SGOTM 13 - Gypsy Kings

Hi all, checking in, just got back from a conference in San Antonio.

First primitive thoughts before I read what is here ... seems either settle in place for a super GP farm (move the capital later) [but rules out the west island for a second city], or on the plains hill for just a great GP farm (faster early production) [waste a forest]. Also, do we really think 3 WB beats a worker first (farm plus mines early could get the WB done in fewer turns, should be an easy test if not already decided)

dV
 
I had a go at Oracle + Pyramids on my second map. There's a window of 25 turns after we get the Oracle in which the second city must complete a Forge (and preferably be at size 2 to keep growing while running the Engineer for the next 50 turns), else we pop the Great Prophet first. Also we need to be as quick as possible, because (say) T95 oracle + 20 forge + 50 Great Engineer + several to get him back home is dangerously many turns. So that means at least three 30:hammers: chops, at least one 45:hammers: whip and another source of :hammers:. That means the capital has to put out several more workers ASAP after Oracle. The good news is that if we fail the Pyramids, then Colossus is still there and we have not yet committed our strategy either way. And, if we Oracle MC and trade it around freely, AIs may start Colossus rather than Pyramids.

Anyway, settling the second city on a decent-sized island with some forests needed a warrior escort (we should check when barbs will enter our borders, though) so that delayed Oracle until T97. Then I got Pyramids on T170. The Industrious AIs had 27 and 14 turns to go on it, but they didn't have stone. I think this was a lucky result, but we need to gather some data. I could probably have shaved a few turns with better management of the second city.
 
I had a go at a pure capital-pyramids SIP strategy trying to whip food-things into overflow-:hammers:. It was doable on my first test save, but my first attempt was not much faster than the above. I got Pyramids T161 (185BC).

I organized the 4 WBs, worker, corn mine, monument, lighthouse and galley by T71, and overflowed 19:hammers: onto the 'mids, whipping down to size 3. I then re-grew to 6, put 3 turns on a settler, did a 3-pop whip overflowing 30-44:hammers:. I repeated this while the two mines and :health: trades gradually came online, which increased the rate of return. I chopped the 3 forests when they made sense. At the end, I did a 2-pop whip on a settler followed by another 2-pop to finish. Neglecting the chops, the next table shows my :hammers: count after overflow accrued, the :hammers: increment since last datum, and details of the final push:
Code:
Turn   :  72  87 102 117 132 147 (3 chops) 158 159 160 161
Hammers:  19 117 215 316 429 542       632 654 654 706 751
Incr.  :  19  98  98 101 113 113        90  22   0  52  45

So I started off doing 6.5:hammers:/turn and ended more like 7.5:hammers/turn:. (Check: 7*90+3*30 = 720, which is about right) Meanwhile, I'd also built 6 settlers! So total output was effectively 17.1 :hammers:/turn, not counting chops. Not bad for a city with two hills!

My first two cities were on 1-tile islands with access to single seafood, since I had no defense for settling on larger things. Those cities organized whipping some warriors and WBs so I could start defending other places. My galley and worker were ridiculously busy, of course! The point at which we trade for archery makes a big difference to the defensibility of the later cities. By T155, I was seeing barb axemen, and settling the 6th settler was impossible. Perhaps two workers will be a better idea than a settler at some point.

The second city was close enough to share the GHmine while Paris was re-growing, which was good for it.

I do not think a granary is necessary for this strategy. It does speed the re-growth, so you get to spend more time building stuff at your maximum size. That might let us 2-pop-whip some settlers and thus sneak the odd worker out next cycle. However it costs 90:hammers: or 15 turns before we get going, and I don't think that's affordable.

Erratum - I posted this stuff below before I realized that for a 4/2 cycle we need to put just below 104:hammers: on the settler (not just below 59:hammers:) before a 2-pop whip, and this is not doable fast enough. A 4/2 cycle can't work better than 6/3. It's hard to see us having enough happiness to make 8/4 workable - but a granary would surely be required.

It's conceivable the cycle could work between sizes 4 and 2. In this case, the monument is not required, so we could get started earlier. Regrowth will be several turns faster (less food needed per population, one fewer population point) but the most we can put on a settler is 15/turn. Towards the end of the 6/3 cycle, I was putting 19/turn on a settler, which gets us to a near-ideal 57:hammers: in 3 turns. With the 4/2 cycle, we'd have to do 15-15-14-14 before whipping, so that's a turn slower (but perfect!). We'd still be working the two mines in the gap between regrowth and settler building - the question is whether on average we work the mines for the 'mids more often than we do for the 6/3 cycle. My gut feeling is that we do, and this should mean faster 'mids for the 4/2 cycle. However, the final whip-push will not be as effective.
 
Can you explain further? IMO, growing unhappy can be OK (all it costs is 2 per ,
I was just talking about growing into RED faces....not over whipping! I used to worry about growing red faces, but C63 convinced me that they do no harm, and eventually get converted to hammers anyway!
 
So I started off doing 6.5:hammers:/turn and ended more like 7.5:hammers/turn:. (Check: 7*90+3*30 = 720, which is about right) Meanwhile, I'd also built 6 settlers! So total output was effectively 17.1 :hammers:/turn, not counting chops. Not bad for a city with two hills!
That is outstanding output!
(I need a special Devil's Advocate hat, this perhaps => :satan:) Imagine however, if in addition to those 6 settlers we also had 750 :hammers: worth of galleys, warriors workboats and a granary. Eg, 3 galleys, 6 warriors, 6 WB, granary.

Ok... back to the Oracle idea - CoL vs MC:
We are Organised right? Cheap courthouses. it strikes me that a better barb defense than triremes is quality fogbusting. And what better way to fog bust than to build cities. Code of Laws enables two excellent methods for continued expansion: 1) cheap courthouses, 2) Caste system merchants/scientists. 1) is great for the whipping phase, 2) for after that.

Another thought - CoL so fast will almost certainly give us the confu religion. What sort of whipping output might we get if we had the :) to grow up to 8 each cycle?

Edit: I tried to investigate the 8/4 whipping cycle for fast expansion using a spreadsheet. Starting from size 4 with a granary and lighthouse, in 16 turns the city can produce 112 :hammers: for whatever, 1 worker and 1 settler. Works out at 22 :hammers:/turn total. I think I might be looking too far to the future, but that's pretty exciting :D
 
@mabraham :goodjob:

excellent work. I will have to give it a few goes before I can comment intelligently.

If I understand correctly I think going for the forge generated Great engineer is safer than whipping into the pyramids. In both cases there is some chance we fail, but with the Great Engineer we have can use him for something else.

Maybe we can whip into the Oracle get several settlers without slowing it down too much, plan to have 2nd city at 8 pop for immediate whip of forge? (probably not doable but worth investigating)

I won't have much time today to do real testing. I will dig deep into mabraham's results later this week.
 
The second city was close enough to share the GHmine while Paris was re-growing, which was good for it.

Bah. I should have thought to put the second city on the same island. That way, it can work the fourth food for the 3 or so turns per cycle the capital is at size 3, and the PHmine for some more turns, its own GHmine all the time, and random coast for the rest. Maybe it can run its own 4/2 cycle on workers and warriors/workboats.

Moral: don't play Civ after midnight :eek:
 
(I need a special Devil's Advocate hat, this perhaps => :satan:) Imagine however, if in addition to those 6 settlers we also had 750 :hammers: worth of galleys, warriors workboats and a granary. Eg, 3 galleys, 6 warriors, 6 WB, granary.

It is situation-dependent. If we have a pile of small islands each with 1-2 seafood (e.g. Grifftavian's test save), then the situation is ideal for the Pyramids cycle. Each new settler gets marooned and fends for itself. The first couple over-whip to get out warriors and WBs for the next couple. They'll get happy again during the Caste phase, no big deal.

If we have a bunch of large islands/continents (e.g. my first save) then we need more infrastructure support (workers, warriors). This is attenuated if the land is jungle and we can't work it until after we trade for IW - all we're doing is staking a claim with the settler. Then again, if the second city is on our (sic) mainland, it can do a fair job of getting out infrastructure support, because it has access to improved :hammers:-tiles that new cities won't.

One defect of the Duckweed-Pyramids is that we have time during the building to tech CoL or MC, but maybe not both. Barb galleys seem to start arriving from about T150, and they're worse in the many-islands scenario that suits Pyramds, per above. That suggests MC first, but OTOH, doing CoL second risks losing the religion, and risks some mad religious zealot (yes Izzy, I'm looking at you) starting Philo and jeopardising our run at Liberalism.

Ok... back to the Oracle idea - CoL vs MC:
We are Organised right? Cheap courthouses. it strikes me that a better barb defense than triremes is quality fogbusting. And what better way to fog bust than to build cities. Code of Laws enables two excellent methods for continued expansion: 1) cheap courthouses, 2) Caste system merchants/scientists. 1) is great for the whipping phase, 2) for after that.

Yeah, I did some of this following on from my early Oracle tests. I was getting the Drama/Lib(Nat)/Gunpowder/Optics trees all teched at around 1000AD with a marginal navy built from 7 or so cities. That doesn't feel fast enough.

Another thought - CoL so fast will almost certainly give us the confu religion. What sort of whipping output might we get if we had the :) to grow up to 8 each cycle?

Edit: I tried to investigate the 8/4 whipping cycle for fast expansion using a spreadsheet. Starting from size 4 with a granary and lighthouse, in 16 turns the city can produce 112 :hammers: for whatever, 1 worker and 1 settler. Works out at 22 :hammers:/turn total. I think I might be looking too far to the future, but that's pretty exciting :D

From 8/4, assuming sufficient :)/:health: and granary, one gets 180:hammers: from the whip, regrowth at (effectively) 21/23/24/26 with :food: excesses 14/13/11/11 working the four food tiles plus nothing/GHmine/both mines/mines+coast, so growing at 1.5/1.8/2.2/2.4 turns. Call that 8 turns. During growth, the GHmine got worked for 6.5 turns and the PHmine for 4.5 turns, so 20+18:hammers:. Then there's 7 turns working the mines for 7:hammers: while building stuff. We work the central :hammers: tile throughout. So every 15 turns we produce 180+38+49+15=282:hammers:. So I estimate at most 19:hammers:/turn. We do get some dribbles of :commerce: from the 7th and 8th people working coast tiles. We do continue to grow through the 7 turns (16:food:/turn), so we'll snowball faster regrowth and eventually a 5-pop whip - on what?. It's better to channel that food into workers/settlers. This increases the effective :hammers:/turn somewhere higher than 19... so I'm roughly agreeing with that 22:hammers:/turn estimate - perhaps it's a bit low.

The marginal return from whipping cycles should decrease with increasing population, because the re-growth phase takes longer for more food per population, and so the conversion of :food: to :hammers: is less efficient. It's worse in our case because our 5th to 8th population are not working great tiles, and 7 and 8 are really somewhat poor.

For comparison, a 6/3 cycle with granary has 135:hammers: from a whip, regrowing at 20/21/23 with :food: excesses 11/14/13, so growing at 1.8/1.5/1.8 turns. Call that 5 turns. We got 2 turns on the GHmine for 6:hammers:. Then we work 10 turns on the four food and two mines for 70:hammers:. We get 15 from the central tile. Total 135+6+70+15=226:hammers:, so 15/turn. Plus food excess of 15/turn during a longer post-growth phase. So the 6/3 cycle is somewhat less effective in total, but requires less :)/:health: support, starts faster, and cooperates much better with a second city sharing tiles (because the second city gets a food tile for 2 turns/cycle).

I'm pretty confident a 6/3 Paris and 4/3 Orleans will outperform an 8/4 Paris and 2/1 Orleans.

Obviously, I'm now a convert to the "second city on mainland" idea :)
 
About the 8/4 cycle:
Spoiler :
From 8/4, assuming sufficient / and granary, one gets 180 from the whip, regrowth at (effectively) 21/23/24/26 with excesses 14/13/11/11 working the four food tiles plus nothing/GHmine/both mines/mines+coast, so growing at 1.5/1.8/2.2/2.4 turns. Call that 8 turns. During growth, the GHmine got worked for 6.5 turns and the PHmine for 4.5 turns, so 20+18. Then there's 7 turns working the mines for 7 while building stuff. We work the central tile throughout. So every 15 turns we produce 180+38+49+15=282. <snip>
It's better to channel that food into workers/settlers. This increases the effective /turn somewhere higher than 19... so I'm roughly agreeing with that 22/turn estimate - perhaps it's a bit low.
I assumed only the health resources we could see + granary + lighthouse and enough happiness (and zero initial whip :mad:) to be happy at size 7 (size 8 being unhappy), ie, requiring only either one happiness resource or a religion.
My thinking was that just before we grow to size 8 we build/invest into a worker, so that excess food was also being converted to hammers.
Total :hammers: = 112 + 90 from worker + 150 from settler = 352 = 22/turn.


Important: Turn 0 Decision!
A second city on the mainland is directly at odds with settling on the plains hill. This is a decision we need to make on turn 0.

If we are aiming for an early wonder build in the capital, then I think it is superior to settle city #2 on the starting landmass so it can contribute to our expansion ASAP. Perhaps we can have as many players as possible voice their opinion on this option?
 
Important: Turn 0 Decision!
A second city on the mainland is directly at odds with settling on the plains hill. This is a decision we need to make on turn 0.

If we are aiming for an early wonder build in the capital, then I think it is superior to settle city #2 on the starting landmass so it can contribute to our expansion ASAP. Perhaps we can have as many players as possible voice their opinion on this option?

I agree that this is a very important decision. I think the approach that got T85 Oracle settling on the PFH (with no galley or settler) sacrifices too much empire-expansion for certainty of the one-off pile of :science:. It's far from clear to me that the free :hammers: from settling on the PH outweighs the ability to work the PFH in the early game - it's our best :hammers: tile. (But please do show me why I'm wrong! :)) Once we have two mines up and worked (near) full-time, both SIP and settle-PFH return 8:hammers: from mines+central tile, so long-term I don't see a difference.

IMO, the key advantage of SIP is that it allows for one city to use all 4 food resources when we want it to (wonder push, Globe drafting). Secondarily, it allows for more effective tile sharing (SIP shares PHmine, corn and clams with settle-3E; settle-PFH at best shares one clam and an improved grassland with settle-2W). Settle-PFH and settle-W-island are probably comparable with SIP and settle-3E except in these regards.

So I vote SIP and plan for a second "mainland" city, probably 3E unless we happen to learn interesting things up north.

However, as previously discussed, if our warrior sees seafood from PFH, then the decision needs to be reconsidered. I think our esteemed captain should move the warrior the instant we get the save :)

I might be biased - who wants to make the case for settle-PFH?
 
I think we have to wait until we move the warrior but I definitely think a 2nd "mainland" city and SIP is the way to go currently. I can be moved to change my mind with evidence, test games etc. and in game developments.
 
I think we have to wait until we move the warrior but I definitely think a 2nd "mainland" city and SIP is the way to go currently. I can be moved to change my mind with evidence, test games etc. and in game developments.

About the 8/4 cycle:
snip...
Important: Turn 0 Decision!
A second city on the mainland is directly at odds with settling on the plains hill. This is a decision we need to make on turn 0.

If we are aiming for an early wonder build in the capital, then I think it is superior to settle city #2 on the starting landmass so it can contribute to our expansion ASAP. Perhaps we can have as many players as possible voice their opinion on this option?

What the warrior sees will go a long way towards determining where to settle as bcool says. The 27th seems a long ways off....we have lots of time for testing!
 
For the record, I am also leaning toward the SIP scenario.

But as Ron says, we do have time for testing, and unless there are XOTMs going I generally don't play any other kind of game :p (BOTM39 is quite time consuming at the moment...)

Perhaps a test game where an extra seafood resource is revealed by the warrior? How much stronger would that be if it could settle an off-shore city before building any wonders? How much stronger would it be if that seafood was fish or crabs for the extra :health:?

Something funny: I don't have Civ installed on my work computer... so I'm writing my own civ simulator game in C# :lol: #civfanatic
 
Let me go macro and strategic rather than micro and tactical for the moment ...

First question, if folks are simming deep in a test game, doesn't that depend a lot on what the test game has in the fog that may or may not be there in the real game? Or are folks limiting their inferences to what the tiles we can see in the start will yield?

Second, if we want two cities on the starting island (calling it a land mass makes my head hurt ... :lol:), it has to be in place (Portugal) and the only other tile available is SE (or Italy). In contrast to the plains hill option (the Alps) [our island is a mini-western Europe, you see ... ;)).

We just barely have enough "Elba room" for a second city! :eek: :lol:

Presumably the logic of two cities on the island is that we can use all 4 food all the time, even whiping the cities to pop 2. That logic makes sense to me, it's and extreme version of a shared resources approach that I sometimes have used.

At some point that whipping gets limited by red faces, right? So we'd have to slow down the whip rate eventually ... or are you looking at whipping down more than 2 pop at a time, to max hammers yielded per red face produced?

Which gets to my question of what an 8/4 cycle means ... you whipping down 4 pop at size 8?

Third point ... whatever we do with Elba, there ain't much commerce to be had there. Seems unlikely we keep the capital on that island, with the lack of land to cottage (or would we maintain a specialist capital on Elba?)

Fourth point, probably not much in resources on Elba either. Hmm ... do we need early archery to be safe colonizing off the island?

Points three and four make colonization critical, and thus if two cities on the island give us whipped settlers warriors and galleys the fastest, that is the way to go.

Elba is not our promised land, but is has to be the springboard that gets us to the promised land.

dV
 
First question, if folks are simming deep in a test game, doesn't that depend a lot on what the test game has in the fog that may or may not be there in the real game? Or are folks limiting their inferences to what the tiles we can see in the start will yield?

Clearly the usefulness of test games will depend on whether the map resembles any of the tests. What they can do is give experience of what decisions work well in what contexts. For examples
  • I observed that getting a forge in a second city for a post-MC-Oracle Great Engineer will require a site capable of generating 180:hammers: ASAP after the Oracle. That means forests and food and whipping in some combination. If we're jungle-bound or on small bare islands, that strategy is dead in the water.
  • If we are on small islands, then our early expansion does not need defenders, which is very useful because we are :hammers:-limited early on.

Second, if we want two cities on the starting island (calling it a land mass makes my head hurt ... :lol:), it has to be in place (Portugal) and the only other tile available is SE (or Italy). In contrast to the plains hill option (the Alps) [our island is a mini-western Europe, you see ... ;)).

The second city being founded before a galley has been built has some merit in a non-wonder scenario.

SIP has the primary advantage that one city has the option of using all four :food: tiles. That's good for wonder-building, running as a GP farm, and as a Globe Theatre drafting farm.

In a wonder scenario, IMO it seems best to follow a strategy of 3-pop whips on Paris building settlers, overflowing max :hammers: to the wonder - this requires an early galley, either to get the off-island mine going (SIP) or to settle the co-generated settlers (settle-PH). Thus I don't see any advantage to settling PH unless we see things in the darkness.

Presumably the logic of two cities on the island is that we can use all 4 food all the time, even whiping the cities to pop 2. That logic makes sense to me, it's and extreme version of a shared resources approach that I sometimes have used.

Yup.

At some point that whipping gets limited by red faces, right? So we'd have to slow down the whip rate eventually ... or are you looking at whipping down more than 2 pop at a time, to max hammers yielded per red face produced?

I think it is possible for both cities to run around size 4 whipping to size 2 on what we can see already. Capital growing to 6 to whip settlers to 3 needs all the food for long enough that the other city basically doesn't whip.

Which gets to my question of what an 8/4 cycle means ... you whipping down 4 pop at size 8?

Yep, that's what the nomenclature means. I introduced it about 15 posts ago.

I'm not sure that 8/4 is a good cycle for us to run, given our strategic considerations.

Third point ... whatever we do with Elba, there ain't much commerce to be had there. Seems unlikely we keep the capital on that island, with the lack of land to cottage (or would we maintain a specialist capital on Elba?)

So far, the leading suggestions are some degree of Moai/Colossus/HR/Bureaucracy capital, or some kind of NE/Pyramids-Representation/Caste/Pacificism GP farm running 11 specialists.

I'm skeptical that a plan to move the capital will bear fruit fast enough - we have to find a good enough site within about the first 100 turns, settle it, cottage it, build a palace, switch to Bureaucracy, grow the cottages, and then later maybe switch to Nationalism.

Fourth point, probably not much in resources on Elba either. Hmm ... do we need early archery to be safe colonizing off the island?

Depends on the size and distance of settle-able land. Exploration will be key, as many of us have now said.

Points three and four make colonization critical, and thus if two cities on the island give us whipped settlers warriors and galleys the fastest, that is the way to go.

Two 4/2 cities has to be fastest for crazy-REX (more tile usage more often, more efficient whipping at smaller sizes), however that's no good unless we have a plan to be able to parlay that REX into enough :science: lead to conquer the world, and enough :hammers: to make it work. There's no Great Lighthouse, and just Currency is not enough. That means we have to
  • Oracle some critical tech that will do the job,
  • build the Pyramids,
  • build the Colossus, or
  • run HR with a zillion warriors while working a pile of coastal tiles when we have nothing better to do
  • find some real Bureaucracy capital site and move there

Elba is not our promised land, but is has to be the springboard that gets us to the promised land.

Agreed. I'm expecting the promised land to belong to an AI and/or be far away :evil: My third test game explores one such scenario. Grifftavian's test game explores another.
 
If we are on small islands, then our early expansion does not need defenders, which is very useful because we are :hammers:-limited early on.
Hehe, I'm sure dV has plenty to say on the topic of assuming safety from barbarian invasion :lol:
 
Agreed. I'm expecting the promised land to belong to an AI and/or be far away
This is exactly what I would bet on. I think this game is sort of modeled after the XOTM a few months back where we had the isolated start and the other island was working together. The difference here is that this start is truly isolated (we started on a continent then), AND we don't have much land to work with. The AI's will have plenty of room to develop and my guess is that an early rush will be next to impossible if not physically impossible. The only question for me is weather we will need Astronomy or not? If not, there is the possibility of capturing great locations early, but that is why I really think we will need Astronomy.

Where is Griff's test game??? I can't seem to find it anywhere. :(

Hehe, I'm sure dV has plenty to say on the topic of assuming safety from barbarian invasion
:lol::lol::lol::lol::lol::lol::lol::lol::lol::lol::lol::lol::lol::lol::lol::lol::mischief:
 
Ok finished BOTM39 tonight ... I should be able to test a bit now with some of the posted saves in the coming days. But grades are due and the wife is sick so maybe not a ton.

I actually think if we see seafood with the warrior in range of the proposed 2nd city that makes SIP even better.
 
Top Bottom