 MA's, 14 Armies. Rome's military is completely wiped out though.

ShuShu -

I haven't seen many people talk about searching for a better starting location, except in the OCC and 5CC threads. How many turns in do you feel is the 'breaking' point where further movement will just cut into efficiency, regardless of the future city site?

[Velociryx]

Awesome job, Aeson!

Just for the record...I'm not out-and-out against spending a couple turns looking for a better place to build, but there's an opportunity cost associated with that....the early turns are *hugely* more important than the later ones, and if you spend too much time scouting with your settler, you'll find yourself playing catch up....of course, given this particular game, that was the whole idea, but in general terms, if you've got some decent raw materials in your initial site zone, it's pretty efficient to build and run with it (not to mention that should you explore, you *could be* in for the shortest game in history...being defenseless, if a rogue barb should appear in those first couple turns, you are pretty much toasted). Best case, you'll lose most of your starting treasury, worst case, restart.

Had it been me, I'd have prolly just built there, used the fish, mined the hill, and once I had the starting island mapped out, set the centrally located city to building the palace.

From time to time, I'll find myself shifting the ol' settler over a tile, but there's gotta be a pretty compelling reason for me to do so (ie - one tile over, and I can plant my settler on a hill, with a gold mine, adjacent to a river....OR....I'm sitting on a shielded grassland tile, one tile over is plains...I'll let my worker go ahead and start working the tile, and move the settler....stuff like that).

Ahhh, another possibility....if you're playing expansionist, it only makes sense to move your scout first. Especially if you can end his first round's movement on a mountain, the expanded view might offer a compelling reason to scadaddle. Otherwise, I'm generally happy building where I drop.



-=Vel=-

[Aeson]

A couple of points about wandering around and barbarians.

Over at CivFanatics there is a thread about goody hut probabilities. The most interesting part of the thread deals with how the game treats huts opened before the foundation of your first city. It seems that barbarians are never an outcome before you found your first city (pseudo expansionist effect).

I was wondering how long I could wander around on that iceberg before encountering barbarians. I opened up the 4000BC save and just wandered around. The 3 goody huts of course didn't spawn barbs, it wasn't until the AI started building cities on the iceberg that barbs finally showed up (1275BC). Maybe there is some sort of continental city check made before barb camps are first placed.

Of course even if there is, it would still be a gamble on whether or not the landmass is shared with anyone else. I've noticed a barbarian grace period in every game I can remember though, always at least 10 turns before they start showing up. The exceptions are on edited maps where the barb camps are placed by hand. They are active from the get go.

[ShuShu]

It should be a simple max min problem...

But my math chops seem to be on the fritz today. But here's what I started with.

Toss out population growth as a variable and treat it as production (via rush and via additional works squares).

Forget about a formula that incorporates production growth rate and just utilize Average production.

Pick a target production and you know you have spent too long when you cross it.

Assume I am racing to Great Library: I need 540 production (2 warriors, a temple and granary). If I have a nice start but no river, I can average maybe 7 shields which will take me 77 turns. If I spend some time to find the perfect river site, I'll average 20 shields, which will take me 27 turns. That leaves me 50 turns to find the perfect spot. If however, I find an average river spot in 10 turns I'll average 15 shields and hit 540 in 36 additional turns which means I only have 9 turns to find a better location. (PS my numbers are off as it usually takes me 40 turns to the first wonder from foundation not 27 as stated above, but 16 is not unheard of for my second. The point is the relative averages)

Bottom line is, I will always search for at least an average river for my starting city, but will stop searching once I find one. I will Move to a better than average river site if it takes fewer than 9 moves to reach the better of the two river sites.

If I am playing REX. My capital will not grow over 6, so the river does not matter. Food is more important than production because when all else fails, rush the settler. The only need for exploration is to find grassland or better. Given that I only expect to use 3-4 tiles out of 21 anyway, all sites will be more or less equal. Hence, no need to search.

[Barfus]

are the romans the red or blue on that map? or are they completely gone already?

[Flight]

lol aeson
your game kinda sums up civ3 from my pov atm - start behind and end up way, way in front being a bit bored at the end
need better AI scaling

very very well done though, for patience more than anything
how does it end....

[Aeson]

Rome: red,
Germans: blue
English: orange
Russia: brown
Egypt: yellow

I agree Flight. The AI, regardless of the 'head start' they are given, doesn't do very well at maintaining a lead. Taking the initiative in every situation is important in Civ3, and the AI rarely does it, they always react to the player. Once parity is achieved by the player the games usually play out the same way. This game was interesting to me because given equal starting locations (and a less crowded map) I usually gain parity on Deity by ~1500BC. This game took significantly longer than that.

One thing that would help the AI is to produce much more mobile offensive units, and not so many defensive. MI's are powerful, and usually can eat up a MA or two while in a metropolis, but it turns into a war of attrition that the AI can't win. I just keep producing MA's and take city after city with losses, and the AI has no chance to retake their cities (especially if I raze them ). My production capacity remains the same while the AI's is further and further reduced. If Rome had had the number of MA's that they had MI's, I doubt things would have been so easy. The other real problem with the AI is that distance doesn't enter into their target aquisition. The 'easiest' target is what they go for, rather than try to break through and destroy my own offensive capacity. They just end up exposing their troops and allowing me to keep the initiative.

I'll finish the game eventually, but there really isn't a whole lot of incentive left. This is just the point where it's already won, all that's left is to continue producing units and throwing them against remaining cities. Also my brothers left some new games (for me) when they came visiting, trading for some of mine. One of the games is Baldur's Gate II which is eating up most of my computer time right now.

[Arrian]

Aeson,

I totally agree about the AI's unfortunate tendency toward building very defender-heavy armed forces, and their really unfortunate tendency of going on ill-advised flank marches to try at get at a size 3 piece of crap rather than striking for the jugular.

The AI has concentration of force down pretty well. But it still builds the wrong units (or rather the wrong ratios of unit x and unit y) and has real issues with recognizing a target or a threat (human invasion stack).

I must say, however, that I recently came up against a relatively large AI mounted force. Keep in mind this was on a standard map. The Americans built up perhaps 20-25 knights, and only had a couple of longbowmen (which all died attacking my entrenched musketmen/cannon/cavalry, but that's beside the point). Normally, we see the opposite: the AI fielding masses of longbowmen and pikes, and later on riflemen or infantry, with practically no attack troops.

One more thing: ealier I said the AI has concentration of force down. What I should have said is that it has been taught concentration of force when on the offensive but fails to do it defensively. Large human invasion stacks often march unhindered to a target city, only to encounter 2-3 defenders, 1 attacker, and a bombard unit or two, when your spies are reporting 100+ AI infantry.

[Disclaimer] No, this does not mean I think the CivIII AI is bad in comparison to other games, such as CivII. But anything can be improved [/Disclaimer]

-Arrian

[Geekinstein]

Wow, just read thru this thread...

...as so many others have said, I'm happy when I can win at Deity in a well selected start with everything else carefully set by me.

'Way cool stuff in here...

On being defender heavy...I tend to do that by grabbing rough terrain and daring the AI to attack (rarely does) and then I tend to launch attacks off those powerful points. Aron Nimzovich (in Chess) liked to use this; called it "overprotection" ... nobody else in the Chess world much cared for it though...

...obviously they didn't play enough Civ III.

[Theseus]

On the topic of concentration of force:

I'm playing a game where Japan is the killer AI civ, having destroyed America for control of a mid sized continent. We're neck-and-neck for the tech lead, and both have 1) monopolies on 2-3 luxuries, and 2) excess of all strategic resources.

I was gonna wait for Tanks to deal with them, partially 'cause I didn't want to bother with a Galleon attack, but they pissed me off with a semi-random attack, and lo and behold, their closest city has Iron and Coal, and none of mine do!!!

Well, gotta have the IW.

So I go on a pretty heavy-duty attack for the time, although limited by the number of galleons immediately available:

8 Rifle
8 Cav
1 Army (1 Rifle + 2 Immortals)

This was just meant to be a placeholder while I brought in reinforcements:

8 Rifle
12 Cav

OK, at this point I take Boston. One more set of reinforcements:

2 Armies (same configuration)

The first Japanese attack was not meaningful... some Cavs, and Rifle and leftover Swords reinforcing a mountain adjacent to Boston.

Next turn I get Rep. Parts, so I upgrade all to Infantry, and also add an Infantry to each of the Armies. I also have a bunch of Cannon on the way, but not there yet.

So I'm feeling pretty good:
3 strong Armies
17 vet Infantry
20 vet Cavs

Let me tell you about the AI's concentration of force!!!!

I lost count of how many Cavs attacked... suffice it to say, that when the carnage was over, I had LOST 2 Armies, the third was at 1 hp, all but 2 Infantry at 1-2 hp, and 5-6 Cav!!!!

Thank god I got further reinforcements in... more Japanese Cavs showed up, but the Arty and another 8 Infantry saved the day (a real squeaker... I transported them in as Cannon and upgraded that turn).

I did get a GL on counter-attack, and built the IW for culture.

[King of Rasslin]

Unbelievable...

[Aeson]

I was playing GOTM09 over at CivFanatics and I finally ran into an AI that seemed to 'get' the advantages of mobile attack units.

I won't post any spoiler info, but one of the AI had built up a huge force of Horsemen. When I invaded I was expecting counterattack by Spearmen/Archer stacks (they had no Iron), but instead I was fighting off hordes of Horses. My War Chariots and Swordsmen actually had a very difficult time making any headway against them. It wasn't until I upgraded to Knights that I finally broke through.

So it is possible that the AI can use these tactics, but why don't they more often? My only guess is that they built Horsemen because they lacked Iron, but most of the time without Iron they just build Archers.

[King of Rasslin]

Did you win by conquest? Let's see what happens!

[Theseus]

I'm firmly convinced that it's all about the AIs having money.

This is just a gut feeling, but it seems that when an AI is rich, it doesn;t prioritize dual-use units (ie, Riflemen, Infantry, etc.), and will go ahead with building attack-only units.

This is one of the prime goals in my quest to generate killers.