The last challenge -- Team game -- Deity trial

[bestsss]

C'mon there is a full open slot... and no one want to try her hand at?
(no bumps any more).

Now serious: will inland sea map be flat or not?

 

[PaulisKhan]

flat is default, I would vote to keep it that way unless compelling reason is presented =)

See if TMIT wants to play? We could watch him spend 4 hours playing 5 turns LIVE!!!!

 

[obsolete]

I think a good choice would be The Rook (sp?). I don't know how he did it exactly, but he got an army of cannons and Reds in that deity Elizabeth game by 1020 and that was damn impressive.

I guess my vote will go for Ghandi, though why do I keep feeling Monte would be nice to have? One less unit spammer to worry about, and abusing sacrificial alter with spritual makes me wonder...

Map I'd go with archipelago because I havent' played it for a long time and at least I know what I'm in for. The other choice sounds like something lop-sided is waiting to occure.

BTW, I think 10 turns is good enough. For those who want a shorter turn-set, a few months ago I asked around for deity players if they wanted to do a REALLY-MICRO of all micro team-games. 1 turn per set.

 

[PaulisKhan]

Ooh Montey, I played him not that long ago in an immortal whip fest SG but I'm open to another round with him.

 

[Duckweed]

Here's the settings we agreed.

and the start

a sheep and 4 FPs (1 under the settler ), pretty good capital with 11 riverside tiles. There could be at least 1 strategic resource in BFC if each FP is couted as 0.5 resource.

Your tech path and build order?

 

[The Rook]

I think a good choice would be The Rook (sp?). I don't know how he did it exactly, but he got an army of cannons and Reds in that deity Elizabeth game by 1020 and that was damn impressive.

Thanks for the voucher. I'm willing to give this a go, if that's ok with the rest of the team. If I sink the ship, I'll just blame obsolete for inviting me.

The challenge sounds interesting, but I'm ignorant about playing against AI teams, so must ask some possibly obtuse questions:

1. How does WFYABTA work with teams of AIs?

2. How does WHEOOH work? Does the team go into WHEOOH, or are both AIs independent with the peaceful member presumably dragged into war?

3. Do AI teams coordinate their research?

4. Assuming that teams automatically share techs, could this potentially help with the "we are not ready to trade..." mechanic?

5. How do war bribes work? If there is a backstabber in a team, could you potentially instigate a bribe through them and force their partner to follow?

 

[Dirk1302]

^Welcome, i liked your Liz game too.

 

[bestsss]

I will check all the questions and you get proper answer, Rook.

Research: the current researched value (I think beakers, need to check again) of any ongoing tech. is added to the weight when calculating what to research next.

Looks like we start off

 

[bestsss]

>>1. How does WFYABTA work with teams of AIs?

>>4. Assuming that teams automatically share techs, could this potentially help with the "we are not ready to trade..." mechanic?

Most diplo deals (tech. including) are teams based. Actually most of the stuff is team based but since the usual case (regular games) is one civ = one team it doesn't matter.
WFYABTA is per team, not single civ.

==========
>>2. How does WHEOOH work? Does the team go into WHEOOH, or are both AIs independent with the peaceful member presumably dragged into war?
Again it's solely team based action.

========
>>5. How do war bribes work? If there is a backstabber in a team, could you potentially instigate a bribe through them and force their partner to follow?

I illustrate that w/ some code. not for that case but for declaring war.

Spoiler :

int CvTeamAI::AI_noWarAttitudeProb(AttitudeTypes eAttitude) const
{
	int iProb;
	int iCount;
	int iI;

	iProb = 0;
	iCount = 0;

	for (iI = 0; iI < MAX_PLAYERS; iI++)
	{
		if (GET_PLAYER((PlayerTypes)iI).isAlive())
		{
			if (GET_PLAYER((PlayerTypes)iI).getTeam() == getID())
			{
				iProb += GC.getLeaderHeadInfo(GET_PLAYER((PlayerTypes)iI).getPersonalityType()).getNoWarAttitudeProb(eAttitude);
				iCount++;
			}
		}
	}

	if (iCount > 1)
	{
		iProb /= iCount;
	}

	return iProb;
}

The defined values are the mean of for all players.
The attitude is similiar, with one bad part of us, it's rounded down. I.e. one at pleased, one at cautious yields "cautious".
Thus it depends on the both if they are backstabbers and so.

Spoiler :

AttitudeTypes CvTeamAI::AI_getAttitude(TeamTypes eTeam, bool bForced) const
{
	int iAttitude;
	int iCount;
	int iI, iJ;

	FAssertMsg(eTeam != getID(), "shouldn't call this function on ourselves");

	iAttitude = 0;
	iCount = 0;

	for (iI = 0; iI < MAX_PLAYERS; iI++)
	{
		if (GET_PLAYER((PlayerTypes)iI).isAlive())
		{
			if (GET_PLAYER((PlayerTypes)iI).getTeam() == getID())
			{
				for (iJ = 0; iJ < MAX_PLAYERS; iJ++)
				{
					if (GET_PLAYER((PlayerTypes)iJ).isAlive() && iI != iJ)
					{
						TeamTypes eTeamLoop = GET_PLAYER((PlayerTypes)iJ).getTeam();
						if (eTeamLoop == eTeam || GET_TEAM(eTeamLoop).isVassal(eTeam) || GET_TEAM(eTeam).isVassal(eTeamLoop))
						{
							iAttitude += GET_PLAYER((PlayerTypes)iI).AI_getAttitude((PlayerTypes)iJ, bForced);
							iCount++;
						}
					}
				}
			}
		}
	}

	if (iCount > 0)
	{
		return ((AttitudeTypes)(iAttitude / iCount));
	}

	return ATTITUDE_CAUTIOUS;
}

 

[bestsss]

Excuse the mutli post. I guess open tech would be agriculture-ah->bw (edit, notice i swapped ah w/ bw) and standard worker, warrior opening.
Hopefully we can find some resource nearby and avoid archers in the beginning.

 

[obsolete]

The next question someone is going to ask is how does the spy pt targeting work in teams. The answere is simple. You don't target a single player on a team, you target that entire time simultaneously.

Anyhow one question of my own... I forget now if you settle on a floodplain, if that tile still contributes 0.4 pollution. I would hope not, since you lose any benefit of that tile just by settling on it.

 

[obsolete]

Duckweed, does everyone get 50 turns for the first save before we decide who's to use for the SG?

 

[PaulisKhan]

Settling on floodplain turns it into a desert tile so you don't suffer the penalty.

(now waits for someone to come in with the code proving me wrong >.<)

 

[The Rook]

>>1. How does WFYABTA work with teams of AIs?
>>4. Assuming that teams automatically share techs, could this potentially help with the "we are not ready to trade..." mechanic?

Most diplo deals (tech. including) are teams based. Actually most of the stuff is team based but since the usual case (regular games) is one civ = one team it doesn't matter.
WFYABTA is per team, not single civ.

So does this mean WFYABTA is averaged? Hence civ A with threshold 15 and civ B with threshold 5 would make the team base WFYABTA 10?

>>2. How does WHEOOH work? Does the team go into WHEOOH, or are both AIs independent with the peaceful member presumably dragged into war?
Again it's solely team based action.

So in the case of a warmonger and peacemonger, the two extremes would cancel each other out and create a moderately aggressive team?

>>5. How do war bribes work? If there is a backstabber in a team, could you potentially instigate a bribe through them and force their partner to follow?

The defined values are the mean of for all players.
The attitude is similiar, with one bad part of us, it's rounded down. I.e. one at pleased, one at cautious yields "cautious".

If one team member accepts war bribes at cautious, and the other friendly, would this round to pleased?

I'm probably getting ahead of myself here. Was just curious about how diplomacy in this game would differ from regular play.

 

[The Rook]

Just some thoughts...

I think settle in place and worker first, followed by warrior whilst growning would be safest. It would have been nice to see what settling options there are to the east, but that doesn't seem to be possible without wasting valuable turns. Also, as Duckweed points out, there is likely something useful hidden in the BFC.

Our best food square seems to be sheep. Unfortunately we can't research AH straight away so need Hunting or Agriculture.

With Hunting-AH research would take 20 turns (assuming timings are the same as standard Deity), enough time to build a mine before starting the pasture. We may need Hunting soon anyway for Archery, unless we are lucky and find horse in the BFC (quite possible).

Alternatively, Agriculture-AH would take 23 turns, but we could farm a FP in the mean time, and be nicely ready for the pasture. Also Agriculture is a prerequisite for Pottery, and I suspect cottaging a some FP could be useful considering all the surplus food squares, and the lack of any visible commerce resources. If we go this route and don't find horse in the BFC, then we must head for Hunting-Archery immediately after AH, unless we are happy defending with warriors.

I think the main difference between Hunting-AH and Agriculture-AH is that we could potentially head for BW earlier with the former if there is no horse in the capital. E.g Hunting-AH-BW-Archery.

Could we let the cultural borders absorb the hut? Surely an AI won't swipe it in 5 turns. Then again, I've never seen a warrior get a bad yield from a hut on turn 1.

 

[PaulisKhan]

Warriors can't get badness from a hut on turn one but they can get xp which would otherwise be rerolled (and potentially become a tech) if city borders were to pop the hut.


Re the game: Is the intention to all play the start?
I'm stuck in Japan at least until the 3rd, and potentially up to the 8th.
If I'm home by the 3rd I'd like a chance to play the start too, else just go on without me and I'll check in as soon as I can.

Re start: There's potentially stone or marble on the plains to the SW so the first thing Id be doing is moving my warrior SW. I'm struggling to do any fog gazing with the resolution and contrast on my crappy old laptop.

What are the 3 tiles 3 east of the settler?
As best I can tell, it's plains, desert and forest, so moving 1E is unlikely to gain anything.

Sheep are "meh" at best but it looks like it's all we've got to work with.

I'd go Irrig->AH->BW
and then decide if I had a shot at Oracle (unlikely with AI teams but may depend on which AI plays the turn first) at which point it would be med/poly->PH->.Writing for CoL, else pottery->writing and push for an early GLib to get those GSci's rolling.

In an AI team game, I don't know what process the first AI uses to determine it's tech choice, but the second Ai in the team usually follows their lead.
While we're digging up code, is it concievable that if the first AI in a team is non-religious, they could drag a more religious partner away from religious techs?

 

[bestsss]

Overlays (like floodplains/oasis) are removed upon settling, should be no health diff.

Yes, espionage targets the team which starts w/ 2 palaces, sucks indeed.

>>Then again, I've never seen a warrior get a bad yield from a hut on turn 1.
Becuase you settled 1st (8 tiles away from the capital, if alone, doesn't result in bad huts), let the culture absorb. Can net technology it settled 1st. The best w/ warrior before settling is getting gold.

Here is the code for DENIAL_TECH_WHORE (known as WFYABTA), in short it work by averaging the team member defined values.

Spoiler :

if (eAttitude < ATTITUDE_FRIENDLY)
{
	if ((GC.getGameINLINE().getTeamRank(getID()) < (GC.getGameINLINE().countCivTeamsEverAlive() / 2)) ||
		  (GC.getGameINLINE().getTeamRank(eTeam) < (GC.getGameINLINE().countCivTeamsEverAlive() / 2)))
	{
		iNoTechTradeThreshold = AI_noTechTradeThreshold();//look before it's the mean

		iNoTechTradeThreshold *= std::max(0, (GC.getHandicapInfo(GET_TEAM(eTeam).getHandicapType()).getNoTechTradeModifier() + 100));
		iNoTechTradeThreshold /= 100;

		if (AI_getMemoryCount(eTeam, MEMORY_RECEIVED_TECH_FROM_ANY) > iNoTechTradeThreshold)
		{
			return DENIAL_TECH_WHORE;
		}
	}
.........
}

And the interested function which again results in mean between the team members:

int CvTeamAI::AI_noTechTradeThreshold() const
{
	int iRand;
	int iCount;
	int iI;

	iRand = 0;
	iCount = 0;

	for (iI = 0; iI < MAX_PLAYERS; iI++)
	{
		if (GET_PLAYER((PlayerTypes)iI).isAlive())
		{
			if (GET_PLAYER((PlayerTypes)iI).getTeam() == getID())
			{
				iRand += GC.getLeaderHeadInfo(GET_PLAYER((PlayerTypes)iI).getPersonalityType()).getNoTechTradeThreshold();
				iCount++;
			}
		}
	}

	if (iCount > 0)
	{
		iRand /= iCount;
	}

	return iRand;
}

>>So in the case of a warmonger and peacemonger, the two extremes would cancel each other out and create a moderately aggressive team?
Almost yes, but unit amassing is player based.

Spoiler :

int CvCityAI::AI_buildUnitProb()
{
	int iProb;

	iProb = (GC.getLeaderHeadInfo(getPersonalityType()).getBuildUnitProb() + AI_experienceWeight());

	if (GET_PLAYER(getOwnerINLINE()).AI_isFinancialTrouble())
	{
		iProb /= 2;
	}

	return iProb;
}

>>If one team member accepts war bribes at cautious, and the other friendly, would this round to pleased?

Yes. Also, 'friendly' and 'pleased' combo results into 'pleased'. As a rule of the thumb you should expect averaged behaviour for the team. Tech, production decision are player based. Virtually all the diplo is team based. I think voting (UN/AP) was player based, though (have not checked, I will, if need)

 

[Duckweed]

Duckweed, does everyone get 50 turns for the first save before we decide who's to use for the SG?

I think we'll keep it in SG format from the beginning, 1st 2 or 3 players may play a longer set to a point worth of discussion, such as critical tech (AH and BW) completion and time for 2nd and 3rd city claim.

 

[Duckweed]

Welcome, The Rook.

Regarding your questions, I think Bestsss has given good answers already.
and the answer to your 3rd question is in post #1.

 

[Duckweed]

To maximize the chance of winning this tough challenge, I'd strongly suggest everyone post a plan before playing -- regarding the following issues when applicable

1. tech path
2. civic change
3. city placement
4. diplomatic decision -- especially tech trade, joining the war
5. war plan -- target, force composition, march route.
6. ? feel free to add more.

If there are other opinions than your plan, please include your reasons behind. I'll call a vote when there's major and strong divergence, of course the player has the final decision when there's a tie.


I settled in place since I don't see other choice and there's nothing else.

Here's my plan for opening

Tech:

Agri->Hunting->AH -- we have 2/6 chance to have copper or horse in BFC, unfortunately this also means there's 4/6 of chance we won't have them, and to make things worse, without a commercial tile, we won't have time to find out both, so 5/6 of the chance we have to research Archery for defense against barbarian. Hunting before AH save us ~30 beakers and 3 turns to archery most of the time.

Build:

Worker->warrior

The worker is going to farm one FP and mine the green hill and be ready to pasture the sheep about the time when AH is completed. We'll face a small choice then, which tile do you want your 3rd citizen to work -- unirrigated FP or mined hill (my choice), I'll stop there and leave this choice to next player.

BTW, I'll let warrior claim the hut in the 1st turn, we might get Agri.

@ Pauliskhan

leaving the river costs us a 11 hammer levee and +2 health, which >> having stone or marble in BFC.