[BTS] SGOTM 25 - Home Slices

[babybluepants]

I concur. How is that possible? I'm currently playing a game with Kremlin. I often get confused on things like overflow gold, etc. but I thought I had the Kremlin calcs down, 100% understood. It should be simply 45 base hammers and +10% is 49. So a 2 whip should be impossible on a 50h unit. So unless the SP 10% works differently than OrgRel 25% or forge 25% or IND trait 50%, etc, I'm confused.
Are you sure you tested that correctly? Not sure how you could test it wrong, but I really thought I understood Kremlin and the rounding involved.

I just tested and it is 47h. Don't understand the math either.

 

[WastinTime]

I just tested and it is 47h. Don't understand the math either.

upload test pls. Save me a few minutes.

 

[babybluepants]

upload test pls. Save me a few minutes.

Yan'an is set up.

 

[LowtherCastle]

I concur. How is that possible? I'm currently playing a game with Kremlin. I often get confused on things like overflow gold, etc. but I thought I had the Kremlin calcs down, 100% understood. It should be simply 45 base hammers and +10% is 49. So a 2 whip should be impossible on a 50h unit. So unless the SP 10% works differently than OrgRel 25% or forge 25% or IND trait 50%, etc, I'm confused.
Are you sure you tested that correctly? Not sure how you could test it wrong, but I really thought I understood Kremlin and the rounding involved.

klarius briefly explained the mechanics in SG8:

Well, I'm not really calculating all the truncations .
In principle it's like:
The number of hammers per population (or per gold in the case of cash rushing) is multiplied by 1/2 for WP or RC.
Then its multiplied by 100 divided by 67 for Kremlin.
The result is then multiplied by the hammer modifiers in the city (200%-225% in your examples IIRC). So we are talking something like 44-50 hammers per pop.

Edit:
For Khazak there is also another thing to consider:
Pop-rush theatre from scratch, before you start your wonder.
As one pop will already complete the theater even from scratch with 300% modifier, you get one full turn of production (+ a few coins) overflow.


Yes. Yes.

and incorporating the Wonder penalty and multiple pop whips:

LC, it's not quite as bad with pop-rushing. You don't lose a hammer for every pop-point due to truncation, but only on the sum.
That is the game first multiplies the population whipped by 30 and does the various calculations on this.
E.g.: 12 pop in a 1/3 wonder (or 1/2 from scratch) with Kremlin is done as:
((12*30)*100)/((300*67)/100) = 179h
This is then multiplied by the production modifiers:
E.g.
(179 * 335)/100 = 599

A special case is the 2/3 penalty. There the denominator is calculated as (150*67)/100 truncated to 100. So Kremlin does exactly cancel this penalty.

which should give 30*100/67=44*1.1=48h, as Kait calculated. That it turns out as 47h could be an artifact of the SP 10% modifier. klarius didn't address that.

For example, if the 10% is on the original 30h...

Wonder divisors:

Just looked again. It's different for different wonders (world and national) in BtS . The early wonders are still 1/2 world 2/3 national, but:

Wall Street 1/2
IW 1/2
West point 1/2
Rushmore 1/2
Red cross 1/2
Kremlin 1/3
Eiffel 1/3
Liberty 1/3
Broadway 1/3
Rock'n'Roll 1/3
Hollywood 1/3
Three Gorges 1/3
Pentagon 1/3
UN 1/4
Space Elevator 1/4
Moai 1/2
Apostolic Palace 1/4

Dammit! I only checked Versailles and Hermitage . I had no idea that the rush cost varied! Thanks

Just to complete the list:

Versailles 1/2
Forgotten Palace 2/3
Hermitage 2/3
National Park 2/3

LC - There are two factors affecting rushing

Without (A) or with (B) any hammers invested
Without (1) Kremlin or with (2) Kremlin

A1 is then without hammers invested and without Kremlin
A2 is then without hammers invested and with Kremlin
B1 is then with hammers invested and without Kremlin
B2 is then with hammers invested and with Kremlin

Each pop provides base hammers, just as hammers from tiles are considered base hammers.

When rushing a building without hammers and without Kremlin (A1) you get 20 hammers / population.
When rushing a building without hammers and with Kremlin (A2) you get 30 hammers / population.
When rushing a building with hammers and without Kremlin (B1) you get 30 hammers / population.
When rushing a building with hammers and with Kremlin (B2) you get 45 hammers / population.

These base hammers are further modified when you rush a wonder. The modifier could either be

2/3 (Forgotten Palace)
1/2 (Red Cross)
1/3 (Broadway)
1/4 (UN)

Base hammers for wonders then range from 5 to 30.

Example: Poprush Red Cross without hammers and with Kremlin, and no production bonus
1 pop = 30 * 1/2 = 15 hammers. Red Cross costs 600 hammers => 40 nurses are distributed among the world to repair the war damage we produce

The modified base hammers are then multiplied with the production bonus in the city.

Example: Above Red Cross is rushed without hammers and with Kremling (as above), but the city has a forge, factory, and power. Bonus = 100%.
1 pop = 30 * 1/2 * 200% = 30 hammers. Only 20 nurses are sent.

Worst rush case is then rushing UN without any hammers and without Kremlin in a city with no production bonus:
1 pop = 20 * 1/4 = 5 hammers. It then takes 200 pop to make peace in the world.
Best rush case is National Park in a capital with IW, factory, powerplant, while running bureaucracy, organized religion and state property (bonus 25 + 25 + 50 + 100 + 50 + 25 + 10 = 285 %) with Kremlin and hammers invested:
1 pop = 45 * 2/3 * 385% = 115 hammers.

 

[WastinTime]

Nope, it's 49.

Seems like a bug to me, but here's how it works:
As I expected, using my normal Kremlin math, we get 49h
That's : 30 * 1.5 * 1.1 = 49
or in the case of a 2 pop whip:
60 * 1.5 * 1.1 = 98

And in the test save, confirmed we get 49h. (98h for 2 pop)
(Thx bbp, for the test save)

The bug is that it let's you 2-whip at 2/50 or even 1/50. There must be a couple rounding issues going on there.

Another example: whip a 1/15 warrior. You get 50/15. +49 again.

P.S. when you do more than 1 pop, you have to multiply in the right spot...
It's not (30 * 1.5 * 1.1) * 2 even tho, in this case we still get 98.

On epic (base 45) and marathon (base 90) speed, it often matters that you use it the right way:

base * pop * 1.5 * modifiers

 

[LowtherCastle]

Nope, it's 49.

Seems like a bug to me, but here's how it works:
As I expected, using my normal Kremlin math, we get 49h
That's : 30 * 1.5 * 1.1 = 49
or in the case of a 2 pop whip:
60 * 1.5 * 1.1 = 98

And in the test save, confirmed we get 49h. (98h for 2 pop)
(Thx bbp, for the test save)

The bug is that it let's you 2-whip at 2/50 or even 1/50. There must be a couple rounding issues going on there.

Another example: whip a 1/15 warrior. You get 50/15. +49 again.

P.S. when you do more than 1 pop, you have to multiply in the right spot...
It's not (30 * 1.5 * 1.1) * 2 even tho, in this case we still get 98.

On epic (base 45) and marathon (base 90) speed, it often matters that you use it the right way:

base * pop * 1.5 * modifiers

This is true. When I had 2/50h and 2whipped it became 100/50h. According to klarius' explanation, the 1whip would be 48h and the 2whip would be 97h, but still shouldn't 2whip on 2/50h. Weird.

 

[LowtherCastle]

Is it possible that one of the BUG/BUFFY fixes added a hammer when they tried to solve the hammer overflow problem? That is, changed it from 100/67 to 1.5 or something like that?

 

[LowtherCastle]

The bug has to do with SP. Without the Kremlin, we can 2whip with 17/50h to 83/50h. That 2whip shouldn't happen either. At 18/50h it's a 1whip.

With Kremlin but without SP, 2whip at 5/50h goes to 94/50, which is 89h, as predicted by klarius (60*100/67=89h), 1whip at 6/50h to 50/50, which is 44h, again, as per klarius. (30*100/67=44h)

Kremlin does not give a 1.5 multiplier, it gives a -33% Hurry Production Cost (read it!), which in the code is 100/67.

 

[Kaitzilla]

The bug has to do with SP. Without the Kremlin, we can 2whip with 17/50h to 83/50h. That 2whip shouldn't happen either. At 18/50h it's a 1whip.

With Kremlin but without SP, 2whip at 5/50h goes to 94/50, which is 89h, as predicted by klarius (60*100/67=89h), 1whip at 6/50h to 50/50, which is 44h, again, as per klarius. (30*100/67=44h)

Kremlin does not give a 1.5 multiplier, it gives a -33% Hurry Production Cost (read it!), which in the code is 100/67.

In Soviet Civ 4, the math calculates you!

Seriously though, state property whips ought to be 33
Which means a 2 pop whip at 17/50 should be completely impossible.
I can't understand at all.

Thanks for you and WT pinning down the problem.

Also, thanks Bbp for making the test save.
I know how long it takes to reproduce that many tiles and rivers.

 

[LowtherCastle]

Hypothesis: SP goes before Kremlin:

30*1.1*100/67 = 49
60*1.1*100/67 = 98

In theory, what happens with a forge when building a library:

Trunc(30*1.35)*100/67 = 59
Trunc(60*1.35)*100/67 = 120, putting the SP 10% and the forge 25% before the Kremlin

Trunc(30*1.1*100/67)*1.25 = 61
Trunc(60*1.1*100/67)*1.25 = 122, putting only the forge 25% after the Kremlin

Trunc(30*100/67)*1.35 = 59
Trunc(60*100/67)*1.35 = 120, putting both boni after the forge

30*1.5*1.35= 60
60*1.5*1.35= 121, WT's formula

Testing:

2whip library at 29/90h => 149/90, which is 120h
2whip library at 30/90h => 150/90, which is 120h
1whip library at 31/90h => 90/90, which is 59h

Appears that SP somehow puts the boni before the Kremlin. That's hard to believe. Could also be after, per that data, but that wouldn't explain the HA data.

 

[WastinTime]

Appears that SP somehow puts the boni before the Kremlin.

This was what I suspected. When I asked to get a test save, I was going to figure out if that was true, but then I lost interest after I got 98h from a 2 whip. I was happy. Thx for the extra analysis. Back to the game.

 

[LowtherCastle]

Axe 2whips at 2/35h
Sword 2whips at 7/40h

SP somehow miscalculates by a hammmer when the cutoff is.

 

[WastinTime]

Axe 2whips at 2/35h
Sword 2whips at 7/40h

This is what really matters. Keeping track of the cutoffs. We should keep a list of all the usual suspects (like HA)

 

[LowtherCastle]

Back to the game, we're on the question of what to tech. We could
1. tech HBR for AI obliteration (instead of the manly chariot solution)
2. hold off until we decide whether we need Astro, then beeline that
3. tech to Literature for some wonder bread spamming
4. tech Currency for the TR and gold trades
5. tech Monarchy for HR and big cities

 

[babybluepants]

Hong Kong / Gua

I think the optimal play for Gua is to whip is something like the following before gift and recapture: granary / wb / galley / chariot / LH.

Capturing Hong Kong with chariot+archer on a galley is a bit dicey. Churchill is PRO. I tried 2t bomb and 2t strike with the bomber, and still got only about 40% odds with the chariot. The archer is pretty much useless attacking.

The problem is, Churchill could also whip an archer and we’re hooped.

With some MM, we may be able to get one more whip out of Gua before it has to be gifted. If that extra whip was a sword, our odds would be much better.

In terms of timing, we want to gift and recapture Gua on the turn that the last whip is complete to not waste turns with it, but we also need it as the bomber base for 3-4t in a row and we lose some turns ferrying back and forth with Hong Kong.

The timing is quite tricky. I guess I’m not sure that Hong Kong works as the first capture, but I would love to capture it as early as possible and ideally before it gets a second city defender...

Edit: xpost with LC - I think we can hold off a bit.

 

[WastinTime]

I'm going to be sad if we don't use a Choko at some point.
I think the HBR thing can wait since we wouldn't run over to Europe first. But I do loves me some HAs. Even attacking close neighbors they're just faster and can snag workers. However, I think we can get by with the manly chariots for now.

I'd choose 2. hold off on tech

Getting HK (presumably for TRs) is not such a big deal rt now while we have only 3 cities. We're getting foreign TRs. Let's go for a different target.

 

[LowtherCastle]

Hong Kong / Gua

I think the optimal play for Gua is to whip is something like the following before gift and recapture: granary / wb / galley / chariot / LH.

Capturing Hong Kong with chariot+archer on a galley is a bit dicey. Churchill is PRO. I tried 2t bomb and 2t strike with the bomber, and still got only about 40% odds with the chariot. The archer is pretty much useless attacking.

The problem is, Churchill could also whip an archer and we’re hooped.

With some MM, we may be able to get one more whip out of Gua before it has to be gifted. If that extra whip was a sword, our odds would be much better.

In terms of timing, we want to gift and recapture Gua on the turn that the last whip is complete to not waste turns with it, but we also need it as the bomber base for 3-4t in a row and we lose some turns ferrying back and forth with Hong Kong.

The timing is quite tricky. I guess I’m not sure that Hong Kong works as the first capture, but I would love to capture it as early as possible and ideally before it gets a second city defender...

Edit: xpost with LC - I think we can hold off a bit.

My first thought is to skip the LH for now. Second, I like the idea of going down on Shanghai first so it can get started on its LH-GLH. Swords from Gua attack on T+4 (or T+3 with 2 roads). How about granary / sword / sword / sword / chariot / chariot? Or something like that. Our worker could road Gua-NE-2N2E before farming the rice maybe. And maybe even skip the chop if possible.

 

[LowtherCastle]

Note that if we time the Gua gift with CHurchill finishing a wb, then he might net the fish for us. We don't even have to delay the re-capture, right? Takes a couple turns to get the fish back, right?

 

[Kaitzilla]

Manly chariots work until a 2nd strong defender appears.
Then the defenders take turns healing and taking the big hit from the bomber.

I'm ok with Gua whipping a bunch of Swords to take France's city first since it is already connected to our Iron even without roads.
The Lighthouse can wait since 2 pop to get +1 is a huge expense with so many good AI cities and city spots to settle.
In that case, the Galley can also wait.
Even if barbs pillage undefended fish, it would have already turned a profit before being destroyed.

We'd also need 1 Chariot to take back our gifted city right away too.
Having a granary city with +12 per turn out of our hands for even a single turn is a huge loss.

A France DOW will sever Gau from our trade network, so maybe use 2 turns to chop a pair of garrison warriors before the France city falls and reunites our empire's trade network?

 

[WastinTime]

My first thought is to skip the LH for now

yea, I thought I objected to that already, but I'll say again, no lh yet.

skip chop also sounds good, but I haven't studied the mm.

Even if barbs pillage undefended fish, it would have already turned a profit before being destroyed.

u sure? even if we have corn & rice available? How much time will we spend at size 3?