The Whipping Thread

[Blake]

Originally posted by Dog of Justice
Incidentally, a fairer design would have the freshly built Granary provide 1 food for every 2 food produced, up to the (10+n) cap. But this isn't a big deal.

It's not like that?
I've always thought you have to go through a full bar of food to gain the full effect from a granary.

[couerdelion]

Originally posted by Dog of Justice
Optimize! I tried to time the granary reaching 30 hammers and the city reaching size 2. Clearly this optimization was a failure and it's extremely hard to go wrong with simply maximizing food.

Um, that's not how you optimize. A freshly built Granary in a size n city will double the first (10+n) food (i.e. half the food bar) you produce at the current population level, and do absolutely nothing for any additional food. Thus, when optimizing, you want about half the food bar filled when your Granary is completed. (If you have significantly more than half the food bar filled when the Granary is completed, you may want to go out of your way to create a large food overflow when your population grows, because every unit of overflow food will be worth 2 food.)

Incidentally, a fairer design would have the freshly built Granary provide 1 food for every 2 food produced, up to the (10+n) cap. But this isn't a big deal.

One thing I am not clear about is how the granary works during the first population growth after it has been built. This seems to suggest that, for every food surplus produced in a city, it stores away another up to a maximum of 10+n which would imply that, it applies a +100% factor to food in most situations.

If this is the way the granary works - rather than storing half of every second food surplus – then the conclusion would seem to me to be correct. Unfortunately I’m in a situation where I cannot play-test

[Arrian]

Blake,

Have you played around with the Aztecs in warlords yet?

-Arrian

[Quillan]

Well, either I wasn't paying attention or something, because I thought the granary still worked the way it has always worked: when the city grows instead of the food bar emptying, it drops to half full instead, thus you only need half as much food for the next growth. If it has any effect on food prodution directly, I was unaware of it.

[couerdelion]

Originally posted by Quillan
Well, either I wasn't paying attention or something, because I thought the granary still worked the way it has always worked: when the city grows instead of the food bar emptying, it drops to half full instead, thus you only need half as much food for the next growth. If it has any effect on food prodution directly, I was unaware of it.

The question was rather how the granary kept the food. For one thing, if you finish a granary just before you grow, you do not start the next pop growth with a half-full granary.

I was originally under the impression that it stored half surplus food produced ie as the granary fills, 50% of it is put aside as “stored”. Now it seems that all surplus food produced goes towards this “stored” amount until it reaches a maximum of half the amount required to grow.

[fed1943]

In vanilla, the first grow after the build of the granary triggers its

power (if in the same turn is not after).

So, the granary does nothing to said first grow.

Best regards,

[Dog of Justice]

Originally posted by fed1943
In vanilla, the first grow after the build of the granary triggers its

power (if in the same turn is not after).

So, the granary does nothing to said first grow.

Best regards,

What do you mean? (And do you have 1.61?) I'm confident that Blake, couerdelion, etc. are correctly interpreting what I said about Granary behavior, and that my description is accurate for 1.61.

[Krill]

IIRC the granary starts storing food from the turn it is built (ie food produced the turn the granary is built is not stored, but from every turn after food is stored). Though I remember this from such a time that it may have changed and then changed been back only to be altered a thrid time

[khumak]

Originally posted by couerdelion


The question was rather how the granary kept the food. For one thing, if you finish a granary just before you grow, you do not start the next pop growth with a half-full granary.

I've noticed this too. Because of that it's probably optimal to complete the granary ASAP if you're at or past the halfway mark for the next pop growth. The exception to this would be that it's not worth sacrificing 2 pop to finish it faster if you can wait a few extra turns and only sacrifice 1. I've actually started purposely devoting at least 2 tiles to absolute maximum food production in all of my cities specifically to get maximum benefit from pop rushing. So every city in my empire gets at least 2 farms unless it has food specials. A city with 2 fish or something doesn't need any farms unless it has wheat or corn or some other special that requires a farm.

Cities that are set up this way grow fast enough that you can pretty much keep pop rushing stuff as fast as possible, constantly knocking those cities down to a pop of 1 or 2 until they have all of the important stuff built. Then you can let them grow normally to their pop cap with all of the buildings they need intact. If my big cities need to build a few buildings I might poprush barracks in my outlying cities and pop rush some units while my big cities build their buildings the old fashioned way.

[Possibility]

What about the long term benefits of commerce from cottages or other sources? Keeping a city at size 1 or 2 for a long time means no commerce coming in, which means no science or money. Plus, keeping your population working those cottages will advance the cottages to towns producing lots of commerce.

I would like to see a study of the long term benefits/negatives of constantly pop-rushing and keeping your city small versus letting the city grow large and advancing the cottages to towns. Your city might have a few more structures built, but will you be a few techs behind in research?

Possibility

[fed1943]

The divergence, I think, it is just in my poor English:

Granary doesn't make food; keep half of the food, when it becomes

full; and,when it becomes full, the city grows.

So, granary doesn't speed first grow.

Best regards,

[khumak]

Originally posted by Possibility
What about the long term benefits of commerce from cottages or other sources? Keeping a city at size 1 or 2 for a long time means no commerce coming in, which means no science or money. Plus, keeping your population working those cottages will advance the cottages to towns producing lots of commerce.

I don't keep cities at size 1 or 2 for a long time even when I do pop rush. They're generally in the 3-4 range most of the time with short periods where they drop to 1 or 2 when I poprush something. Keep in mind that if you set your cites up right for poprushing, they have at least 2 decent food tiles (3 or 4 food each) or 1 excellent food til (5 or 6 food). That allows them to grow back to their former size quickly. Cities with zero potential for good food production I don't pop rush at all unless I'm at the pop cap or need to rush a culture related building for border expansion.

Yes you lose access to some tiles that your people could be working for a few turns but would you rather have that coastal city at the pop cap working fish/cottages and still working on that granary seemingly forever? Or would you rather have one that will grow back to the pop cap quickly and already has a granary, library, harbor, lighthouse, forge, market, bank, university, etc so that once it does grow back it will massively outproduce the non poprushed city in every way?

Obviously poprushing is not always the answer and blake gave a pretty good outline of when to use it. In general you can be more and more generous with the whip the more food your cities produce. High production, low food cities are usually poor choices for poprushing anything other than maybe a granary.

I just started a game with a ridiculously good 1st city location that I poprush every building I get access to as soon as the cost is low enough that I don't drop down below whatever population I need to work all of the crab/fish tiles. It's got 1 fish, 4 or 5 crabs (can you say future GPP?), 1 beaver in a plains forest, 3 plains hills, and the rest is grassland surrounding a freshwater lake.

Imagine how fast a city grows when it's working all of those fish/crabs with a lighthouse and a granary built? I just switch to pure food production after the poprush until I'm back at the cap again. This city is usually at the pop cap despite all of the poprushing it does and since it's not working on buildings it has plenty of time for wonder building or unit pumping if I don't have any wonders I want to build.

[Blake]

The newest test I've done has been this case:

We have a new city on the floodplains. This city has 3 floodplains to work, some river grasslands and a bit of forest (2 grassland 1 plains hill)

This city needs to be self contained, it's capital is off building the Oracle or gods knows what, so this city has to train a worker, build a granary and also produce some defense - a barracks, and after that, units.

I tried 6 different strategies and played it out for 50 turns, here is the screenshot of 4 of them:



The 6 strategies used:
1) Train worker at size 1.
2) Grow to size 2 and train worker.
3) Grow to size 2, put 30h into worker, whip worker.
4) Grow to size 4, whip worker.
5) Grow to size 4, whip granary, grow back to 4, whip worker.
6) Same as 4, but have the worker run away and join the circus (to help calculate the value of a worker turn).


I recorded the relevant information, tracking commerce through the use of 100% culture tax and a control city (to subtract free commerce from trade and such). I counted up the number of turns spent improving cottages and noted everything else.



I calculated Commerce Score by considering every turn spent working a cottage to be worth +2 commerce. For production score, I used the assumption that every pop is worth 30h (whip whip) and each food is worth 2h (whip whip).
For the Point Value I used h = 0.55 c = 0.4

I also noted when the worker was trained and thus the number of turns of work it did.


Important Notes:
Two cities completed workers on turn 15, these were train at size 1 and grow to size 2 then whip. The whip city had additional yield of +8h, +2f and +6c, so this is an interesting little penalty-free optimization - if a city has a 3 food tile and you can whip, then always grow to size 2 then half complete the worker and whip it.

Remarkably, growing to size 4 then whipping the worker was only one turn slower than training the worker at size 2, accounting for the one turn difference it generated about 8h extra plus some commerce.

The highest commerce score strat was Grow to Size 2, then whip. That generated the most commerce and you can see how the small boost from the "optimization" increased yield considerably over #1, which got the worker at the same time.

The highest production strat was whipping the granary at size 4, however not far behind was whipping the worker at size 4, with that strat generating a lot more commerce.

Clearly the best strat is the one which yielded the highest point value - growing to size 4 then whipping the worker yielded a whopping +50h over the next best, at the cost of 36c. However in order to experience the full power of that strategy it is necessary to have the 3 floodplains (or other 3f tiles) and when the worker will do growth-enhancement improvement it is likely having the worker out earlier will yield more turn advantage.

[couerdelion]

Once again, some sound experimental work Blake though I’ll have to read to through a couple of times to fully digest what it means. It is perhaps worth noting that the earlier arrival of the worker will increase commerce far into the future since every cottage that it built will not only be worked earlier but will also grow earlier. So those times where the worker is arriving earlier, will tend to be slightly undervalued.

[Blake]

Actually that's why I include the bonus +2c for working cottages which are still improving. That +2 is basically "future income".

Working a cottage-type improevement for a single turn turn also increases income in the future for a single turn (the cottage will eventually upgrade, working it this turn just moves the timeline forward by one turn):
Cottage:
Also gives +1c as a hamlet, +1c as a village, +1c as a town. (+3 total)
Hamlet:
Also gives +1 as a village, +1 as a town.
Village:
Also gives +1 as a town.

(Later in the game the village and town numbers get increased by +3)

It is obvious that it is not better to work a 2-1-5 village (pretend it's grassland elephant :P) than a 2-1-6 gems mine, in fact it's better to work the gems mine because you get the +1 extra c now, rather than in the future after the village has upgraded.

I think to be totally accurate it'd be closer to this:
Working a cottage is worth +2 extra.
Hamlet is worth +2
Village is worth +0.5

I might use these numbers in the future.