Map size and game difficulty?

[Campo]

I encountered an odd situation and wonder if it's generally known.

A month ago I won my first diety game -- small world, 7 civs, raging hordes. I played a perfectionist-aggressive approach -- founded and grew 10 cities, accumulating the best wonders, until gunpowder. Then I set off on conquest and won that way. I played two more games the same way, refining my technique a little, with the same results.

So then I decided to play on a slightly larger map (all other settings the same). Due to time limits, I didn't want too big a map. So I customized map size to 50x50. I lost the game. (More accurately, I abandoned it when it became clear that I was unlikely to win.) I then went through a dozen more attempts, with the map size always 50x50 or 50x60. I lost all of them. The AI was just beating me to some wonders (that I had gotten in my earlier victories), and it stayed much closer in techs. I led in techs, but the AI was only a few turns behind me.

I couldn't figure out why I had succeeded before and was failing now. But two days ago I started a game on a small map. It's not quite over, but I'm winning handily. I got all the wonders I was after, and I'm conquering with cannons against phalanx everywhere I go.

Since the only significant variant was map size, I've concluded that something about either a larger map, or perhaps a custom size map, makes the game more difficult. Is this consistent with anyone else's experience?

[Ming]

Actually, it's no surprise. On a small world, the AI's stop expanding sooner. They don't have as many cities, and your ten city perfectionist approach will always beat the snot of them. On a larger world, they can keep up better because they keep expanding, to the point where they have far more cities than you do.

Plus, isn't the science rate different depending on the size of the world

[Scouse Gits]

Some AI civs will always expand if there is enough territory. In the AI Headstart Succession Game, played on a giga map, I thought I had found the Russian capital. However, this Moscow (size 2), was the second time around the build list. They had 54 cities!

----------

SG(2)

[Campo]

That does make sense, but the effect seemed to happen so quickly. The AI would stay much closer to me (technologically) even early in the game, before they had time to expand much. I'd get beaten to either Colossus or Hanging Gardens, for example, which were my first wonders. At that point they probably didn't have any more cities than I did, and maybe less, because I put down those 10 cities fast.

In the normal size map, I get both of them built before anyone else, the last game before anyone even started working on them (at least there was no announcement).

I could see the expansion explanation if the effect kicked in after a few hundred game years, but it seems to start right from the beginning. You might be right but I'm not sure that's all there is to it.

[DaveV]

Here's my guess: the smaller the map, the more broken and varied the terrain. The AIs have fits building cities on a very small map, and also have fits producing much of anything from those cities. As the map grows larger, the terrain grows more homogeneous and the AI is able to cope better.

[Ming]

good observation DaveV... there is probably a lot of truth in what you say. The terrain on the real small worlds is a nightmare for the AI's to try to build on

[Scouse Gits]

I have played most games on Medium/Large/Giga Maps. Is the terrain that bad on the Small(er) Worlds?

I know that techs come faster on a Small World ... is there a formula for this?
------
SG(2)

[debeest]

Heh heh heh. Sure there is. It's just that nobody knows exactly what it is....

Actually, having finally and laboriously and proudly figured out a large chunk of the formula, and then having found that others had already gotten to the same point and no further, I'm kind of attached to the question. So here again is the part of the formula that I do know:

(beakers needed for next advance) = (# of next advance) x (some integer)

(free starting techs don't count toward # of next advance)

Simple, eh? But still too simple, since no one appears to have figured out exactly what determines the integer. Obviously it rises as the number of advances rises, but not tracking smoothly with it. It jumps substantially with the 20th learned tech (and also with the 40th, I think, but I haven't really checked this). It rises faster, the farther you're ahead of the AI in research; but even though everyone says to give all your tech to the AI to take advantage of that, I've seen no one quantify how much difference that really makes, or describe exactly how it works. My own experience has been that it takes an awful lot of giving to make any difference at all. But the best players all seem to agree on it.

Now that I've been pointed to the CivFanatics site with its exact formulation of the power graph determinants, I believe Civ players could find the exact beaker formula within a week if we all put our minds to it.

I almost always play mid-size worlds, and the beaker requirements generally progress something like 10, 20, 33, 48, 60, 78..... In my occasional small-world games, it's generally more like 10, 18, 27, 36, 50...... It's not a formula, of course -- it varies from game to game -- but it gives you a general idea of the degree of difference according to size.
[This message has been edited by debeest (edited March 27, 2001).]

[La Fayette]

quote: Originally posted by debeest on 03-27-2001 11:34 PM Heh heh heh. Sure there is. It's just that nobody knows exactly what it is.... ).]

debeest
What I have tested corresponds exactly to what you mention.
I tested early game on small maps and also some gift of techs.
1) early game (small map)
I suppose that the pattern would be 10,20,30,40,50 if the AI did not discover anything while you are researching.
Formula: (number of the tech you are researching)*(10)
But, as the AI discovers techs meanwhile, you get a bonus, such as:
2*9=18, where 2 means that you are researching tech #2 and 9 means that you have got a bonus for not being too far ahead of the AI.
2) gift of techs
I tested that using savegames of mine in the midgame.
Almost all the time, I got results showing that there was a penalty if I was ahead of the average of the AIs by more than 4 or 5 techs (or a bonus if I wasn't too far ahead - if you prefer).
Example:
24*25=600, where 600 is the number of beakers required to research tech #24 when I am 8 techs ahead of the average AI.
24*23=552, number of beakers required to research tech #24 after gift of techs and being 4 techs ahead of the average AI.
In the midgame the integer (25 or 23 in this example) is close to the number of the tech being researched (24 in this example).
I am afraid this is some kind of 'La Fayette's unfinished' (but I killed hundreds of spies a few weeks ago and my lab is going to remain closed a few more days).

------------------
aux bords mystérieux du monde occidental

[finbar]

Thank you, people, for confirming what I have long suspected. Small maps are good for techs.

------------------

[geofelt]

Your success with wonders may have much to do with the productive capacity of the city that is building one than the world size. I like to build many wonders, and the limiting factor seems to be production, not tech on any size world. If you can get trade early, and build a few caravans, it helps immensely. Gold from tribute can help a bunch. On a large world, you may not meet other civs early enough to trade for needed techs.
For some really ugly terrain and some fast games, try a micro(25x40) world.

[Dis]

25 X 40

actually I think I will try it tonight. I've been bored of civ, but I may like this ( I usually play on large maps)

you folks have been busy researching. I didn't think there was a bonus for small maps. I seemed to be having trouble because I couldn't create the # of cities I usually do, and I could not get ahead in tech. This was actually a medium sized map I think. So I will try a small one now.

[Deity Dude]

oops - sorry duplicate post

[This message has been edited by Deity Dude (edited March 29, 2001).]

[Deity Dude]

quote: Originally posted by La Fayette on 03-28-2001 10:08 AMAlmost all the time, I got results showing that there was a penalty if I was ahead of the average of the AIs by more than 4 or 5 techs (or a bonus if I wasn't too far ahead - if you prefer).

Does this apply to multiplayers games if u get ahead/behind other humans

[finbar]

quote: Originally posted by Deity Dude on 03-29-2001 03:36 PM Does this apply to multiplayers games if u get ahead/behind other humans

No. The software only compensates for the AI factor. And, from memory, it only does it on certain levels of play.

------------------
Founder, ACS Pedantry Institute
Founder, ACS Gourmet Recipe Exchange
Horse and Hydey Wrangler
Mono Rules!
#33984591