• Civilization 7 has been announced. For more info please check the forum here .

SGOTM 14 - Kakumeika

I did a quick check in the testgame and it seems AI traits are fixed. I world built in some great spies and used them to infiltrate the various leaders. This gave me city inspection powers and the expansive leaders still had +2 :health: and the creative leaders still had +2 :culture: and the charismatic leader still had a free +1 :) so I suspect we can safely assume that the AI leader traits are the same.

Thanks for quickly verifying this! Without espionage it may not have been quite so easy to do.

With that settled we may know enough about Random Personalities to get by. I still wonder whether more information about it would be helpful. Certainly stratagems for determining an AI's true personality are lacking and we need more information on this topic. I'm not sure we can wait until we get enough espionage to inspect AI cities to help us in this quest.

Note that whenever we switch civics we will know which AI have it as a favorite civic, since we will see a +1 for Favorite Civic in Diplomacy displays and this will help us identify its personality. For example, there are four AI leaders that have Organized Religion as a Favorite Civic. If any AI gives us a +1 for switching to Organized Religion when they are also running it, we will known their personality must match one of these four, Hatshepsut, Ramesses II, Suryavarman II or Brennus.

Open Borders is another way to tell. Any Civ not willing to Open Borders upon first contact may have the personality of Tokugawa or some AI similar to him. This one needs further research though.

Sun Tzu Wu
 
Note that whenever we switch civics we will know which AI have it as a favorite civic, since we will see a +1 for Favorite Civic in Diplomacy displays and this will help us identify its personality. For example, there are four AI leaders that have Organized Religion as a Favorite Civic. If any AI gives us a +1 for switching to Organized Religion when they are also running it, we will known their personality must match one of these four, Hatshepsut, Ramesses II, Suryavarman II or Brennus.

By regularly monitoring the diplomatic relations between the AI we can also guess their favorite civics. This can be seen on the F4 "glance" screen.
 
summary of differences T74 GLH no pyramids bcool game vs. T74 GLH & pyramids frogdude's game
5 workers vs. 4 workers
5 cities vs. 4 cities
1 settler moving vs. 1 settler almost finished
18 pop vs. 15 pop
2 lighthouses vs. 2 lighthouse
1 granary vs. 2 granaries
more roads, gems hooked up, ivory hooked up in 2 turns, workers set to improve gold on border pop vs. limited roads, still need to hook up gems, ivory 10+ turns away, workers not near gold for anticipated border pop
writing finished & slightly more gold vs. writing finished & slightly less gold

and given Walter Wolf's wonderful research, The earliest pyramids goes is T95 with industrial AI -- I assume that means the AI had stone, but I'm not sure.
Regardless, I think we can delay the pyramids, and delay the forest chops into the pyramid until we get mathematics.

I had another game that was about trying to see the impacts of getting wheel + pottery before sailing + masonry, and driving the economy with the gems.
One of the advantages from this is that I was able to build a road on the tile NW of capital, which speeds up the arrival of the settler by a turn. Similar the two roads NE of capital (and a chop) do the same for the stone city.

I'm much more pleased with the results of this than my previous run, but still feel lots of improvement is possible. I stopped this one at turn 77:
  • Major downside: I didn't get the Great Lighthouse build until turn 77.
  • Tech front, I had Wheel, Pottery, Sailing, Masonry, Writing, and half of mathematics.
  • Much less deforestation.
  • More workers & warriors.
  • 2 libraries, 3 granaries and 2 lighthouses built
  • a third of the pyramid built

There were some awkward early worker timings and a lack of things to build at crucial overflow points as a result of this plan. Still, the injection of more whipping hammers from the granaries was showing good yields. I think experimenting with Wheel -> Sailing -> Pottery -> Masonry -> Writing may be worthwhile. Also, Hunting should be considered in there.
 

Attachments

  • Sgotm14 testgame Turn77.CivBeyondSwordSave
    118.4 KB · Views: 47
Hehe, it is true that slavery isn't amazing before granaries, but I still feel it has some uses.

To grow from 1 to 2 Population takes 22:food:
To grow from 2 to 3 Population takes 24:food:
To grow from 3 to 4 Population takes 26:food:
To grow from 4 to 5 Population takes 28:food:


1) Quick barb defense. "OMG barbarian walked out of the one fogged tile and is headed for my corn" is a cry heard in many Civ 4 games. Slavery is a great counter. It is less dramatic on normal where a unit can be built in 2 or 3 turns, but it still works.

In our game, with good fogbusting on our peninsula and normal speed, barbs should not be a problem.
So point to Mabraham against slavery.

Slavery does allow rapid conversion of food to hammers, and where the alternative costs a lot (like an important improvement being pillaged) then clearly you want to be able to whip.

2) Whip a monument. If the first thing you build is a warrior or monument, whipping that monument when you grow to size 2 turns 22:food: into 30:hammers: instantly. It also lets you grow to size 2 again on an improved food tile while the worker is busy improving the 2nd tile. You get the monument and border pop sooner rather than later.

Yep, an early 1-whip followed by a 2-whip of a granary is normally good management. A worker supplied by the empire is obviously best for kick-starting things.

3)A new city can quickly create a military unit and a worker if it has a good food tile.

Start the city building a unit while growing to size 2, then switch to a worker at size 2. Use the first worker to improve the food and then chop a forest. When the +20 hammers goes into the city, whip the worker complete using 1 population. The #2 city can have another worker and unit out very fast using this method, and it quickly grow back to size 2.


4)Build granaries the turn after researching them :)

Yes, that's another example. Of course, completing a granary is only a high priority if the food box is just under half full. Then it has time to fill during the subsequent growing turns. Completing the granary earlier than this offers no advantages - you may as well not have finished the granary yet, as have it fill early and have to wait for the food box to fill. You still might wish to whip the granary, but doing so immediately might not offer an advantage over a slightly later whip.

5)Whipping at 4 population with a full food bin can also be useful. A 2 pop whip at 27/28 food will reduce the city to size 2 and change the food bin to 27/24 food. The next turn you are size 3 and well on your way to size 4 already thanks to the food overflow. This is mainly useful if the workers are not improving tiles fast enough.

Yes, but had you not whipped you'd be at size 5 the next turn. If the fourth and/or fifth population are not working good tiles (e.g. they are working Gforest) then one may as well whip them.

The main thing is, you can have stuff now rather than later. Sometimes such a thing cannot have a hard value assigned to it. How do you quantify an axeman army taking an enemy's capital the turn before they get longbows thanks to slavery?

Yup, tough.

A good question to ponder is how much natural production a city can have to outshine slavery both before and after granaries. Is it +12:hammers: per turn? +16:hammers:?

Whips are most efficient at quick and marathon when whipping down to even city sizes, but we're at normal.

Otherwise, I think this question should be addressed in terms of the food-to-hammers conversion rate of a given city size, rather than in terms of the raw hammer output of the unwhipped city. Food is the basis of the whole multifaceted Civ4 economy.

Lets consider a candidate city that is at size 6 working a wet grassland corn (6:food:), 4 Gmines (each 1:food:3:hammers:) and 1 Pmine (4:hammers:). Including the central tile, that's sustained 12:food:17:hammers:1:commerce: per turn - however it is the hammers that interest us here.

Now suppose we wish to sustainably whip military units using 2-population whips every 10 turns, regrowing working the corn, a stray Gforest and then the Gmines. (Doing sustainable 3-population whips would require higher base food.) Starting from a turn upon which we whip with a half-full food box (i.e. 16 food) and assuming we have a full granary, the city looks like
Code:
Turn Size Food Hammer Food Net
          Rate Rate   Box  Hammers
0    4    12    8     20    68
1    4    12    8     24    76
2    4    12    8     28    84 (growth occurs after tiles are worked)
3    5    13   11     14    95
4    5    13   11     17   106
5    5    13   11     20   117
6    5    13   11     23   128
7    5    13   11     26   139
8    5    13   11     29   150 (growth occurs after tiles are worked)
9    6    12   17     17   167 (grow onto Pmine, switch Gforest to Gmine)

So this city is a break-even proposition for 2-population whips from size 6 to 4. Key in this result is that working a Gmine provides 1 food, and converts a food produced elsewhere into 3 hammers. This efficiency is needed to overcome the need to work a Gforest just to get food into the system. If more Pmines were being worked, you'd need another food supply.

I can't see the above does much more than point out that a city needs to be more efficient than 4 Gmines and 1Pmine plus supporting food to benefit from not whipping.
 
One of the biggest differences in mine compared to STZ, is that I think the stone/crab/gold site should be settled first. There are a few reasons for this.

This city can immediately start working the oasis. It provides a not insignificant boost to our research. It is also easier to establish a road link to this site if we research the wheel first. (Which by the way doesn't slow down the The GLH because we can whip the workboat and the excess hammer go into the lighthouse when we finish sailing). It can probably finish the GLH faster than the gems site can if we build the stone quarry quickly and aggressively chop forests near the city.

The game just seems to flow better by settling this city first. Even if we don't go for the great wall or Stonehenge.

I just don't see that a stone city first can be better. The wheat site gets to work the wheat and gems at size 2 building lighthouse->TGL straight out of the box without researching Wheel, and with the deer camp coming in with Hunting. Stone has to get itself a workboat and masonry to get into the game at all. However, I'll test it.

Frogdude pointed out that the stone site gets a Sailing trade route clockwise round the island, and so does the wheat site when we clean out the fog down south with Columbus.
 
Well, you answered the question yourself. I guess you were right about industrious + required resource gave the fastest times. Exploring all of that would require a whole week of playing :D.

Settle, set to build poly-archer and press Enter to end turn 100 times, occasionally fortifying a warrior sounds like it should take only a couple of minutes per run through!
 
I just don't see that a stone city first can be better. The wheat site gets to work the wheat and gems at size 2 building lighthouse->TGL straight out of the box without researching Wheel, and with the deer camp coming in with Hunting. Stone has to get itself a workboat and masonry to get into the game at all. However, I'll test it.

OK, I did. Have chopped stone and capital clean (apart from spices and silk). Teched Sailing, Masonry, Wheel, Hunting, Pottery, Writing, Maths. Should probably have settled SE of cows fourth, rather than SW of pigs. Switched to slavery while first settler was walking. 2-whipped subsequent settlers.

T72 TGL done in stone city - could maybe be a turn or two faster
T74 4 cities 10 pop 4 workers, just done Writing, sustainable beakers per turn 22
T78 4 cities 11 pop 4 workers, 21 bpt
T82 5 cities 17 pop 4 workers, half of Maths, Pyramids done T82 (could definitely be a turn faster), bpt 31 (granaries have kicked in, libraries are finishing)

I'd been whipping settlers and workers overflowing onto the Pyramids (mostly with stone), but I could certainly have been doing that onto a granary for much of the time, and then mass-chopped the Pyramids around T88 hopefully with maths.
 

Attachments

  • Sgotm14 testgame BC-1040 T74.CivBeyondSwordSave
    117 KB · Views: 36
  • Sgotm14 testgame BC-0925 T78.CivBeyondSwordSave
    124.9 KB · Views: 39
  • Sgotm14 testgame BC-0825 T82.CivBeyondSwordSave
    128.9 KB · Views: 44
Earlier I was talking about scouting the coastline by looking back. The attached screenshot illustrates this. In order to scout the western coastline while generally heading north, the warrior can move 1N. He does not need to go 1W.
 

Attachments

  • scouting coastline looking back.JPG
    scouting coastline looking back.JPG
    98.4 KB · Views: 86
Now suppose we wish to sustainably whip military units using 2-population whips every 10 turns, regrowing working the corn, a stray Gforest and then the Gmines. (Doing sustainable 3-population whips would require higher base food.) Starting from a turn upon which we whip with a half-full food box (i.e. 16 food) and assuming we have a full granary, the city looks like
Code:
Turn Size Food Hammer Food Net
Rate Rate Box Hammers
0 4 12 8 20 68
1 4 12 8 24 76
2 4 12 8 28 84 (growth occurs after tiles are worked)
3 5 13 11 14 95
4 5 13 11 17 106
5 5 13 11 20 117
6 5 13 11 23 128
7 5 13 11 26 139
8 5 13 11 29 150 (growth occurs after tiles are worked)
9 6 12 17 17 167 (grow onto Pmine, switch Gforest to Gmine)
So this city is a break-even proposition for 2-population whips from size 6 to 4. Key in this result is that working a Gmine provides 1 food, and converts a food produced elsewhere into 3 hammers. This efficiency is needed to overcome the need to work a Gforest just to get food into the system. If more Pmines were being worked, you'd need another food supply.

I can't see the above does much more than point out that a city needs to be more efficient than 4 Gmines and 1Pmine plus supporting food to benefit from not whipping.

Just to clarify your point for myself ...So you get 167 net hammers by 2 pop whipping every 10 turns, or you could produce 170 net hammers working the tiles for 10 turns. The benefit of whipping would be you would get the military units a few turns earlier. So in terms of just producing hammers and units whipping is slightly better for this city. Anything more productive hammer wise than this city might not want to be whipped.

The situation gets more complicated when you factor in the reduce maintenance costs by whipping and the possible extra commerce not whipping could produce if the city was on a river or otherwise getting commerce.
 
T72 TGL done in stone city - could maybe be a turn or two faster
T74 4 cities 10 pop 4 workers, just done Writing, sustainable beakers per turn 22
T78 4 cities 11 pop 4 workers, 21 bpt
T82 5 cities 17 pop 4 workers, half of Maths, Pyramids done T82 (could definitely be a turn faster), bpt 31 (granaries have kicked in, libraries are finishing)

So are you concluding that it is easier to get the GLH built in the stone/crabs/gold site? I think it has numerous advantages that I didn't fully describe.

1) I think it is faster -- although this isn't that important unless we think we are going to get close to AI dates
2) The GLH Culture gives us the gold without having to worry about a monument
3) It saves forests around the gems/deer/wheat site for the National Epic or the Great Library.
4) It leaves open the possibility of the Great Wall

Other comments
1) As you suggest or imply: I think we do not want to put any hammers in the pyramids unless we have stone, and I think we want to wait until we have mathematics to chop any forests into the pyramids. We can finish the pyramids before the earliest dates that AI usually finishes it.
 
Major downside: I didn't get the Great Lighthouse build until turn 77.
Tech front, I had Wheel, Pottery, Sailing, Masonry, Writing, and half of mathematics.
Much less deforestation.
More workers & warriors.
2 libraries, 3 granaries and 2 lighthouses built
a third of the pyramid built

As I implied for mabraham's tests, I don't think delaying the Great Lighthouse is that big of problem unless we expect there is a significant chance we lose the race. So maybe teching roads and pottery first is ultimately better. Of course with that said, in Walter_Wolf's tests the AI do sometimes built it earlier than that. We might want more tests similar to Walter_Wolf to see how big a risk we might be taking by delaying the TGL.

The TGL is a little less useful on a hub map than other maps. There are likely no islands (which means a lot of our trade routes will only be 1 for quite some time. So how much do we want to sacrifice for it? Or how aggressive should we be to build it asap? (Assuming it is a hub map, and assuming that it isn't significantly modified...)
 
Conclusion: We don't understand Random Personalities as well as we should.

Sun Tzu Wu


It is true Tachywaxon mentioned CIV4LeaderHeadInfos.xml values before. If "Random AI" is indeed purely a shuffle of .xml values, then careful attention to AI attitudes and eventual favorite civics will be our only clues.

I found this thread last night that mentioned such .xml values:
http://forums.civfanatics.com/showthread.php?t=161282

The information relevant to us was posted on page 2 by Ori:
http://forums.civfanatics.com/downloads.php?do=file&id=3359


I do not know how to interpret all those numerical values that make up the personality of the AI. Maybe you guys know. :dunno:
 
Settle, set to build poly-archer and press Enter to end turn 100 times, occasionally fortifying a warrior sounds like it should take only a couple of minutes per run through!

Being a civfanatic, I couldn't stand looking at that unimproved corn. The next thing I know, I was building a cottage ... :D
 
I've run a few test games. This is the only one that was interesting. I stopped at T80. This is a GW, TGL, Pyramid test.

T63 Built GW in Stone City.
T80 Built Pyramids in capitol. Built with chops and working tiles (poorly).
Due on T86 TGL in Stone city. (I played it out and did get it.)

At T80
5 Cities
18 Pop
One settler 3 turns away (or a two pop whip away.)
Mathematics 98/390:science:

Research: Wheel->Masonry->Sailing->Hunting->Writing->Pottery->Mathematics (Partial)
Initial build from T33: Settler->Worker->Settler->warrior...

I settled Deer Gems city first, then Stone City. Sailing trade routes were not available, so roads were built immediately to Stone City site. The PHF NW of capitol chopped, then roaded, then worker went to Wheat, while microing a road on tile between wheat and gems. Then on to gems. Stone was hooked up ASAP and GW chopped after a 1 pop WB whip in Stone city.

I went back to T74
Pop 16
5 Cities
4 workers
1 Lighthouse, no granaries
 

Attachments

  • Sgotm14 shulec testgame Mids T80 BC-0875.CivBeyondSwordSave
    126.9 KB · Views: 42
So are you concluding that it is easier to get the GLH built in the stone/crabs/gold site? I think it has numerous advantages that I didn't fully describe.

1) I think it is faster -- although this isn't that important unless we think we are going to get close to AI dates
2) The GLH Culture gives us the gold without having to worry about a monument
3) It saves forests around the gems/deer/wheat site for the National Epic or the Great Library.
4) It leaves open the possibility of the Great Wall

IIRC some early tests I did build TGL at the wheat site T72, so I am not necessarily concluding that either is faster. The wonder culture is more useful at the stone site. Deforestation is moot - the site that doesn't build the first wonder still has its forests. GW is a sound upside of Wheel->Masonry and settling stone site first. Ideally we sneak in Pottery too... I'll run such a test.

Other comments
1) As you suggest or imply: I think we do not want to put any hammers in the pyramids unless we have stone, and I think we want to wait until we have mathematics to chop any forests into the pyramids. We can finish the pyramids before the earliest dates that AI usually finishes it.
 
Deforestation is moot - the site that doesn't build the first wonder still has its forests.

I don't think this is moot since the stone/crabs/gold site doesn't make a good great person farm while the gems/wheat/crabs/deer site does. So saving the forests around the gems/wheat/crabs/deer for the National Epic or other more :gp: intensive wonders like the Great Library is more useful.
 
OK, I tried going for GW, TGL and Pyramids teching Wheel->Pottery->Masonry->Sailing->Hunting->Writing->Maths->Currency. This requires settling a bit later in order to get more tech done before things start to slow down, however you can use the time to get another worker and get some roads towards the stone and wheat sites, so some of the settler walking time is saved. Settled stone then wheat.

I got the three wonders T62, T86 and T92 respectively. Stats:

T74 5 cities 19 pop 4 workers done Writing, 23 bpt, 3 granaries, 2 lighthouses
T82 5 cities 20 pop 4 workers half done Maths, 23 bpt, 4 granaries, 2 lighthouses, 1 settler
T87 done maths, finish chops into Pyramids
T92 7 cities 26 pop 4 workers half done Currency, 33 bpt 5 granaries 3 lighthouses 3 libraries

At T82, this is a granary, 2bpt and 3 pop ahead of my #587 test games, two wonders behind, and several forests still unchopped.
 
I don't think this is moot since the stone/crabs/gold site doesn't make a good great person farm while the gems/wheat/crabs/deer site does. So saving the forests around the gems/wheat/crabs/deer for the National Epic or other more :gp: intensive wonders like the Great Library is more useful.

Fair enough
 
I've run a few test games. This is the only one that was interesting. I stopped at T80. This is a GW, TGL, Pyramid test.

T63 Built GW in Stone City.
T80 Built Pyramids in capitol. Built with chops and working tiles (poorly).
Due on T86 TGL in Stone city. (I played it out and did get it.)

Excellent ETAs for these three Great Wonders.

It is generally better to built The Great Wall in the Capital, because the Palace generates 4 Ept that can be multiplied with any settled Great Spies via Scotland Yard and other espionage multiplying buildings. Because of this extra 4 Ept in the Capital, it is usually the best Espionage City.

Again, great dates for the three great wonders and respectable REX of five cities considering that.

Sun Tzu Wu
 
I my tests, I did see a game where the Great Wall went around T64. If we commit to the Great Wall, one turn could be a difference maker. mabraham's test of T62 vs my T63 could be significant. If you look at WW's test games, the average was T52.

I don't have much experience with the Great Wall and running espionage economies. I will have to defer to those of you who regularly use espionage. I know Kaitzilla is a master spy, are any of you other guys big into espionage?

I have no opinion as to whether or not we need to prioritize the Great Wall. I will defer to you guys. If we decide we need it, I don't think it matters if we settle on the stone or gems first. But I believe we seriously risk losing it if we do not research wheel->masonry from T33.
 
Top Bottom