SGOTM 15 - Kakumeika

[Manco Capac]

EDIT: Wait, your units are threatening Toku's city. As far as I remember, a threatening unit in city BFC makes the AI refuses the city liberation, thus disabling the option. Try to disband your units, then report if any change. I'm making my own test meanwhile...

OMG! It worked!

EDIT: Friendly WangKon being at war. That's no good.

So basically, +1.5 plusmod for 30 per unit/city liberation (chariots, that is) and 100 for the settler itself is quite few cost for almost infinite diplo manipulation!

So how are gonna call this bug? Infinite city liberation?

 

[Sun Tzu Wu]

EDIT: Wait, your units are threatening Toku's city. As far as I remember, a threatening unit in city BFC makes the AI refuses the city liberation, thus disabling the option. Try to disband your units, then report if any change. I'm making my own test meanwhile...

A city is threatened by enemy units in in the 5 x 5 square centered on the city; not just units in the city's BFC.

Sun Tzu Wu

 

[ZPV]

A city is threatened by enemy units in in the 5 x 5 square centered on the city; not just units in the city's BFC.

Sun Tzu Wu

For a city to be "threatened", yes.
However, this check is for a visible enemy unit on a BFC tile.

 

[kossin]

One easy, although limiting, way of controlling the AP discovery by DanF is simply to stop using team AIs... there are plenty of things to be explored in Civ4 yet

 

[Manco Capac]

For a city to be "threatened", yes.
However, this check is for a visible enemy unit on a BFC tile.

What's the difference? I don't understand why there are two kinds of checking; one being the 5x5 square and the other a BFC.

Honestly, it makes no sense a chariot 2 tiles away diagonally is not threatening a city...

 

[ZPV]

What's the difference? I don't understand why there are two kinds of checking; one being the 5x5 square and the other a BFC.

Honestly, it makes no sense a chariot 2 tiles away diagonally is not threatening a city...

For things like not moving your defenders out to grab a bait-worker, the check is the 5x5 square (rather, a function of the city is called, asking if it is threatened).

For this, the code loops through the BFC tiles, and checks for any visible enemy units of the recipient (rather than enemies of the city owner, which is how the other check is coded).

 

[Manco Capac]

To my Kakumeika's fellows:

I wonder if it could be great to resume what we learnt in this SGOTM as generalistic inner game mechanisms, AI behaviours, etc.

Some aspects, some will remember without fault as AP bug, make peace with AI ends with cease fire, scouts are excellent baits because they are critical moments of the game, but other aspects were perhaps less obvious and perhaps forgotten because not stirring enough our minds.

If any contributions, I'll gather then in a modified post of the first page.

 

[Kaitzilla]

OMG! It worked!

EDIT: Friendly WangKon being at war. That's no good.

So basically, +1.5 plusmod for 30 per unit/city liberation (chariots, that is) and 100 for the settler itself is quite few cost for almost infinite diplo manipulation!

So how are gonna call this bug? Infinite city liberation?

I vote to call it the "No City-Razing Infinite Liberation Exploit" I am very glad we chose not to use it.

Tachywaxon's test game illustrates just how powerfully abusive it can be. I'm pretty sure it cost 12 Impi's their lives, but it made an enemy into a permanently friendly ally in a single turn and there is nothing that could have stopped it.

Spoiler :

As for folks saying that we should not have used holy war, I didn't realize it was a bug until after the first time we used it when I could not replicate what we were doing in a test game for our 2nd holy war. Once I realized the problem was that the test was missing a Team_AI, I saw that only by making a 2nd civ in the test game a full Hindu member by converting them could I get a holy war going against a heathen.

We noticed the bug earlier, but assumed it only applied to open border resolution, not holy war too. By the time it dawned on some of us that it was affecting holy war too, it was too late to stop using it.

We were definitely worried at the beggining of the game about having any Gawa opponent for the AP election, but by the end we saw that we could convert Shaka and it would be no big deal. If the bug had not existed, then we would have been mystified why the our first holy war resolution against Shaka didn't come up and our whole game would have been 10 turns slower as we hurried to flip Toku to Hinduism in Trojan city I think.

Very sorry if we did something wrong, but most of us don't know the AP inside and out . Even the guide in the Strategy forums is wrong on some of the major parts.

 

[Kaitzilla]

To my Kakumeika's fellows:

I wonder if it could be great to resume what we learnt in this SGOTM as generalistic inner game mechanisms, AI behaviours, etc.

Some aspects, some will remember without fault as AP bug, make peace with AI ends with cease fire, scouts are excellent baits because they are critical moments of the game, but other aspects were perhaps less obvious and perhaps forgotten because not stirring enough our minds.

If any contributions, I'll gather then in a modified post of the first page.

Let's see, off the top of my head.

1) "Show enemy movements" allows you to see if a stack is "grouped" or has a straggler or two. If there is a stragler, then the stack can attack and die and the straggler can retreat.
2) Cities can be gifted to another civ if your gift city is within 9 tiles of their closest city. Diagnal tiles are counted as 1.5, and any 0.5's left over get truncated.
3) You can (Liberate) a city to another civ if it is within 9 tiles of their capital and you haven't generated any culture in it yet.
4) If you and a hated civ are at war together against a 3rd civ, you can open borders with the hated civ no matter what the relations are at.
5) The open borders through shared war can be abused to teleport units. This is because canceling an open border results in "not right now, maybe in a few years" before they will sign a new one, which is like 25 turns for most civs, but with shared war this too is bypassed. So close borders to teleport, then sign a new open borders treaty right away if there is a shared war with the civ.
6) Trade Embargos are a perfect way to get two civs to stop having open borders or liberating cities to each other . They simply refuse to talk to each other for 20-30 turns I think.
7) DanF is an awesome team advisor. He has the combined Strategy Forum's knowledge from 2003-2033 in his head.

 

[Kaitzilla]

It was very sticky. I didn't mention that Toku had gone in WHEOOHRN and probably was targeting us.

Also maximizing food in all our cities.

Plus before I made a ceasefire with Shaka I swear Shaka wasn't the worst enemy of Hammuragawa. However when I checked after the ceasefire he was. There was no option so I gave Shaka back his hindu city but did not get any relations hit with Hammuragawa. We could have weathered a -4 hit but I was still surprised that it didn't show up.

Also we didn't get diplo hits from 2 spies that got caught trying to flip hammuragawa over to Hinduism.

I also poisoned the Southern Team AI's capital and made it unhappy with our 2 spies there... might have reduced its pop by 2. (but I think he might have just whipped it)

I was poised to poison FishTown too, but liz had trade routes to it so it had plenty of health and happiness. She did whip it for us though.

I'm pretty sure if we had not won T186, we would have had to go for a different victory condition (or conquered an AI completely and converted their cities to hindu).

Yes! Huge praise again .

I was so deflated from the T165 failure that I could not have played Sunday. It was really down to you and Mabraham to get us any Laurels at all.

 

[Kaitzilla]

Did we ever figure out why Shaka stayed in Hindiusm after we flipped him T154 even though he had only 1 and sometimes 0 Hindu pop?

If he had been kind enough to flip back to Buddism like Liz did, then Shaka/Hamma would never have been worst enemies and we could have easily gotten a cease fire with Shaka instead of that darned 10-turn peace treaty.

 

[Manco Capac]

Let's see, off the top of my head.

1) "Show enemy movements" allows you to see if a stack is "grouped" or has a straggler or two. If there is a stragler, then the stack can attack and die and the straggler can retreat.
2) Cities can be gifted to another civ if your gift city is within 9 tiles of their closest city. Diagnal tiles are counted as 1.5, and any 0.5's left over get truncated.
3) You can (Liberate) a city to another civ if it is within 9 tiles of their capital and you haven't generated any culture in it yet.
4) If you and a hated civ are at war together against a 3rd civ, you can open borders with the hated civ no matter what the relations are at.
5) The open borders through shared war can be abused to teleport units. This is because canceling an open border results in "not right now, maybe in a few years" before they will sign a new one, which is like 25 turns for most civs, but with shared war this too is bypassed. So close borders to teleport, then sign a new open borders treaty right away if there is a shared war with the civ.
6) Trade Embargos are a perfect way to get two civs to stop having open borders or liberating cities to each other . They simply refuse to talk to each other for 20-30 turns I think.
7) DanF is an awesome team advisor. He has the combined Strategy Forum's knowledge from 2003-2033 in his head.

Nice for the first reply. Once I get up to 4 replies, I'll start to compile in a modified post of mine (the first) with big title: tricks we learnt! That way, next SGOTM, if someone wants to remember, then we won't scratch our heads.

Reply to 1) I mentioned it once, but wasn't useful at the time.
Like Hatshe SoD during peace moments, if one observes suspicious movement from AI and has a close unit, put the unit on the same tile as the SoD and right-click the pale unit icons; they show instantly the sub-SoD's.

Reply to 5) In fact, OB treaty is once again available in a random way; it can be as fast as one turn of waiting.

• Tribute in wars is out of worst enemy system; quite useful to give something without bad diplo kickback.
• Culture over mountains gives two tiles of sight.
• Barbs are idiots; a single unit (whatever the kind) can block a chokepoint and no barb will come near mainlands if that chokepoint is the only path to mainlands. AI units doesn't follow that rule. And barb next to the unit will follow a rule that goes over this one: KILL.
• No-city-razing-infinite-city-liberation...

 

[Manco Capac]

I vote to call it the "No City-Razing Infinite Liberation Exploit" I am very glad we chose not to use it.

Tachywaxon's test game illustrates just how powerfully abusive it can be. I'm pretty sure it cost 12 Impi's their lives, but it made an enemy into a permanently friendly ally in a single turn and there is nothing that could have stopped it.

He he. Sacrificial way to indefinitely please an AI. Should work nicely to get along with Monty.

 

[Manco Capac]

• We have learnt teleportation over an ocean pre-Astro is possible thanks to map cylindrical wrapping.
• We have learnt a wounded caravel (85% of total health?) with a missionary is a way to spread a religion wihout a city or/and OB.
• Animals don't go on resources tiles unless attacking a unit on this tile; useful on how to avoid animals.
• We have learnt one angry citizen doesn't eat food while building a settler or a worker. (credit: mabraham)
• We have learnt beaker investment comes before hammers in builds. Sometimes useful for earlier use of marble or stone if settled on it.
• Banking beakers in the 5 first turns for a tech with a modifier means some added beakers because the modifier goes for the 5 turn lump of beaker at once at T5.
• In No Tech Trading, an AI refusing a tech for trade with an excuse means he has reached over 50% of said tech and is unwilling to lose its trade capacity.

 

[bcool]

2) Cities can be gifted to another civ if your gift city is within 9 tiles of their closest city. Diagnal tiles are counted as 1.5, and any 0.5's left over get truncated.
3) You can (Liberate) a city to another civ if it is within 9 tiles of their capital and you haven't generated any culture in it yet.

2*)Cities can only be gifted if they are on the same land mass and within 9 tiles of their closest city and if they are not in financial crisis.
3*) I believe city liberation is more complex. A city will liberate to the AI with the nearest capital. So I believe it can be greater than 9 tiles. Once there is culture of any kind in the city then it changes but I don't remember now what the formula DanF gave. Also I assume this means you can't liberate a city that is closer to your capital than the AI's capital even with 0 culture.

Did we ever figure out why Shaka stayed in Hindiusm after we flipped him T154 even though he had only 1 and sometimes 0 Hindu pop?

1) AI rarely flip out of religion if they are in a war (and their cities are threatened?) I think the AI must evaluate the value of flipping compared to 1 turn of anarchy. I've noticed this in other games, an AI will rarely flip out of the AP religion when at war.

2) Queued missionaries can be used to build wealth. If you erase a queued missionary from the queue and then build 3 missionaries using other cities, then the hammers invested in the missionaries in the erased queues will get converted to wealth. A way to build wealth before currency. Not sure if it has much use since you have to build 3 missionaries as well. Building wealth through fail gold with a wonder that you have the resource for is clearly better but has its own problems.

3) Never trust a map maker!

4) The whipping debate
whipping away grass hill mines to build settlers and workers is clearly better than not whipping (assuming you have a decent food surplus)
Whipping away grass hill mines for non-settlers and non-workers is only better only in some circumstances. Obviously if you need the units for a critical war it is better. If you have a good food source and farms to work while you regrow. Whipping vs non-whipping results in very similar total production outcomes if you are whipping away grass hill mines. However this assumes you have the happiness to deal with the unhappiness and you have grass farms to regrow the city while it recovers from the whip. --mabraham might remember more details from his tests.

5) You can get by with a lot fewer fast workers, I think we ran with fewer fast workers than cities for the entire game.

6) threatening more cities than an AI threatens dramatically improves the peace deal you can get. There is a cutoff point for the formula that calculates peace deal value if you can get at least 5/6? of their power. You can threaten a city by having a unit in the 5 by 5 square centered on the city.

7) with no tech brokering on, you can gift an AI a chain of technologies that are prerequisites for the next one. In a regular game you have to gift the technologies that have prerequisites one per turn.

8) barbarian units will move out of the way if they are on the tile you are going to teleport to. Although the barbarians might have been the reason Moses failed. If the barbarians have no valid tile adjacent to them for them to move to then maybe your units teleport to a different tile.

9) watch out for the isolated civilization with the great wall! They will teleport a horde of barbarians over the ocean at you!

 

[Manco Capac]

Nice ones, bcool. I shall compile them tomorrow.
I know this pedagogy of what we learnt looks a bit childish, but doing a brainstorming (because one recalls this and another that) of what we learn, especially details how the game works could mean
less imbroglio in the next SGOTM and focus more on strategy with all parameters known.

----------------------------------------------

Cute excel for city maintenance: http://forums.civfanatics.com/showpost.php?p=11263730&postcount=563

 

[babybluepants]

Congratulations on the excellent game!

2*)Cities can only be gifted if they are on the same land mass and within 9 tiles of their closest city and if they are not in financial crisis.

Unless they have only 3 cities or less, right?

 

[Bebekija]

Thank you very much for the kind of list you're attempting, it will be another fast bookmark once you've compiled it !!!

For the sake of readers, there's a little technicality about this one :

• We have learnt one angry citizen doesn't eat food while building a settler or a worker. (credit: mabraham)

Actually, though the food eaten by the angry citizen still gets added to settler/worker production, you still have to sustain enough food for this extra citizen, or you'll starve
For instance, if your happy cap is 5 and you're pop 6, you do get a +2F->+2H production bonus compared to what you would expect, *but* you still have to work at least 12F worth of tiles or your food bar will shrink.
This leads to a rather confusing city screen where the food bar is decreasing while at the same time the production bar shows something like 1F-13H, so actually indicates a food surplus

 

[Manco Capac]

^
Bebejika est un espion impudent des Canards en Plastique!

Thanks for the clarification.

 

[mabraham]

6) threatening more cities than an AI threatens dramatically improves the peace deal you can get. There is a cutoff point for the formula that calculates peace deal value if you can get at least 5/6? of their power. You can threaten a city by having a unit in the 5 by 5 square centered on the city.

These apply only to the Aggressive AI setting - and not to leaders with the Aggressive trait.