Strategy Thread - for quick strategy questions and answers

[Dhalphir]

That was the main thing I wondered - whether the food that an unhappy citizen consumed was costing production on a settler or worker.

So aside from slightly increased maintenance, there's no real problem with having unhappy citizens as long as you can still work your good tiles, and they can always be whipped away.

[suttree]

NobleHelium is going to accuse me of being obvious again, but unhappy citizens ALWAYS have a cost. Even if you are stagnating on a worker/settler, that unhappy citizen cost you food to build and now its doing nothing. If another micro plan allows you to put that food to better use, then you are incurring a cost.

There is always a cost, but sometimes it is worth the cost even if you are not building a worker/settler. Imagine a city with +5 food surplus small enough that you can convert 1 food to 2 hammers via the whip. Even with an unhappy citizen thats (5-2)*2=6h vs 5fh stagnating on a settler. There's more to it than that, but you get the idea.

The fun of learning civ4 is learning WHY things work - so ignore the rules of thumb people like to throw about and work things out for yourself.

[WilliamLP]

Dhalphir, this is a great idea for a thread! There are a lot of things I've wanted to ask, but didn't want to pollute the general forum with small posts, and this solves the problem.

Does whipping a city the turn after it grows into unhappiness have a cost, other than starting the whip unhappiness timer later? (Vs whipping the turn before growth?) Maintenance costs are calculated before growths?

Another question: Is there an explanation somewhere of the mechanics of random religion spread? I'm looking for enough to calculate exact probabilities.

[suttree]

Not because I know the answers, but because its fun to find the answers:

Does whipping a city the turn after it grows into unhappiness have a cost, other than starting the whip unhappiness timer later? (Vs whipping the turn before growth?) Maintenance costs are calculated before growths?

Maintenance: This suggests maintenance is calculated before growth. So no cost there. Needs an expert to confirm.

I can speak to the cost of whipping after growth, though. It is often more efficient to whip the turn before growth because it trades labour turns for food - the food comes from the reduced cost of the subsequent increase in pop.
This is easy to see from a concrete example, but its certainly worth doing the math in general.

City X is at size 6 working a single +6 food tile and grassland. City X has a granary and the food bar is at 26/32.

(1) If City X whips this turn, the food bar changes to 26/26 before growth. On the next turn City X grows to size 4 with the food bar at 19/28.
(2) If City X grows this turn, the food bar remains at 26/32 before growth. On the next turn City X grows to size 7 with the food bar at 16/34 then whips so the food bar is at 16/28

Waiting delays production and costs you 3 food. Please note that this is a boundary case where the population you are whipping produces nothing. In the standard case, the whipped pop is productive so you trade a turn's work for food. Food is valuable, so this can be a good trade.

Another question: Is there an explanation somewhere of the mechanics of random religion spread? I'm looking for enough to calculate exact probabilities.

I'm too lazy to turn this in to numbers: link

[suttree]

OK, decided to not be lazy.
Expanding on what's written in the comments here:

Num=(HolyCity+Shrine)*ReligionSpreadFactor/MapDistance
ProbabilityOfSpread=Num/1000

HolyCity=1 if the holy city exists, else 0
Shrine=1 if the shrine exists, else 0
ReligionSpreadFactor=100
MapDistance=100*ActualPlotDistance/(MaxDistance-5)

[WilliamLP]

Thanks for the religion spread formula. I guess the gist of it is that it converges to a ten percent chance (twenty percent for a shrine), per city per turn, for a city that is close to the holy city. And then linearly scales to 0 when two cities are as far apart as the map allows? This may mean a city in the middle of a torroidal map has at least a 5% chance per turn of spreading religion to any city on the map with a connection?

Can religion spread without open borders? I've heard conflicting information about this - that article seems to say no.

I actually tested your whip case for fun:




That's the city with 1 hammer into a library and 26 food in the box, ready to triple-whip.

Case 1:






Case 2:






I see a difference of 3 food, in favour of whipping the turn before growth, exactly what you said. This makes sense because the growth from 3 to 4 takes away 13 food (growth at 26 means granary saves half), where the growth from 6 to 7 takes 16 food.

Like you said this is whipping away tiles with no net benefit, so the question is roughly quantified by whether the one turn of the three tiles we're whipping away is worth a cost of 3 food, plus delaying the thing by one turn, plus whip anger that ends 1 turn earlier? (Or just 2 food for a double-whip, or 1 for a single?)

The most common triple-whip is surely a settler, does it change anything that a city can't grow when producing, and the overflow from a settler will go into hammers? What I've done before is: 1. get 10 hammers into it at, say size 6. 2. grow to unhappiness. 3. triple-whip it. (At size 5 you tend to have to do this anyway just to triple whip.) But I wonder if I've been leaving food on the table doing this. I should test this too, but I'm just lazy.

I didn't really consider this mechanic precisely before, so thanks. (I started by trying to always whip the turn before growth because I saw Sullla doing that a lot. )

[SevenSpirits]

The net value of the worked tiles is often higher than 1f per tile. And in fact, take a common case of settler 3-whip from size 6 on quick speed in a city that grows 1 size per turn. If you whip immediately, your tiles worked count goes 3, 3, 4, 5. If you grow first (putting the hammers into your next non-stagnating build that turn), you work 6, 4, 4, 5 tiles, which is 4 more. Now continue that case: in a good city, you'll have enough overflow from that settler (since you put 2 turns of production in it and only needed 5h past the 60 from whipping) to maybe 1t a worker or at least 2t it. Due to this large overflow, you'll therefore want to stagnate for 1-2t further after whipping. So now the calculus is, say, 6, 4, 4, 4, 4, 5 vs 3, 3, 3, 3, 4, 5 - so growth first gets to work *6* more tiles. It may even be better than that: those extra tiles might be the difference between finishing the worker in 1t instead of 2t! (And if you can finish the worker in 1t then you can whip a settler every 6t instead of every 7t which is quite nice.)

On the flipside, the main drawback is getting the settler a turn later. That can be worth quite a bit too.

[Krill]

The calculus is also slightly different for quick speed versus normal as well, because the slavery unhappy runs off in 6 turns as opposed to 10, whereas costs are rounded...better...at 0.67 cost of normal.

Honestly, trying to play two games on the two different speeds is a nightmare if you are doing thorough MM for both. I swear that the strategies for the speeds are slightly different due to the different tipping points from the calculus.

[suttree]

The net value of the worked tiles is often higher than 1f per tile. And in fact, take a common case of settler 3-whip from size 6 on quick speed in a city that grows 1 size per turn. If you whip immediately, your tiles worked count goes 3, 3, 4, 5. If you grow first (putting the hammers into your next non-stagnating build that turn), you work 6, 4, 4, 5 tiles, which is 4 more. Now continue that case: in a good city, you'll have enough overflow from that settler (since you put 2 turns of production in it and only needed 5h past the 60 from whipping) to maybe 1t a worker or at least 2t it. Due to this large overflow, you'll therefore want to stagnate for 1-2t further after whipping. So now the calculus is, say, 6, 4, 4, 4, 4, 5 vs 3, 3, 3, 3, 4, 5 - so growth first gets to work *6* more tiles. It may even be better than that: those extra tiles might be the difference between finishing the worker in 1t instead of 2t! (And if you can finish the worker in 1t then you can whip a settler every 6t instead of every 7t which is quite nice.)

On the flipside, the main drawback is getting the settler a turn later. That can be worth quite a bit too.

Yes, you don't get any savings until you grow - so a settler whip where the food box is not full after you whip 6-3 might be bad (maybe you reaaaally want the city now instead of one turn later)

Seven, a better example would be the case where you are at size 6 and you have 26 or more food in the food box. In that case tiles worked is 3,4,4,... and you are converting 3 labour turns (say 6fh) to 3 food at size 4 - which is a good deal even before you consider the return on founding the city one turn earlier.

Your comments on planning for efficient use of the overflow stand, however.

[SevenSpirits]

Yeah, they are definitely different. Like in the case I just described, if you want a 6-turn 3-whip cycle on quick speed making a settler, worker, unit, then you have to be able to 1t the worker and grow a size per turn, because even with a 1t worker you are spending half of the time stagnating. On normal you don't need to spend 1/3 of the time building the settler so there is more wiggle room.

The biggest difference though is chops. On quick speed your worker can produce 3.25hpt by spending his time chopping. On normal a worker can produce 5hpt chopping. 3.25hpt is so close to not even worth it that you should probably just wait until math for most of your forests. It changes the entire teching strategy on top of changing how you develop.