[BTS] SGOTM 25 - Home Slices

[LowtherCastle]

What will the Shanghai slave do?

 

[Kaitzilla]

Not sure I understand you. I still see the Kremlin completing on T32 (after hitting Enter on T31).
T30 2whip sword
T31 sword
T32 Kremlin

It is possible to complete the Kremlin after hitting enter on T31, but it requires working Iron Mine on T30 and T31.
A few are lost to overflow limit when 2-pop whipping the 6/40 Sword and then producing 11 base in Beijing.

By building the Kremlin 1 turn slower, Beijing can work more food and regrow faster.
The whips I'm eying all happen on Turn 33.
Do we need +2 Spy GPPs?

Might be able to work Kremlin in Yanan for 1 turn for 12 fail.

 

[LowtherCastle]

It is possible to complete the Kremlin after hitting enter on T31, but it requires working Iron Mine on T30 and T31.
A few are lost to overflow limit when 2-pop whipping the 6/40 Sword and then producing 11 base in Beijing.

By building the Kremlin 1 turn slower, Beijing can work more food and regrow faster.
The whips I'm eying all happen on Turn 33.
Do we need +2 Spy GPPs?

Might be able to work Kremlin in Yanan for 1 turn for 12 fail.

Okay. The Shanghai slave could go to help improve the Beijing rice and fp watermill, and so on. Sounds good. I think we should seriously consider building a pavilion in Beijing asap and growing it to work as many tiles as we can improve, running the cultural slider for happies. Unorthodox but who cares. We have it, let's use it.

 

[LowtherCastle]

Maybe even 16h failgold in Yanan. Did you hold growth last turn?

 

[Kaitzilla]

On the road at Manchou-1w on T32, right? SHould be able to get 2 chariots there unless the T31 Man-2s1w chariot needs to chase workers, which is the goal anyway.

How about:
T31 - Archer 2W1S Manchou, Sword+Choko 2W2S of Manchou, 2 Chariots 1W2S of Manchou, 2 Chariots on Rice. Bomber blasts Iron Mine
T32 - If LB attacks Archer, Bomb LB and kill with Sword, 2 Chariots 1E of Manchou, Choko 1W1S of Manchou, 2 Chariots 1W2S of Manchou.
If no attack, Bomb LB and move Archer 1W1S of Manchou.
T33 - Manchou falls, Sword promotes to City Raider I and moves to 1W of Manchou if it miraculously didn't fall , 2 Chariots 1E of Manchou can grab any Workers that didn't get scared into Manchou.

Manchou keeps 20% defenses in all scenarios.

Maybe even 16h failgold in Yanan. Did you hold growth last turn?

Ya, held growth last turn.
2 more is probably more valuable than +4

Okay. The Shanghai slave could go to help improve the Beijing rice and fp watermill, and so on. Sounds good. I think we should seriously consider building a pavilion in Beijing asap and growing it to work as many tiles as we can improve, running the cultural slider for happies. Unorthodox but who cares. We have it, let's use it.

The Shanghai slave could do a lot of help with Rice and improvements near Beijing.

Was gonna just pre-chop near Shanghai for faster for The Great Lighthouse?
Shanghai can do a lot of whips!
Was thinking Library or Pavillion 2-pop whip into Forge to get borders moving, then Forge whip into TGL, followed by enough chops to finish TGL.

If we can get that Tokyo slave, can pair it with gems worker to improve Shanghai, and the current Shanghai slave can really help in Beijing.

 

[LowtherCastle]

Pavilions will also give us +1 from dyes, which we should settle sooner or later.

 

[LowtherCastle]

How about:
T31 - Archer 2W1S Manchou, Sword+Choko 2W2S of Manchou, 2 Chariots 1W2S of Manchou, 2 Chariots on Rice. Bomber blasts Iron Mine
T32 - If LB attacks Archer, Bomb LB and kill with Sword, 2 Chariots 1E of Manchou, Sword+Choko 1W1S of Manchou, 2 Chariots 1W2S of Manchou.
If no attack, Bomb LB and move Archer 1W1S of Manchou.
T33 - Manchou falls, 2 Chariots 1E of Manchou can grab any Workers that didn't get scared into Manchou.

Manchou keeps 20% defenses in all scenarios.

Even better. Just keep watching workers with Spy, just in case they go somewhere weird, like chopping Man-NW.

Green light from me!

 

[LowtherCastle]

Also with pavilion in Bejing, we could run an artist or two if we want to speed up the GP and give ourselves a chance for an early GArtist.

 

[LowtherCastle]

Gau stealing the Fish for 1 turn would also be very beneficial getting Settlers and Workboat out a bit faster.

That's Gua's fish being stolen by HK as far as I'm concerned. Gua gets it whenever required.

 

[LowtherCastle]

Was thinking Library or Pavillion 2-pop whip into Forge to get borders moving, then Forge whip into TGL, followed by enough chops to finish TGL.

Shanghai and Gua are both great for GP spamming, if we so choose. Do we want one for a pure Artist? With Kremlin, SP and a forge, TGL whips at about 30h/pop, so that's something to consider too. Chops are useful but could be kept for AP* or something else. Partly a question of where we want to put the worker-turns right now. Later our workers might be more idle. Yanan and Beijing are desperate for wkr-ts and Mongolia, (Manchou, 2Gold) and roads west are too. You decide. I often decide based on what accelerates us the fastest right now. I think that's Beijing and Yanan.

Pavilion pops the borders 1t sooner, unless the cultural slider helps. Library adds to our HBR beakers.

* AP with Kremlin+SP+forge+OR whips for 18h/pop. Pretty lousy.

 

[LowtherCastle]

There is currently 15 turn of whip anger in Yanan.
We will experience an angry citizen next turn on T31, but since we are building Settler, there won't be any food losses.



Oh!
The free +1, ya I forgot about that.

Yes, that clears up the math perfectly.
So no AI knows Horseback Riding yet.

Extra Credit math homework:

Spoiler :

Making a test game where I know no one has Civil Service, the various slider settings produce the following numbers:

Civil Service - 1352 to finish.
Preqreq bonus, 20% for knowing Code of Laws.
No bonus for not knowing Theology.

0 per turn, 1352 turns
1 per turn, 676 turns
2 per turn, 451 turns
3 per turn, 338 turns.
4 per turn, 226 turns
5 per turn, 194 turns
6 per turn, 169 turns
7 per turn, 151 turns
8 per turn, 136 turns
9 per turn, 113 turns
10 per turn, 104 turns

Having no per turn means Civil Service should never complete, but it does.
Must be a secret +1 per turn bonus!

10 per turn gets things done in 104 turns.
10 * 1.2 * 104 turns = 1236
That is far short of 1352 that Civil Service costs!
Let's add in the +1 that I thought only came with 0% slider.
(10+1) * 1.2 * 104 = 1372.8
That is close, but can't be right.
If the math worked that way, Civil Service would be done in 103 turns, not 104.

As always when I'm stumped by civ 4 math, there is a round down somewhere, or Floor().
Floor((10+1)*1.2)*104 = 1352, exactly what Civil Service costs!

Now we need 1 final test.
Is the formula Floor((10+1)*1.2)*104?
Or is it Floor((10*1.2)+1)*104?

Plug in all 11 values from 0 per turn to 10 per turn into both formulas, and see which gives 11 correct answers!

Floor((??+1)*1.2)*x turns = 1352
Plug in 0 Science, get x=1352/[Floor((0+1)*1.2)]; x=1352 turns. So far so good.
Plug in 1 Science, get x=1352/[Floor((1+1)*1.2)]; x=676 turns. So far so good.
Plug in 2 Science, get x=1352/[Floor((2+1)*1.2)]; x=451 turns. So far so good.
Plug in 3 Science, get x=1352/[Floor((3+1)*1.2)]; x=338 turns. So far so good.
Plug in 4 Science, get x=1352/[Floor((4+1)*1.2)]; x=226 turns. So far so good.
Pug in 5 Science, get x=1352/[Floor((5+1)*1.2)]; x=194 turns. So far so good.
Plug in 10 Science, get x=1352/[Floor((10+1)*1.2)]; x=104 turns. Great!

Floor((??*1.2)+1)*x turns = 1352
Plug in 0 Science, get x=1352/[Floor((0*1.2)+1)]; x=1352 turns. So far so good.
Plug in 1 Science, get x=1352/[Floor((1*1.2)+1)]; x=676 turns. So far so good.
Plug in 4 Science, get x=1352/[Floor((4*1.2)+1)]; x=271 turns. WRONG
Plug in 5 Science, get x=1352/[Floor((5*1.2)+1)]; x=194 turns. So far so good.

So take the visible from your slider, add +1, then multiply by prereq bonus, then round down to the nearest integer.
That should be the true per turn.

That makes me really want to tech Horseback Riding as our 1st tech!
The Kremlin means our cities with Libraries and coastal cities can make 2 Horseback Archers every whip!

I've lost my vigor for Currency.
We can tech it once we get Alphabet and that nice +20% research bonus.

The formula ZPV gave to me, which works all the time as far as I can tell is:

formula: [(1 + BaseResearch)*[100+[30*KnownTeamsWTech/TeamsAtStart]+20*Prereqs]/100]
​

with truncations at each square bracket [ ], three total, and BaseResearch is what you see up in the left-hand corner next to the research slider.

This formula doesn't necessarily distinguish between zero and one AI having a tech very early in the game, btw. That's why I use a spreadsheet. For example, for a one-prereq tech like HBR in our game (9 TeamsAtStart), with our research slider at 100% and showing a BaseResearch of 19b or 20b, both zero and one AI knowing the tech yields the same answer to the formula above, whereas with BaseResearch of 18b or 21b, one AI shows +1b over zero AI.

Note 1: Ignore the color coding in this image. It's clearly all messed up. Once upon a time I had it so it showed where the numbers changed in rows.

Note 2: This spreadsheet also tells you where there are beaker jump points. For example, if you happen to know that one AI knows the tech you're researching, going from 20 to 21 BaseResearch gives you an extra beaker you don't see in the upper corner. So you could MM for an extra coin somewhere and get two beakers in return. Or if you don't necessarily want to go 100% research, you perhaps could MM and adjust the slider down from 24 to 23, because you see that you're getting jobbed a beaker at 24. (Edit: Oops, you can't see that. I'll expand the image...)

 

[WastinTime]

OurDistance > HisCulture/OurCulture*HisDistance, where 0c => 1c.

Except for the offshore bit, which doubles the score for the one not offshore. Also, it presumably means if we're both offshore (a city on the other island to the east), it's back to what you typed above.

When I say 'distance' I mean sid's distance, which means if it's off continent, then it's doubled. So with that baked-in to ourDistance...

OurDistance > OurCulture/HisCulture*HisDistance
(you had the culture flipped)

Not sure about ties. Seems like the first player checked wins, so we would NOT be able to liberate to anyone if we have tied scores.
If two AI tie then it's the first one processed. *not tested.

 

[WastinTime]

2W of the clam.

I was going to say, why not settle right next to clam, but your clever way of settling means the recaptured clam city (ours would have no border pop) will have 2 forests, of which one can be handed off to the fish city. (like you said) Just pointing out how well thought out that was.

I was kinda eyeing up those cities for us to just use ourselves, but I guess this is our way to get Alpha.

It could attack our Fishing boats in Gao, attack our Galley

I don't think the AI normally attacks empty, uninjured galleys with a single galley. Not sure how they know if it's empty (cheaters)
It would also be a first if the AI had a unit to attack amphibiously. Normally their galleys are empty or have a settler/escort.
Going for the fish seems most likely

If we don't grab those Workers the same turn the city falls, they will likely be disbanded by Toku.

Is this because is only other city is off-continent? cus otherwise, I find the AI doesn't disband workers very often. Even if they have no escape.

 

[LowtherCastle]

Is Yanan still planning to build a worker?

 

[Kaitzilla]

Is Yanan still planning to build a worker?

Yanan is gonna do 1 turn of fail gold, then settler whip, then library.

Mongolia is the only city scheduled to build a worker.

 

[LowtherCastle]

Yanan is gonna do 1 turn of fail gold, then settler whip, then library.

Mongolia is the only city scheduled to build a worker.

Okay, I guess that's the one I remembered. Do we need it?

 

[Kaitzilla]

Is this because is only other city is off-continent? cus otherwise, I find the AI doesn't disband workers very often. Even if they have no escape.

When and why the AI disband Workers is a mystery.
Not even Tachwaxon's excellent "Slave Economy" guide talks about it.

The 100% case is when the AI recaptures a Worker that the player stole.
It is always instantly terminated.
Workers tend to flee to the next city when it gets captured, but sometimes they get disbanded.
Workers that flee into a city rarely get disbanded.

I'm sure I've killed off Deity AI with 1 units and not 1 Worker was captured!
I call it the Ghost Case Scenario.


That's why, barring a woodsman 2 unit, I try to scare Workers into a city with horse units before capturing the city.
Tend to get the most workers with the least chasing that way.

 

[Kaitzilla]

Okay, I guess that's the one I remembered. Do we need it?

Yes!

Mongolia has no good 2nd tile yet, and 11 turns into 59 with Kremlin.

 

[WastinTime]

I'm sure I've killed off Deity AI with 1 units and not 1 Worker was captured!

Of course they vaporize if you kill an AI. Unless you use the 'require complete kills' game option.

 

[Kaitzilla]

Of course they vaporize if you kill an AI. Unless you use the 'require complete kills' game option.

More context, the civ had 5 Workers when the war started.
I didn't make any effort to capture them since I had only slow units and the stiff resistance required 1 stack of doom.

By the time I got to the last weak city, I spread out my troops to herd any workers into the last city.
There were no workers.

So I spread out more, but saw nothing.
Even freshly arrived 2 units on neighbor's roads didn't see them.
They must have been disbanded.

I've also seen workers vanish that physically couldn't have run away, so I know some of them are getting disbanded.

Many players think they just run away, and they do, but I'd swear there is a point when the AI hits the skull and crossbones button.

I'm not crazy