Vel's Strategy Thread - Part Three

[Arrian]

Re: Sad but true

Originally posted by pchang
If you really want to maximize your score, forget about building out your civilization. Just try conquering the world as quickly as possible. On Monarchy, I scored 9283 points by conquering the world (massive mounted warrior rush as Iriquois) in 190BC. I don't think you can beat that on Monarchy by maximizing your population and infrastructure.

Yet another example of my problem with the scoring system. Firaxis went out of their way to provide more options to the "builder" gamers, and then left the scoring entirely dependant on territory and population (with happiness factored in, of course). The earlier you conquer, the better. Guess what that means? The highest scores will be ancient despot rushes.

Not that it really matters, I suppose, but it irritates me. It will, of course, continue to irritate me, because it isn't going to change. No big deal.

-Arrian

[Aeson]

Actually scoring is heavily map dependant. The highest scores come from 2050 domination disabled games on huge landmasses, even beating out the early victory bonus on tiny maps. CivFanatics GOTM2 was a perfect example of this. It was a Monarch, Large Pangaea game with domination disabled. I built up to 2050 scored 12900 and came in third place behind 2 others who also built up to 2050 (15k, 13k). That was the first "bloated" game I had played (and for good reason, took about 80 hours to play through). Now that I understand how the scoring works in such scenarios, I could probably push 20k on the same map.

Population doesn't count nearly as much as the landmass you control. The earlier you conquer/settle the world (save for that last city) the higher your final score will be. I wouldn't be suprised to see a score of 50k+ score on a Huge/Pangaea/Deity game with domination disabled. To play out that scenario would take 100's of hours though, why I haven't finished any games of that type. The highest score by early conquest is 36060, and that requires an edited map so that conquest can be achieved by 3950BC (the second turn). I'm attaching a save game if anyone wants to see how that's done, or to further test the conquest bonus by date. Just delay building the first city until the turn before the date you would like to check.

All that being said, the scoring system certainly is out of whack. Early launches, cultural, and diplomatic victories have no chance to score as well as early conquests, or bloated 2050 scores (which can be any victory type). A better way to judge is by date for each individual victory condition. Of course how to compare the different victory types is much more difficult.

[pchang]

So what you're really saying is

Then the real way to maximize your score is to do the earliest possible despot rush until the AI is down to 1 city, and then settler bomb the map as rapidly as possible to get the entire map inside your cultural boundaries as quickly as possible. Then, maintain a sufficient millitary force to entirely block in the AI, set to fortify, automate all workers, allow your governors free reign, and then hit next turn over and over again until the game ends.

[Arrian]

Aeson - Ahah! I thought the map size had something to do with it! Now I don't feel so bad about my highest scores vis-a-vis those 15k ones. The map size factor must be tied to the territory scoring, which as you say is the biggest factor.

pchang - pretty much, yeah.

Thus, I will henceforth ignore score, so long as I achieve "The Magnificent"

-Arrian

[Hermann the Lombard]

Re: So what you're really saying is

Originally posted by pchang
Then the real way to maximize your score is to do the earliest possible despot rush until the AI is down to 1 city, and then settler bomb the map as rapidly as possible to get the entire map inside your cultural boundaries as quickly as possible. Then, maintain a sufficient millitary force to entirely block in the AI, set to fortify, automate all workers, allow your governors free reign, and then hit next turn over and over again until the game ends.

Correction! All true until "allow your governors free rein." Substitute "micromanage until you can't see straight anymore, then throw the computer out the window...but check your score first."

-- Hermann

p.s. That's a "horse" expression rather than a "monarch" expression, thus "free rein." Works the other way too, I suppose.

[gnomos]

Re: Sad but true

Originally posted by pchang
If you really want to maximize your score, forget about building out your civilization. Just try conquering the world as quickly as possible. On Monarchy, I scored 9283 points by conquering the world (massive mounted warrior rush as Iriquois) in 190BC. I don't think you can beat that on Monarchy by maximizing your population and infrastructure.

You can score a lot more points by maxing out the map. I'm on pace to score 17,300 in my current monarch game, and I'm only on 20% land mass. I think scoring 30K is definitely doable on monarch. I do think that like civ2, scoring is boosted by difficulty, the same population and land area would score more on diety.

[gnomos]

Re: Re: Sad but true

Originally posted by Arrian


Yet another example of my problem with the scoring system. Firaxis went out of their way to provide more options to the "builder" gamers, and then left the scoring entirely dependant on territory and population (with happiness factored in, of course). The earlier you conquer, the better. Guess what that means? The highest scores will be ancient despot rushes.

Not that it really matters, I suppose, but it irritates me. It will, of course, continue to irritate me, because it isn't going to change. No big deal.

-Arrian

Well, that's sort of correct as far as it goes, rushing early gets you a faster start. You need to control as much territory as you can for the longest time possible, since scoring is an average of all your turns. But ending the game early is not the best way to maximize your score. I think you can earn more points by maximizing the map than the early victory bonus yields. That's as it should be, maximizing the map is a heck of a lot of work.

[gnomos]

Originally posted by Aeson
The highest score by early conquest is 36060, and that requires an edited map so that conquest can be achieved by 3950BC (the second turn). I'm attaching a save game if anyone wants to see how that's done, or to further test the conquest bonus by date. Just delay building the first city until the turn before the date you would like to check.

3950 BC? Hmmm, how do you get a unit? Ahhh, ok, you have to play expansionist, join the worker, disband the explorer for shields, then pop rush the unit. And you have to start right next to your opponent, so you can conquer him on turn 2. So the map has 2 land tiles only, for your start location and his. You go first, so you get a city, and his settler sits there unable to do anything, because you can't found a city immediately adjacent to another city (this isn't documented, but I figured this out with testing).

[gnomos]

Re: So what you're really saying is

Originally posted by pchang
Then the real way to maximize your score is to do the earliest possible despot rush until the AI is down to 1 city, and then settler bomb the map as rapidly as possible to get the entire map inside your cultural boundaries as quickly as possible. Then, maintain a sufficient millitary force to entirely block in the AI, set to fortify, automate all workers, allow your governors free reign, and then hit next turn over and over again until the game ends.

I don't think it's that linear. If you conquer everyone, then tech development is slow, and I'm scoring more points from population than from land. You need aqueducts and hospitals early to score pop points. Instead, you want to trim your opponents, settle a lot of land while keeping them viable to do your research for you. You need early contact, dominate them through tech trading while taking a lot of land. Then when you get well into industrial and have 4 turn techs you whittle them down to 1 city.

I'm not sure if it's better to have like 8 opponents or 16. With fewer opponents I think you can grab more land early, but contact will be slower, and opponents will be harder to conquer since they've had mroe time to get established. With more opponents they will own more of the map early, but you can conatact them sooner, get tech trading going, and each nation will be smaller, easier to take out. At higher difficulty levels the AI research bonuses can make tech advance much more quickly, as the different factions research different tech lines at times, leading to multiple techs appearing at the same time.

[gnomos]

Re: Re: So what you're really saying is

Originally posted by Hermann the Lombard


Correction! All true until "allow your governors free rein." Substitute "micromanage until you can't see straight anymore, then throw the computer out the window...but check your score first."

-- Hermann

p.s. That's a "horse" expression rather than a "monarch" expression, thus "free rein." Works the other way too, I suppose.

yes, I'm afraid micromanaging is necessary, the AI doesn't gang up workers, it spreads them all over the place. You need to focus workers on tiles you need developed to help cities grow fast.

[Aeson]

Originally posted by gnomos
3950 BC? Hmmm, how do you get a unit? Ahhh, ok, you have to play expansionist, join the worker, disband the explorer for shields, then pop rush the unit. And you have to start right next to your opponent, so you can conquer him on turn 2. So the map has 2 land tiles only, for your start location and his. You go first, so you get a city, and his settler sits there unable to do anything, because you can't found a city immediately adjacent to another city (this isn't documented, but I figured this out with testing).

Actually if you look at the savegame I posted, all you have to do is build your first city to win. This is because the AI is defeated if they dont have a city and there is no valid terrain left to build a city on. The map has 3 non water tiles; 2 mountains, and one grassland (any terrain other than mountains will do) in a diagonal (from a grid perspective) line. I'm sure the way you described would also work, but it would take 1 extra turn. Conquest victories are awarded the turn after the last AI city/settler is destroyed.

As far as scores go, territory seems to be weighted heavier than population from my experience. The faster all the territory is claimed, the higher the score will be. I am basing this on the fact that my CivFanatics GOTM was played maxing population (no pop rushing after the initial 50 horseman army was built) at all stages of the game, while the two that beat me went for faster conquests (to build up to 2050 from). At the end of the game I was scoring more per turn than the others, as my earlier "average" didn't have to be pulled up as much. The game that won, had everything conquered about 100 turns before I did, using a straight despotic rush up to that point. So they were getting the "big" turn scores for a lot longer than I was, even if their early average had to be pulled up more. Other than that our games were almost identical, it was just the timing of the conquest that made a difference. Of course finding that "sweet spot" where conquest time isn't shortened too much, while population is maximized would yeild the best results.

[vmxa1]

Bravo

[Hermann the Lombard]

Re: Re: Re: So what you're really saying is

Originally posted by gnomos


yes, I'm afraid micromanaging is necessary, the AI doesn't gang up workers, it spreads them all over the place. You need to focus workers on tiles you need developed to help cities grow fast.

So what else is new in Civ?

I'm sure this has been suggested elsewhere, but it would sure be nice to be able to "group" or "gang" units to have them move and act together. I'm not looking for army bonuses, here, just a reduction in micromanagement when I have a stack going somewhere, or working somewhere.

[Arrian]

*Paging Vel, paging Vel. Vel to strategy forum, Vel to strategy forum.*

How's that compilation for part IV coming? Maybe one of us could help out, since you've dived into that Candle'bre project?

-Arrian

[Lawrence of Arabia]

bump