SGOTM 15 - Kakumeika

I apologize that I missed shulec's earlier comment comment about ending his turnset early. I appreciate him picking up the remaining in his turnset despite the confusion. If I had realized shulec's turn was over I would have been more explicit about whose turn it was.

We need to debate the next 4 turns and get the game moving. We have been taking a bit too much time early on. If the game goes longer than we hope we definitely want the game moving faster than it has been.

2 big issues in the next 4 turns
Chop workers or not?
The longer term path is a bit unclear, but I'm in favor of chopping at least 2 forests for more workers out of DeerGold.

Where to settle?
If the plan is to build the oracle in Delhi then I think settling by the gems instead of gold/corn/sheep is better in the short term. We can get the gems working a turn or 2 after we settle by the gems. While the gold near gold/corn/sheep takes much longer to get worked.
 
At which turn are you guys seeing third cities pop up in test games? I assume we have plenty of time to make the gold/sheep site our 4th city, but is gold/sheep safe a fifth city?

Don't know, been playing insularly. There's plenty of ~T100 test game saves floating around.
 
I doubt sheep/gold is a safe 5th city, but I haven't noticed when the AI build their 3rd cities.

I think sheep/gold might be a good city to build the hanging gardens, and maybe eventually we could get the corn. Or just settle 1N of gold and get the deer and give up on a deer/cows city in favor of another coastal city.
 
My tests use slavery to get a worker out and then regrow quickly, so that 3 food might be important.

Whipping Delhi without a granary for a worker seems a doubtful proposition to me. At size 5 before we improve copper (which is mostly when we're talking about) we produce 15:food:12:hammers: and eat 10:food:. That's 17/turn on a 60:hammers: worker, which takes 3-4 turns to build, depending on overflow. 7 turns puts out two workers regardless. Setting up a 1-pop whip almost always requires two turns of that 17, for three turns to produce. Then there's three three more turns to produce a second worker at size 4 putting out 15/turn. Then we are at size 4 building some non-worker with a :food: excess of 6 taking 4-5 turns to fill a 28:food: box. If the short-term need for workers is acute enough to consider whipping, then producing one on turn 3 or 4, and then one on turn 7 and no need to re-grow to size 5 seems better than getting one on turn 2 or 3 and one turn 5 or 6. With growth to size 6 desirable to work the copper mine as well as our Gmines, whipping seems less clear again.
 
I doubt sheep/gold is a safe 5th city, but I haven't noticed when the AI build their 3rd cities.

I think sheep/gold might be a good city to build the hanging gardens, and maybe eventually we could get the corn. Or just settle 1N of gold and get the deer and give up on a deer/cows city in favor of another coastal city.

I like deer/cows because it is extremely efficient in the 2-5 size range and that's when we'll have most need to run scientists for beakers or whip units for an army. Long term, it has to run a few cottages and mines and be content to be mediocre, but in an SGOTM race, packing as hard a punch as you can early is normally the way to go. Two cities around size 4-6 grow and produce more efficiently than one larger one (particularly if you happen to have a low happiness cap, or need to run scientists pre-Caste), so you get back the 100:food:+:hammers: cost of the second settler reasonably soon from the extra central tile (at least 3 stuff/turn), faster ability to grow to use new tiles, lower marginal costs of growing to use new tiles, the additional whip cycle, and 1 extra pop with HG. The cost is higher maintenance, of course, but the extra city comes with at least 2:commerce: nearly free, as well as the chance of 1:gold: from the Kashi.
 
Where to settle?
If the plan is to build the oracle in Delhi then I think settling by the gems instead of gold/corn/sheep is better in the short term. We can get the gems working a turn or 2 after we settle by the gems. While the gold near gold/corn/sheep takes much longer to get worked.

In my tests, I settle the gold site and just one turn later I'm working the gold mine. I work gold at size 1 until the Sheep pasture is complete.
I would like to grab this land before it is gone and get this city building an Axe ASAP.
 
Whipping Delhi without a granary ....

You may be right about pre-granary whipping, but I had a hard time with your calculations because in reality, it is a 2 pop whip and we re-grow with +7 or 8 food.
Why not take the hammer boost now, and get the slavery unhappiness turns burning off ASAP so we can whip again?

Give it a quick try. I like how the worker is out fast and can mine Copper/road/mine gold, then quarry.

After the worker whip, I finish the settler and then TGW to grow. I suppose we could whip the settler, but I like the timing of having the worker before the settler.
 
You may be right about pre-granary whipping, but I had a hard time with your calculations because in reality, it is a 2 pop whip and we re-grow with +7 or 8 food.
Why not take the hammer boost now, and get the slavery unhappiness turns burning off ASAP so we can whip again?

Give it a quick try. I like how the worker is out fast and can mine Copper/road/mine gold, then quarry.

After the worker whip, I finish the settler and then TGW to grow. I suppose we could whip the settler, but I like the timing of having the worker before the settler.

Ah, I understand you now. I was still thinking of building a worker after the settler, but the copper discovery changes that. The marginal value of a worker is higher when there's a good tile to improve. Switching to a worker @17 production and 2-pop whipping it next turn (down to 13 production) overflows 30 production back to the settler. So the settler production might not even be slowed down. The worker improves the copper in four turns, and IIRC grassland river copper is 2:food:4:hammers:1:commerce:, which gets settler production at size 3 back to 15 with a bonus :commerce:. Or we can probably overflow onto TGW if that settler timing suits better.

This gets a worker moving with the settler whether we settle gems or gold, for further payoff when we build that mine. bcool won't have been whipping this worker, hence his slower start with that new city.
 
Another trick we should consider for financing our tech is chopping post-Maths forests into whipped walls. With fast workers, stone and OR, this could be amazingly effective.
 
Yes I wasn't whipping a worker in Delhi and that is why I didn't like settling the gold city. If we whip a worker in Delhi then I can see how that city is fine as a 3rd city.

I'm okay with the deer/cows city, I was just suggesting if we stop putting down new cities at 6 then we limit ourselves to 2 coastal cities if we settle deer/cows. If we don't settle deer/cows then we can have 3 coastal cities and still only have 6 cities for lower maintenance. Perhaps there are clever ways to get wealth out of 7 cities and still not slow our tech pace down much. Then the deer/cows city is a no-brainer.

I haven't tested the whipped walls/chopped forest into walls builds in a while. I understand the overflow hammers can produce a lot of gold, but I'm not sure if they changed those mechanics.
 
Another trick we should consider for financing our tech is chopping post-Maths forests into whipped walls. With fast workers, stone and OR, this could be amazingly effective.
+
Spiritual Temple is also an option for the gold-from-overflow trick.

I believe the gold overflow trick doesn't work the way it did in 3.17 because 3.19-BUFFY removes non-generic multipliers from traits and resources.

~~~~

Re whipping the Fast Worker:

A 2 pop whip without a granary leaving Delhi at size 3 when we actually want it to be producing Settlers at size 6?

I have quite strong doubts that the earlier worker turns outweigh the losses of not working our already improved tiles in the long term. I don't think we want to spend turns regrowing while building other stuff right now.
I cannot come up with any hard numbers to prove it but let me quote DaveMcW with a post he gives assuming a granary is present:

It is always inefficient to kill a citizen working a mined grassland hill.

At size 6, it becomes inefficient to kill off a mined plains hill.
At size 6, it becomes inefficient to kill off a plains forest.

At size 10, it becomes inefficient to kill off a mined desert hill.

At size 20, it becomes inefficient to kill off an engineer.

~~~~

I'm worried about the tendency to not push forward our peaceful land grab in the east vs Toku as aggressively as we could -- we might lose Horses or Iron that way.

What was the reason for not getting Oracle in Sheep-Gold again? Was it the perceived risk of a late AP?
With Copper in Delhi's BFC for faster production that danger has decreased. There should be at least 5 forests left to chop into the AP (more with grows).
And I already explained the emergency plan of rushing out the GProphet from Delhi while getting the GS from DeerCow as 2nd GP.

Sheep-Gold with a very soon captured Corn will be a very strong city. I cannot emphasize the importance of food (Corn + green rivers) enough! :)
 
What was the reason for not getting Oracle in Sheep-Gold again? Was it the perceived risk of a late AP?

The main arguements were against the usage of many pre-math chops and concern about having to many workers and not enough work.

Deergold has to be neglected after the workers are chopped out. All workers are used to get the marble and then the Oracle in sheepgold. We will want to get at least one farm and more GrassMines up and running. The Oracle in dehli can allow for this.

We don't risk losing the AP unless we lose the race to Theology. If we found Christianity, the AP can be built with organic hammers in plenty of time. Since we get no resource bonus, the lack of chop bonus means the only benefit of chopping is speed.
 
Updated test game to T46. The test game has a beaker or two more overflow, correct gold, build queues, city, worker and land state. Didn't check for forest status. I added a Mansa city to simulate Osaka. Save attached.

I observed a tell-tale with the tile production display - ocean that has food and commerce shown must be in the BFC of a land tile, even if that tile is a mountain and cannot have a city built on it. There's an ocean square about 7W of DeerGold whose production overlay proves there is land to its south, and that led me to observe a peak on that land. Unfortunately, we can't tell if there is more land there. Screenshot from real game attached. Sorry I didn't take another zoomed out to make the position clear to all.
 

Attachments

  • mountain in the fog.JPG
    mountain in the fog.JPG
    20.9 KB · Views: 53
  • Kakumeika test game T46.CivBeyondSwordSave
    Kakumeika test game T46.CivBeyondSwordSave
    133.7 KB · Views: 57
The main arguements were against the usage of many pre-math chops and concern about having to many workers and not enough work.
There will be enough work, even WastinTime agreed on 5 workers, let the 6th FW get roads towards Hammu for foreign trade routes without Sailing.
Deergold has to be neglected after the workers are chopped out. All workers are used to get the marble and then the Oracle in sheepgold. We will want to get at least one farm and more GrassMines up and running. The Oracle in dehli can allow for this.
I'm happy to temporarily neglect DeerGold to get the yummy 6-yield tiles up earlier in the new cities and our strong formation in the east than a farm and a mine in DeerGold.
I'd like to let DeerGold slow-build a Settler for Hammu's double-liberation city.
 
The main arguements were against the usage of many pre-math chops and concern about having to many workers and not enough work.

There were about 8 chops and 6 workers to achieve T62 Oracle in SheepGold, whereas we can can get away with fewer of both for T62 Oracle in Delhi.

I think one key idea behind Oracle Delhi is that stacking our prophet :gp: means we get the GProphet for bulbing Theol faster, and get the next hybrid GSci/GPro faster also - assuming we're happy bulbing whatever looks good at the time with whatever we get.

Deergold has to be neglected after the workers are chopped out. All workers are used to get the marble and then the Oracle in sheepgold. We will want to get at least one farm and more GrassMines up and running. The Oracle in dehli can allow for this.

Yeah, OR-buffed DeerGold at size 6 (gold, deer, Gfarm, 3 Gmines) can build Pyramids in 16 turns without chops if we take a bit of care of it now...

We don't risk losing the AP unless we lose the race to Theology. If we found Christianity, the AP can be built with organic hammers in plenty of time. Since we get no resource bonus, the lack of chop bonus means the only benefit of chopping is speed.

Chopping is always just a bonus to speed. Maths adds base hammers. A resource/building bonus amplifies the increase in speed from a chop, but doesn't add base hammers.

Come to think of it, three turns chopping a pre-Maths forest near DeerGold for a worker seems like a poor proposition. The hill forest chops are worse, because they need a further worker turn. Three worker turns for the normal forests realises 20:hammers:, but DeerGold put out 24:hammers: on a worker in that time. So we're spending three worker turns to get a worker three turns earlier. This breaks about even, unless there is no ability to do anything now and we will be in a rush later - as we were with the T62 SheepGold Oracle plan - and are willing afford to pay the premium of not having Maths. Anyway, it seems right for Eiffel to abandon DeerGold and head for the copper (gets it up a turn faster even with the worker whip) and then over to help with SheepGold. The first worker out of DeerGold is in plenty of time to organize a Gfarm for it before it's Oracle time (but need testing).
 
I believe the gold overflow trick doesn't work the way it did in 3.17 because 3.19-BUFFY removes non-generic multipliers from traits and resources.

Yeah I'll do some experimenting tomorrow.

~~~~

Re whipping the Fast Worker:

A 2 pop whip without a granary leaving Delhi at size 3 when we actually want it to be producing Settlers at size 6?

I have quite strong doubts that the earlier worker turns outweigh the losses of not working our already improved tiles in the long term. I don't think we want to spend turns regrowing while building other stuff right now.
I cannot come up with any hard numbers to prove it but let me quote DaveMcW with a post he gives assuming a granary is present:

I agree totally with those general conclusions. A Gmine converts food to hammers at 1 to 3 (eats two food, grows one, grows 3 hammers), whereas whipping with a granary converts it at around 1 to 2 depending on population and initial food bin level and the length of time considered.

However always is too strong. Some things can be strong enough to warrant such a whip. Here, the worker gets faster access to the copper tile (in time for the start of regrowth), and then the gold tile at SheepGold, and then the trade route to SheepGold... I guess the way to settle this is to compare the productivity of different plans out to finishing the Oracle or so. I'll do that with my spreadsheet tomorrow.

I'm worried about the tendency to not push forward our peaceful land grab in the east vs Toku as aggressively as we could -- we might lose Horses or Iron that way.

Yeah I think the case for SheepGold next is pretty strong.

What was the reason for not getting Oracle in Sheep-Gold again? Was it the perceived risk of a late AP?
With Copper in Delhi's BFC for faster production that danger has decreased. There should be at least 5 forests left to chop into the AP (more with grows).
And I already explained the emergency plan of rushing out the GProphet from Delhi while getting the GS from DeerCow as 2nd GP.

Sheep-Gold with a very soon captured Corn will be a very strong city. I cannot emphasize the importance of food (Corn + green rivers) enough! :)

Yes the corn is great, but the available evidence is that even with Oracle, monument and perhaps Hinduism, SheepGold doesn't capture the corn until at least turn 95, and one of my test game reports had no capture at T102. So we'd have to commit to a library there as well, and we're going to want an axeman and a granary...
 
Yes the corn is great, but the available evidence is that even with Oracle, monument and perhaps Hinduism, SheepGold doesn't capture the corn until at least turn 95, and one of my test game reports had no capture at T102. So we'd have to commit to a library there as well, and we're going to want an axeman and a granary...
That's due to Mansa having knowledge of both Mysticism and Meditation, right?
With Buddhism and Monument he gets the border pop early, but our Toku probably doesn't have Mysticism yet and Buddhism is very unlikely to spread to Osaka. The first culture generating building is likely to be a Library so he gets the border pop later -- when SheepGold is near the 2nd border pop for a much earlier capture of the Corn.

In one of my tests Toku had control of Corn for only two turns.

BTW, nice find of the lands west of DeerGold. :goodjob:
 
Back
Top Bottom