SGOTM 13 - Gypsy Kings

On weekdays when I don't have a lot of civ time, this thread really does outrun me ...

Ronnie1 said:

This is the key point here....we get to 2 pop whip the Granary with a max overflow of 44H going to the Oracle on the exact same turn the Oracle is available. Then the following turn, we get the Chop/Mine into the Oracle also. Then we get to work max hammers for the time we are building the Oracle so we can whip it to completion on T94. Feel free to take the latest test game and see if you can improve upon that time. If you are successful, we will do it your way!

Click to expand...

I wasn't criticizing or questioning the plan, just making sure I had a clear sense of the logic of it. I don't tend to spend time calculating whip overflow variations in my solo games, but maybe I should start doing that.

mabraham said:

Yeah. A GP farm is useless once we start waging war - we should be glad of a viable conversion plan. We win a medal for "fastest to become last man standing", not for "prettiest last man standing with the most Great People"

Click to expand...

"prettiest last man standing with the most Great People" was the objective of a recent prior SGOTM, wasn't it?



dV

I'm not sure a GP Farm is useless once the war begins by the way. What is a GA worth in terms of extra units? Especially with a large empire a GA (especially with a captured MoM) can provided us with a significant boost in unit production (and wealth if we need it).

Agree, fast Astronomy looks likely to be required. If so, then it makes sense to beeline that. However, do we want to try for Duckweed-GLib before that? Requires a tech path of Masonry-Writing-CoL-Aes-Lit-Machinery-IW-Compass-Optics, with Maths and Calendar worked in somewhere suitable,

Click to expand...

Especially with Marble we should go for the early Great Library for sure I think. That tech path looks very promising right now.

Click to expand...

Currency is notably lacking. I assume Currency would pay for itself pretty quickly.

Currency is notably lacking. I assume Currency would pay for itself pretty quickly.

Click to expand...

Absolutely, especially given that most of our trade routes will be overseas trade even amongst our own cities.

I wasn't criticizing or questioning the plan, just making sure I had a clear sense of the logic of it. I don't tend to spend time calculating whip overflow variations in my solo games, but maybe I should start doing that.

Click to expand...

I know you weren't criticizing dV.....I was just giving you a bit of crap! We really would go about it in a different fashion though if we could save time. I took mab's plan and tried a number of different ways to shave a turn off of it, couldn't do it.

bcool said:

I'm not sure a GP Farm is useless once the war begins by the way. What is a GA worth in terms of extra units? Especially with a large empire a GA (especially with a captured MoM) can provided us with a significant boost in unit production (and wealth if we need it).

Click to expand...

True. I did have one GA at the start of my war effort to switch civics and get GT built, but it did have other effects. The "Mfg. Prod" plot shows doubling of output about that time, but I'm not sure what goes into that plot. +1 for most of our land tiles (mines and workshops) is useful, but doesn't sound like a doubling of total output... The extra go to paying for the next GA, and the extra (?) and lower expenses are also good.

I've had a few Great People spawn "naturally" anyway, and I certainly could have prioritised capturing MoM higher than I did...

bcool said:

Currency is notably lacking. I assume Currency would pay for itself pretty quickly.

Click to expand...

I don't build many GLibs, so am a bit hazy on how soon we need Lit. I know that in the game I played out that I failed to get it the first time (through being lazy), and had to reload to get to the conditions I was trying to test.

Masonry-Writing-Maths-Currency-CoL-Aes-Lit-Machinery-IW-Compass-Optics (with Calendar worked in somewhere suitable) is the most elegant (gets chop bonus faster, gets trade routes up faster, hopefully syncs CoL with Rep switch) but puts a delay on GLib...

da_Vinci said:

I wasn't criticizing or questioning the plan, just making sure I had a clear sense of the logic of it. I don't tend to spend time calculating whip overflow variations in my solo games, but maybe I should start doing that.

Click to expand...

Sure, me neither. However, a few hours spent looking for a small increase in the likelihood of something with a big effect over the several months of the game can be worthwhile.

This SGOTM gave me the excuse to write the planning spreadsheet I've always wanted. Managing the flow of and through tiles, with building, civ and leader effects, via the granary, with whipping and chopping and multipliers, enforcing limits has been scarily complex.* It has made for a good test bed, optimization platform and record-keeping device, however. Making the sliders work properly will be a job for another year...

* (I even found a bug in the Civ4 implementation - when a whip- decays during IBT, the newly-happy person still eats and doesn't do any work, unless you're building a settler, in which case the pre-turn food production estimate is higher by 2 and -production results concur, even when there is no plain 2 tile to work!)

I tried playing out the Test Game from T71 - T94 following bc's spreadsheet. Some things I noticed that might warrant checking during the actual game play...

On T90, when starting research on Masonry (7T), Treasury = 12 (-2/turn). Looks like we may need to put in 1 turn at 0% Science.

On T94, for Galley to pickup Worker, it had to sail 1N from Capitol (i.e., Worker could not make it back to Capitol on that turn). After Worker boards, Galley could sail 1SW, toward the Western Island, or just wait in place with 1 movement point remaining.

I had another look at the progress Score graphs for the game with some of our past saves. At the moment, each population point is about four points (but decreasing with increasing size), a tech is worth about six points (i.e. the bulk of the score) and the Oracle should be worth 5 points. After T51 we whipped twice, costing us about 15 points (actually), but we've settled and regrown to 5 population, adding a tech, so it looks like smooth slow growth. We haven't yet seen the full effect of granaries on our two-city resource-spread empire (i.e. higher total pop and continual growth). I think it will be quite clear from the score graph which teams have followed this approach.

PD looks like they uploaded shortly after whipping, and the Trolls look like they have an early granary or lighthouse (or maybe just three clams?).

From the power graph, it looks like PR and Trolls haven't researched BW as early as the crowd.

Comparing the actual save and the bcool-based-on-Grifftavian test game, the actual save has 933/1192 ratio of available land points, i.e. 25% less. This adds a little bit of weight to the idea that we have oceans and isolation. The estimate for total population is slightly higher - 1348/1271. There is a fair chunk of desert in the test game, so either desert might be missing, or the density of food resources or grassland tiles is higher in the actual save.

* (I even found a bug in the Civ4 implementation - when a whip- decays during IBT, the newly-happy person still eats and doesn't do any work, unless you're building a settler, in which case the pre-turn food production estimate is higher by 2 and -production results concur, even when there is no plain 2 tile to work!)

Click to expand...

It is possibly a feature or a known bug that unhappy people don't eat while producing a settler or worker. I've not directly tested it, but I've seen people comment on this before.

Grifftavian said:

I tried playing out the Test Game from T71 - T94 following bc's spreadsheet. Some things I noticed that might warrant checking during the actual game play...

Click to expand...

Great!

Grifftavian said:

On T90, when starting research on Masonry (7T), Treasury = 12 (-2/turn). Looks like we may need to put in 1 turn at 0% Science.

Click to expand...

City maintenance is population-dependent. In a test game, I am paying 2 city maintenance on T93 at sizes 4+4, but by T95 with 3+2 it is back to 1. My general recollection is that we get Masonry done just in the black .

Grifftavian said:

On T94, for Galley to pickup Worker, it had to sail 1N from Capitol (i.e., Worker could not make it back to Capitol on that turn). After Worker boards, Galley could sail 1SW, toward the Western Island, or just wait in place with 1 movement point remaining.

Click to expand...

It should do something about moving the worker to its destination - I think the discussion is leaning towards improving the western island first. Then we estimate how many turns before the settler is built, and pick a scouting radius for the galley. Rinse, repeat.

I won't be moving the galley on T94 so I leave it up to the next player to sail 1N and decide where to take the worker.

I've updated the PPP to reflect this (sorry I left the old Galley move in after we changed the worker plan).

And even if masonry can't be completed we can do 0% in the middle. I don't think finishing Masonry on a specific turn is necessary even if we decide to take the worker to the stone first--since the only thing he could do is road.

da_Vinci said:

On weekdays when I don't have a lot of civ time, this thread really does outrun me ...

Click to expand...

So true! I only stopped reading for about 15 hours and came back to ~40 posts. I'm not complaining - I love it!

Long term, Vassals are turned off right? Doesn't that make Conquest excessively tedious? Not really something that needs thinking about right now, but keep it in mind for later. We will need a reasonably robust economy, and a GPFarm capital with settled GProph and many Caste System merchants would certainly help.

I'm hoping for another Globe Theatre drafting city. The capital only has a maximum possible food surplus of 14 at sizes greater than 4, so from size 5-6 could draft every ~1.64 turns. That's pretty good, but I wouldn't be surprised if there is a city site with some fish resources and irrigated grassland tiles that can do better.

This is a pretty cool article.

edit: PPP looks good.

Okay on T93 everything is going to plan

stopping per request for workboat scouting input.

https://picasaweb.google.com/Cooley...CMKB6JSLp8vXyQE#slideshow/5582260637430765186

link to screenshot

I'm assuming we move the workboat scout 1S and then 1 SE. Any objections?

Other events
T76 Great Merchant born for GLH AI

T82 Judaism founded

T91 Stonehenge built

I'm assuming we move the workboat scout 1S and then 1 SE. Any objections?

Click to expand...

Keep going!

Okay played out T93 and Oracle is whip ready, the galley waits, the worker waits, the workboat revealed no seafood west of the stone

The capital has lots of hammers to use for the next build.

link to updated screenshot from scouting S and SE

https://picasaweb.google.com/CooleyPhysics/Sgotm13?authkey=Gv1sRgCMKB6JSLp8vXyQE#5582271600345332610

Played to the PPP, no deviations. I have the saves for all but 1 turn (T78) let me know if anyone wants to look at them. Then I'll spend the time to figure out the best way to link the file.

autolog

This SGOTM gave me the excuse to write the planning spreadsheet I've always wanted. Managing the flow of and through tiles, with building, civ and leader effects, via the granary, with whipping and chopping and multipliers, enforcing limits has been scarily complex.* It has made for a good test bed, optimization platform and record-keeping device, however. Making the sliders work properly will be a job for another year...

Click to expand...

Is this something you are willing to share

bcool said:

Is this something you are willing to share

Click to expand...

I'd be keen to see it to.

For the record, I've written an entire Civ4 simulation program - it was a good project for me to practice my C# programming. It reads the XML and everything

For what it's worth - we know that GLH owner is not Philosophical. We already knew they were Creative, so I guess that rules out Pericles... but we already ruled him out .
I'd be interested to see the Demographics screenshots of the turns around 74-77, as this is where most non-creative AI would get their 2nd border pop.

Judaism founded... no sure way to know who it was, but I'd be willing to bet it's Izzy.

adrianj said:

Long term, Vassals are turned off right? Doesn't that make Conquest excessively tedious? Not really something that needs thinking about right now, but keep it in mind for later.

Click to expand...

Vassals are off. Conquest was tedious. Settling a million small islands to eke out the land percentage is also tedious. We may need to do this if an AI gets too big for (our) boots.

adrianj said:

We will need a reasonably robust economy, and a GPFarm capital with settled GProph and many Caste System merchants would certainly help.

Click to expand...

... somewhat. CC has enough food for 7 specialists + GProphet, i.e. 23 from merchants. We'd have found time for grocer and market in this scenario, so call that 34. The empire income in my played-out game is 813, so our GP farm running merchants is nice, but only a small fish. Some more GPs to give me a Golden Age for Free Market for +3-4 in each of 40 cities, however...

adrianj said:

I'm hoping for another Globe Theatre drafting city. The capital only has a maximum possible food surplus of 14 at sizes greater than 4, so from size 5-6 could draft every ~1.64 turns. That's pretty good, but I wouldn't be surprised if there is a city site with some fish resources and irrigated grassland tiles that can do better.

This is a pretty cool article.

Click to expand...

True. Two fish, one clams and three grassland farms gives the ideal 16 excess required...

bcool said:

Okay played out T93 and Oracle is whip ready, the galley waits, the worker waits, the workboat revealed no seafood west of the stone

Click to expand...

Good job.

bcool said:

Other events
T76 Great Merchant born for GLH AI

T82 Judaism founded

T91 Stonehenge built

Click to expand...

Excellent. T91 was a very typical Stonehenge time in my test games. Our odds of getting the Pyramids have gone up!

bcool said:

The capital has lots of hammers to use for the next build.

link to updated screenshot from scouting S and SE

https://picasaweb.google.com/CooleyPhysics/Sgotm13?authkey=Gv1sRgCMKB6JSLp8vXyQE#5582271600345332610

Played to the PPP, no deviations. I have the saves for all but 1 turn (T78) let me know if anyone wants to look at them. Then I'll spend the time to figure out the best way to link the file.

Click to expand...

Windows 7 has a zip utility built in - select the files, right click, send to compressed folder, upload that. Need another tool for Windows XP, though.

bcool said:

Is this something you are willing to share

Click to expand...

Yes. It's still a "work in progress", but I did design it with users other than me in mind. I expect it to be very useful for hacking out our next series of builds in these two cities, but I haven't yet proved that it does Duckweeding exactly right, for example.

I still have a few ideas on the drafting table, but some time next week it might be in business. It's in OpenOffice Calc (freely available for every platform) since that had regular expressions. I've no idea if it will port to M$ Excel, and no interest in finding out!

Ideally, a two-city "tile governor" would be awesome, but I'm not holding my breath for that one... I have real code to write!