SGOTM 14 - One Short Straw

[shyuhe]

I believe LC is up, unless he's too busy.

 

[Dhoomstriker]

No, we need to get both resources; otherwise the second one is there for an AI to trade for a mystery 3rd resource.

Possibly, possibly not. Anyway, the idea is not to trade away any Resources that we would need in the next 10 turns... I have no access to the game to check, but I imagine that it might be possible that we could do away with Resources for such a period of time. I see no point in trading into Unhappiness or Unhealthiness.

I don't quite see where the link is between resource trading and gold limits.
The amount of gold they'll trade us depends on their population, gold per turn, etc, and how much gold they've traded to us, but not resource trade turns.
I don't believe us giving them resources will let them gift us more gold - as far as I can tell it's just a running total of how much they've given us free.

Well, when I was testing before, gifting or not gifting Resources made a difference. It shouldn't be too hard of a test to set up:

Spoiler :

1. Get an AI (or an AI team in this case) to be Pleased with us (such as using the World Builder Diplomacy Attitudes editing functionality or gifting techs)
2. Give yourself a bunch of extra connected Resources (a quick way is to put them on the same squares as your Cities are placed, but you can also use the World Builder to build an improvement)... just make sure that you have the relevant tech (such as Calendar) to be able to access said Resources
3. Optionally, use the World Builder with an improvement type selected and right-click on many of an AI's Resources, just to remove their Camps, Farms, Pastures, etc, giving you more trading options
4. Save the game as "save A" (call it whatever you want, I'll just pick that name here so that I can be clearer in what I am suggesting)
5. End the turn 10 times and save the game as "save B"
6. Load "save A," gift a bunch of Resources (like 5 of them) to an AI--maybe to both team members, even, since it's a team-based game, save the game as "save C" although you probably won't need to load this game
7. End the turn 10 times and save the game as "save D"
8. Play around with "save B" and "save D" to find out what the Request a lump sum of Gold threshold is and see if the value is different between the two saved games. My testing in the past showed that the values were different.

Yeah, sometimes code-diving can be more conclusive but this testing shouldn't take too long to set up (for anyone that has access to Civ 4, which I do not have).

If we don't need the health and definitely won't, then fine, but the only benefit is +0.2 attitude for one AI, which is worthless unless there is a specific AI we want to please.

Please correct me if I am wrong, but there is at least one team that won't trade with us until we get them up to an average of a Pleased attitude toward us, right? If that situation is the case, then said AI team is our preferred target team of AIs to please.

You're right. I took getPrereqAndTech() to mean any of the prereqs, when it's really just the first one. (The other ones are hiding behind the function getPrereqAndTechs() )

So, does that fact change the weighting to be in favour of Horseback Riding being easier to get in trade than Construction?

Begging for gold? We need to be really careful with that, since we don't want a peace treaty with the wrong person at the wrong time.

Certainly, we wouldn't just go making a Demand/Request all willy-nilly, particularly sine once you ask, you can't ask again and expect to be successful for a minimum of 10 turns and more likely to be closer to 30 to 50 turns later. However, we can also potentially use such "no war for 10 turns" timing to our advantage to avoid being dog-piled at the right time, should an opportunity present itself.

 

[babybluepants]

My understanding of what ZPV was saying yesterday was that HBR is more difficult to get in trade. TBH, that doesn't necessarily correspond with my playing experience. It's much easier to trade for in most games, though that possibly has to do with more AIs getting it sooner.

The main reason I want HBR first is so we can build a stable in Ivory or Washington and get an 8XP unit capturing the barb city, though. The trade stuff is a crapshoot anyway, since we don't actually know which will be available or when.

LC played just before me.

 

[Mitchum]

That means mdy is up, right? I say we give him 24 hours to "get it" before moving on to shyuhe. Since he hasn't posted for a while, you may want to start preparing, shyuhe...

 

[shyuhe]

you're right bbp. I'm going to be out of town this weekend so I won't be able to really dig in until early next week.

 

[ZPV]

My understanding of what ZPV was saying yesterday was that HBR is more difficult to get in trade. TBH, that doesn't necessarily correspond with my playing experience. It's much easier to trade for in most games, though that possibly has to do with more AIs getting it sooner.

The main reason I want HBR first is so we can build a stable in Ivory or Washington and get an 8XP unit capturing the barb city, though. The trade stuff is a crapshoot anyway, since we don't actually know which will be available or when.

LC played just before me.

So, does that fact change the weighting to be in favour of Horseback Riding being easier to get in trade than Construction?

Yeah, my earlier calculation about HBR was wrong, and HBR is a little easier to get in trade than Construction. I still want the early stable for the 8XP unit though.

Possibly, possibly not. Anyway, the idea is not to trade away any Resources that we would need in the next 10 turns... I have no access to the game to check, but I imagine that it might be possible that we could do away with Resources for such a period of time. I see no point in trading into Unhappiness or Unhealthiness.

Well, when I was testing before, gifting or not gifting Resources made a difference. It shouldn't be too hard of a test to set up:

Spoiler :

1. Get an AI (or an AI team in this case) to be Pleased with us (such as using the World Builder Diplomacy Attitudes editing functionality or gifting techs)
2. Give yourself a bunch of extra connected Resources (a quick way is to put them on the same squares as your Cities are placed, but you can also use the World Builder to build an improvement)... just make sure that you have the relevant tech (such as Calendar) to be able to access said Resources
3. Optionally, use the World Builder with an improvement type selected and right-click on many of an AI's Resources, just to remove their Camps, Farms, Pastures, etc, giving you more trading options
4. Save the game as "save A" (call it whatever you want, I'll just pick that name here so that I can be clearer in what I am suggesting)
5. End the turn 10 times and save the game as "save B"
6. Load "save A," gift a bunch of Resources (like 5 of them) to an AI--maybe to both team members, even, since it's a team-based game, save the game as "save C" although you probably won't need to load this game
7. End the turn 10 times and save the game as "save D"
8. Play around with "save B" and "save D" to find out the Request a lump sum of Gold threshold is and see if the value is different. My testing in the past showed that the values were different.

Yeah, sometimes code-diving can be more conclusive but this testing shouldn't take too long to set up (for anyone that has access to Civ 4, which I do not have).

Please correct me if I am wrong, but there is at least one team that won't trade with us until we get them up to an average of a Pleased attitude toward us, right? If that situation is the case, then said AI team is our preferred target team of AIs to please.

Ragnar/Asoka is the team that won't OB or trade tech until pleased. I just don't think it's feasible to get them Pleased, and we're going to be DoWing them sooner rather than later.

The power ratio, the number of turns we've known them, and whether or not they are a land target are what determine how much gold they are willing to gift us.

Certainly, we wouldn't just go making a Demand/Request all willy-nilly, particularly sine once you ask, you can't ask again and expect to be successful for a minimum of 10 turns and more likely to be closer to 30 to 50 turns later. However, we can also potentially use such "no war for 10 turns" timing to our advantage to avoid being dog-piled at the right time, should an opportunity present itself.

 

[babybluepants]

I think we need to work out how much infra we want in each city already. I'd encourage everyone to quickly play this out to Astronomy. It comes up fast (25-30t from now), and our secondary cities are hardly even getting base infra (granary, lh, maybe library and barracks) done by then.

 

[LowtherCastle]

We still don't know where Oz is, of course, so our exploring wb should scout the perimeter in addition to establishing TRs. After connecting to Gandhi, I think it should go back north and continue looping around. Unless a baarb city blocks it, this will also give us TRs in the event of DoWing Asoka.

Currently we have exposed 59 of the 64 east-west tiles and 25 of the 40 north-south tiles.

Spoiler :

 

[LowtherCastle]

Ivory needs stables. Our galleon-producing city does not need barracks.

 

[LowtherCastle]

That narrows Gandhi down to the personalities of Roosevelt, Asoka or Cyrus and confirms Montezuma behaves like Mansa Musa.
It also means that Genghis behaves like Brennus or Montezuma.

Brennus has OrgRel as his favorite civic. The chart of attitudes doesn't show him giving anyone plusmods for that, though 5 AIs also have it. I don't know if the chart hides that because of random personalities, but maybe you can figure out Genghis based on this. It would be nice know if the chart reveals that detail or not, because AIs will start getting Monarchy soon and switch to HeredRule.

Speaking of HeredRule, all but Gandhi/Monte can research Monarchy. That's a problem for us if we want them to research HBR or Construction for us instead. If we can figure out who has HeredRule for a favorite civic we can get a better idea how many are less likely to research Monarchy.

IF we have a strong chance of getting both COnstr and HBR in trade, going Compass next would be great. It's important to decide now because we're gaining non-trivial beakers even at 0% slider. Furthermore, at 30bpt, we get more value on Compass than HBR or COnstruction.

ZPV, based on your latest calcs for HBR, how many AIs will need to know it to trade it (your best guesstimate).

 

[ZPV]

Brennus has OrgRel as his favorite civic. The chart of attitudes doesn't show him giving anyone plusmods for that, though 5 AIs also have it. I don't know if the chart hides that because of random personalities, but maybe you can figure out Genghis based on this. It would be nice know if the chart reveals that detail or not, because AIs will start getting Monarchy soon and switch to HeredRule.

Speaking of HeredRule, all but Gandhi/Monte can research Monarchy. That's a problem for us if we want them to research HBR or Construction for us instead. If we can figure out who has HeredRule for a favorite civic we can get a better idea how many are less likely to research Monarchy.

IF we have a strong chance of getting both COnstr and HBR in trade, going Compass next would be great. It's important to decide now because we're gaining non-trivial beakers even at 0% slider. Furthermore, at 30bpt, we get more value on Compass than HBR or COnstruction.

ZPV, based on your latest calcs for HBR, how many AIs will need to know it to trade it (your best guesstimate).

It doesn't hide favourite civic bonuses, once they are there. Montezuma/Brennus just doesn't feel right though. I'll think about it some more.

Just one other for Gandhi/Montezuma, probably one other for the others, but it could be two.

 

[Dhoomstriker]

Won't a Favourite Civic selection come from what appears on the F4 Info screen?

I.e. Rather than from their personality, won't the Favourite Civic, just like Leader Traits, come from the Leader whose picture we see? I.e. Leader picture + Traits + Favourite Civic should all be known, while it is only the aspects of a personality (monopoly trading threshold, willingness to talk after a war has begun, willingness to Open Borders, etc) which we have to figure out.


As for trading for Horseback Riding and/or Construction, I guess the important determining factor is "how soon do we need those techs for the respective buildings or units that they unlock"? If can figure out the earliest date that we would want to build a Stable and/or a Catapult, then we can work backward from there and only switch to researching the relevant tech instead of Compass "at the last minute possible."

 

[LowtherCastle]

F4 Info Screen shows questions marks, so I think it's considered personality.

Btw, another phenomenon I noticed today. Genghis is pleased with us and thus we can ask him for friendly favors, whereas Gandhi is cautious so we can only threaten him to give us stuff. For some reason that is not affected by the average of their two attitude characteristics on that...

 

[Dhoomstriker]

F4 Info Screen shows questions marks, so I think it's considered personality.

Ahhh, so that fact WILL make for a relatively easier way of helping to confirm whether a leader has the personality that we suspect them of having.

Btw, another phenomenon I noticed today. Genghis is pleased with us and thus we can ask him for friendly favors, whereas Gandhi is cautious so we can only threaten him to give us stuff. For some reason that is not affected by the average of their two attitude characteristics on that...

Very neat... if we want peace with a team, we can always hit up a single member of said team, as long as said member still likes us sufficiently. It's probably worth testing just to confirm that we can get 10 turns of peace being enforced with both team members, confirming whether or not we can request a gift from BOTH team members during overlapping times, and confirming whether or not the team member that isn't Pleased with us will get upset for our Demand.


I still think that gifting a Resource will increase the amount of lump sum Gold that we can request. The code may not show so directly but such a gift may increase one variable that gets used in the Gift Requesting function.

 

[ZPV]

I still think that gifting a Resource will increase the amount of lump sum Gold that we can request. The code may not show so directly but such a gift may increase one variable that gets used in the Gift Requesting function.

The gift requesting code is very simple. I'll paste it below:

Spoiler :

if (iOurValue > 0 && 0 == pTheirList->getLength() && 0 == iTheirValue)
	{
		if (GET_TEAM(getTeam()).isVassal(GET_PLAYER(ePlayer).getTeam()) && CvDeal::isVassalTributeDeal(pOurList))
		{
			if (AI_getAttitude(ePlayer, false) <= GC.getLeaderHeadInfo(getPersonalityType()).getVassalRefuseAttitudeThreshold()
				&& GET_TEAM(getTeam()).getAtWarCount(true) == 0
				&& GET_TEAM(GET_PLAYER(ePlayer).getTeam()).getDefensivePactCount() == 0)
			{
				iOurValue *= (GET_TEAM(getTeam()).getPower(false) + 10);
				iOurValue /= (GET_TEAM(GET_PLAYER(ePlayer).getTeam()).getPower(false) + 10);
			}
			else
			{
				return true;
			}
		}
		else
		{
			if (AI_getAttitude(ePlayer) < ATTITUDE_PLEASED)
			{
				if (GET_TEAM(getTeam()).getPower(false) > ((GET_TEAM(GET_PLAYER(ePlayer).getTeam()).getPower(false) * 4) / 3))
				{
					return false;
				}
			}

			if (AI_getMemoryCount(ePlayer, MEMORY_MADE_DEMAND_RECENT) > 0)
			{
				return false;
			}
		}

		iThreshold = (GET_TEAM(getTeam()).AI_getHasMetCounter(GET_PLAYER(ePlayer).getTeam()) + 50);

		iThreshold *= 2;

		if (GET_TEAM(GET_PLAYER(ePlayer).getTeam()).AI_isLandTarget(getTeam()))
		{
			iThreshold *= 3;
		}

		iThreshold *= (GET_TEAM(GET_PLAYER(ePlayer).getTeam()).getPower(false) + 100);
		iThreshold /= (GET_TEAM(getTeam()).getPower(false) + 100);

		iThreshold -= GET_PLAYER(ePlayer).AI_getPeacetimeGrantValue(getID());

		return (iOurValue < iThreshold);
	}

The only possible thing gifting a resource could change is AI_getPeacetimeGrantValue. This happens in CvDeal.cpp:addTrades and CvDeal.cpp:doTurn, but it only affects the player receiving a free gift. The doTurn variant will be affected, if we are receiving a free resource from them, but that's it.
(it's also affected by gifting a unit, but that's somewhat irrelevant).

 

[Dhoomstriker]

The only possible thing gifting a resource could change is AI_getPeacetimeGrantValue. This happens in CvDeal.cpp:addTrades and CvDeal.cpp:doTurn, but it only affects the player receiving a free gift. The doTurn variant will be affected, if we are receiving a free resource from them, but that's it.
(it's also affected by gifting a unit, but that's somewhat irrelevant).

Well, it's late and I'm tired, but if an AI "received" a gift, wouldn't that mean affecting their getPeacetimeGrantValue variable, which translates to them wanting to give us more? I'll be happy to be proven wrong but it can't hurt to at least try it out, right?

 

[ZPV]

Well, it's late and I'm tired, but if an AI "received" a gift, wouldn't that mean affecting their getPeacetimeGrantValue variable, which translates to them wanting to give us more? I'll be happy to be proven wrong but it can't hurt to at least try it out, right?

Us giving them a gift increases "them.AI_PeacetimeGrantValue(Us).
The value that is used is "us.AI_PeacetimeGrantValue(them)", and we don't want to increase it since it is subtracted from the threshold.

 

[Dhoomstriker]

Us giving them a gift increases "them.AI_PeacetimeGrantValue(Us).
The value that is used is "us.AI_PeacetimeGrantValue(them)", and we don't want to increase it since it is subtracted from the threshold.

Are you sure? That kind of coding doesn't make any sense... I'll give you some things and be nice to you so that you will be meaner to me in the future.

 

[ZPV]

Are you sure? That kind of coding doesn't make any sense... I'll give you some things and be nice to you so that you will be meaner to me in the future.

It's "Have I already given you lots of nice stuff?
Then I'd better not give you any more."

 

[Mitchum]

shyuhe, are you planning to play the next turnset?