SGOTM 10 - Xteam

Can you decipher how good the odds have to be for them to attack one of our units?

Will a pillaging unit avoid other possibilities to beeline to pillage?
 
Can you decipher how good the odds have to be for them to attack one of our units?

Will a pillaging unit avoid other possibilities to beeline to pillage?

Not sure exactly what odds Gandhi will attack with, but he is very risk averse in attacking units according to the XML. By my experience, when you send units to choke the AI, the new units they build wont leave their cities to pillage. Rather, the AI will accumulate units to dislodge the intruders.

I created a new test map, and I chose the option "New Random Seed on Reload" so that every time the test is reloaded, new random stuff can happen. I'm finding that Gandhi usually improves cows first and then the eastern silk. From there, sometimes he roads the western silk, and sometimes he roads directly toward Bombay. Sometimes when he roads toward Bombay, he exposes his worker NW of Bombay. A little later, he exposes multiple workers (2-3) building a fort on the wheat.

Try for yourself.
 

Attachments

"Not sure exactly what odds Gandhi will attack with, but he is very risk averse in attacking units according to the XML. By my experience, when you send units to choke the AI, the new units they build wont leave their cities to pillage. Rather, the AI will accumulate units to dislodge the intruders." That's consistent with my experience. I would add that he will use units in excess of two to escort a settler as well. However, not sure I've any experience with G as an agressive AI. Do you?


Very late here, and RL will interfere until Friday night, but I'll try to test sometime before that.

Does G consistently expose workers on the wheat? Would we need three units to capture three workers? Is this after ivory and even iron are hooked?

Guessing no one has any better idea than I as to the possibilities of acquiring G's map after going to war.
 
I don't think there's much point in worrying about getting Gandhi's map in or after a peace deal. We need to stay at war with him longer than it will take him to circumnavigate. Gandhi will trade maps at any attitude as long as we aren't his worst enemy. We might need to gift him our map after the war to avoid that distinction.
 
"I don't think there's much point in worrying about getting Gandhi's map in or after a peace deal. We need to stay at war with him longer than it will take him to circumnavigate." Well how about during a brief cease fire after, say, 7 turns of war?
 
Did a little testing.

Gandhi built a 2nd worker and started roading to Bombay followed by improvement of the Ivory. In the test I declared and captured a worker on the ivory tile. Gandhi started building barracks so I gave him a caravel when the barracks were done.

I noticed the following:
  1. Gandhi does not attack longbows that are fortified in a forest or on a hill.
  2. He will attack longbows on flatland - particularly if he has a catapult.
  3. He will also attack an explorer on a forested hill.
  4. War declaration will get him to stop building settler and start building military. In the test his first settler was postponed by ~15 turns.
  5. Getting peace again is not a problem and Gandhi will go back to cautious. But he will not trade open borders or map after peace.
  6. He will eventually start sending longbows towards our cities but will keep 2-3 longbows in his own cities.
  7. Gandhi's caravel doesn't sail with any clear purpose confirming my initial guess that the AI doesn't know how to get circumnavigation - it just happens by coincidence. The caravel didn't turn back either - perhaps it has been given an explore command it will stick to.

I think we still have a decent chance of getting circumnavigation because Gandhi is most likely not going the direct route. He might as well go north or south at some point.

I don't think we need to declare immediately - I'm sure there will be more chances to capture a worker. But it does seem like we should declare before Gandhi gets access to ivory and/or iron.
 
Did a little testing.

Gandhi built a 2nd worker and started roading to Bombay followed by improvement of the Ivory. In the test I declared and captured a worker on the ivory tile. Gandhi started building barracks so I gave him a caravel when the barracks were done.

So we should send our other LB to Gandhi's northern borders if we aren't declaring immediately.

I noticed the following:
He will also attack an explorer on a forested hill.

He will probably lose this battle right? The explorer would have a 140% defensive bonus with fortifying for 5 turns.

[*]Getting peace again is not a problem and Gandhi will go back to cautious. But he will not trade open borders or map after peace.

This is probably because we will be his worst enemy, at least until we gift him something valuable or the negative diplo points start to build with the other AI because of religious differences.

Gandhi's caravel doesn't sail with any clear purpose confirming my initial guess that the AI doesn't know how to get circumnavigation - it just happens by coincidence. The caravel didn't turn back either - perhaps it has been given an explore command it will stick to.

It most likely has an EXPLORE command. A caravel can't pillage after all. I still don't trust your conclusion that Gandhi isn't trying for circumnavigation. From what our map trading shows, he has at least one and probably two ships beelining east/west. Maybe in the real game, Gandhi randomly decided to prioritize circumnavigation and then consequently decided to build the ships. Giving Gandhi a ship in worldbuilder might not produce the same behavior.

I think we still have a decent chance of getting circumnavigation because Gandhi is most likely not going the direct route. He might as well go north or south at some point.

If Gandhi starts going north/south, then it won't hurt us to first contact the Americans with our caravel and then turn west.
 
"I don't think there's much point in worrying about getting Gandhi's map in or after a peace deal. We need to stay at war with him longer than it will take him to circumnavigate." Well how about during a brief cease fire after, say, 7 turns of war?

If we get peace/CF and then redeclare, we'll have -6 diplo with Gandhi for declaring war on him. That will be hard to recover from with religious difference thrown in soon. I don't think we should burn our bridges with such a willing tech trader.
 
ShannonCT said:
So we should send our other LB to Gandhi's northern borders if we aren't declaring immediately.

Yes I think so. From there we can also monitor when Gandhi starts improving the ivory tile.

ShannonCT said:
He will probably lose this battle right? The explorer would have a 140% defensive bonus with fortifying for 5 turns.

Yes, the explorer won a couple of battles but finally lost to a dual attack with first a cat and then a longbow.

ShannonCT said:
It most likely has an EXPLORE command. A caravel can't pillage after all. I still don't trust your conclusion that Gandhi isn't trying for circumnavigation. From what our map trading shows, he has at least one and probably two ships beelining east/west. Maybe in the real game, Gandhi randomly decided to prioritize circumnavigation and then consequently decided to build the ships. Giving Gandhi a ship in world builder might not produce the same behavior.

I think you are attributing too much intelligence to the AI. The truth is that the AI is extremely dumb. As you say the caravel must have been given the UNITAI_EXPLORE command and if you look in the world builder you will see that there doesn't exist any UNITAI_CIRCUMNAVIGATION command. Thus it is highly unlikely that the caravel(s) are doing anything but exploring.

I suggest that you play another turn and re-buy Gandhi's map so that we can identify his caravel movement. I think we can play for both circumnavigation and slowing down Gandhi by a well timed DOW.
 
I'm not a C++ expert, but I see something in the CvUnitAI.cpp file that suggests that the availability of the circumnavigation bonus affects how AI exploring units decide to move.

Spoiler :

bool CvUnitAI::AI_exploreRange(int iRange)
{
PROFILE_FUNC();

CvPlot* pLoopPlot;
CvPlot* pAdjacentPlot;
CvPlot* pBestPlot;
CvPlot* pBestExplorePlot;
int iSearchRange;
int iPathTurns;
int iValue;
int iBestValue;
int iDX, iDY;
int iI;

iSearchRange = AI_searchRange(iRange);

iBestValue = 0;
pBestPlot = NULL;
pBestExplorePlot = NULL;

int iImpassableCount = GET_PLAYER(getOwnerINLINE()).AI_unitImpassableCount(getUnitType());

for (iDX = -(iSearchRange); iDX <= iSearchRange; iDX++)
{
for (iDY = -(iSearchRange); iDY <= iSearchRange; iDY++)
{
PROFILE("AI_exploreRange 1");

pLoopPlot = plotXY(getX_INLINE(), getY_INLINE(), iDX, iDY);

if (pLoopPlot != NULL)
{
if (AI_plotValid(pLoopPlot))
{
iValue = 0;

if (pLoopPlot->isRevealedGoody(getTeam()))
{
iValue += 100000;
}

if (!(pLoopPlot->isRevealed(getTeam(), false)))
{
iValue += 10000;
}

for (iI = 0; iI < NUM_DIRECTION_TYPES; iI++)
{
PROFILE("AI_exploreRange 2");

pAdjacentPlot = plotDirection(pLoopPlot->getX_INLINE(), pLoopPlot->getY_INLINE(), ((DirectionTypes)iI));

if (pAdjacentPlot != NULL)
{
if (!(pAdjacentPlot->isRevealed(getTeam(), false)))
{
iValue += 1000;
}
}
}

if (iValue > 0)
{
if (!(pLoopPlot->isVisibleEnemyUnit(this)))
{
PROFILE("AI_exploreRange 3");

if (GET_PLAYER(getOwnerINLINE()).AI_plotTargetMissionAIs(pLoopPlot, MISSIONAI_EXPLORE, getGroup(), 3) == 0)
{
PROFILE("AI_exploreRange 4");

if (!atPlot(pLoopPlot) && generatePath(pLoopPlot, MOVE_NO_ENEMY_TERRITORY, true, &iPathTurns))
{
if (iPathTurns <= iRange)
{
iValue += GC.getGameINLINE().getSorenRandNum(10000, "AI Explore");

if (pLoopPlot->isAdjacentToLand())
{
iValue += 10000;
}

if (pLoopPlot->isOwned())
{
iValue += 5000;
}

if (!isHuman())
{
int iDirectionModifier = 100;

if (AI_getUnitAIType() == UNITAI_EXPLORE_SEA && iImpassableCount == 0)
{
iDirectionModifier += (50 * (abs(iDX) + abs(iDY))) / iSearchRange;
if (GC.getGame().circumnavigationAvailable())
{
if (GC.getMap().isWrapX())
{
if ((iDX * ((AI_getBirthmark() % 2 == 0) ? 1 : -1)) > 0)
{
iDirectionModifier *= 150 + ((iDX * 100) / iSearchRange);
}
else
{
iDirectionModifier /= 2;
}
}
if (GC.getMap().isWrapY())
{
if ((iDY * (((AI_getBirthmark() >> 1) % 2 == 0) ? 1 : -1)) > 0)
{
iDirectionModifier *= 150 + ((iDY * 100) / iSearchRange);
}
else
{
iDirectionModifier /= 2;
}
}
}
iValue *= iDirectionModifier;
iValue /= 100;
}
}

 
Just have no time to test, but glad Fred had a go at it -- useful info once war starts. Please be more specific re: "eventually start sending longbows towards our cities." Hope leif is able to have a go as well.

"In the test I declared and captured a worker on the ivory tile." How did you manage this? Was one ivory already hooked up, and then you attacked when the worker moved to a perimeter tile? Were there two workers at this stage?

"If we get peace/CF and then redeclare, we'll have -6 diplo with Gandhi for declaring war on him. That will be hard to recover from with religious difference thrown in soon. I don't think we should burn our bridges with such a willing tech trader." Concur.

Don't understand why getting G's map for only one more turn would be very useful. Indeed, I'm wondering how many turns before it would likely be useful.
 
Just have no time to test, but glad Fred had a go at it -- useful info once war starts. Hope leif is able to have a go as well.
Interesting discussion, and I have been following it. Sorry I haven't had much time to test either as I have been trying to get WOTM25 out the door. Another day or so and I should be finished.

While I know we would like to achieve both Circumnavigation and a war to hold Ghandi back, I would like to ask one dumb question. :hmm:
Which one of these tasks is the more important one? It would seem to me that holding Ghandi back and gaining space for city sites would be more important than circumnavigation? If this is true, than our primary consideration should be Ghandi and if we get circumnavigation, that would be a bonus?
I have been known to be wrong before... :blush:
 
While I know we would like to achieve both Circumnavigation and a war to hold Ghandi back, I would like to ask one dumb question. :hmm:
Which one of these tasks is the more important one? It would seem to me that holding Ghandi back and gaining space for city sites would be more important than circumnavigation? If this is true, than our primary consideration should be Ghandi and if we get circumnavigation, that would be a bonus?
I have been known to be wrong before... :blush:

Holding Gandhi back is more important but we should try to do both. G has only built roads on the cows and silk so he's still quite a few turns from hooking ivory. Our explorer can keep tabs on road progress so we will know when we absolutely must declare war to prevent ivory hookup. What we're really talking about is betting a free worker against circumnavigation. If we can keep trading maps with Gandhi for the next 5 turns, our chance at circumnavigation is greatly improved, but our chance for a free worker is diminished (although there is some chance for 2 free workers by delaying war).

I feel like having an extra movement point for ships for the rest of the game will prove more useful than getting the extra worker now. The stolen worker wont help hook iron or our ivory faster so it doesn't have much strategic importance. We can whip an extra worker for 1 pop in Moscow whenever we need one so stealing a worker is like gaining 1 pop in Moscow.

CP, I agree that getting G's map for one more turn isn't much use. We already know that G's caravel is sailing east (it has been for the last 5-6 turns) and it's highly likely that there's another boat sailing west. Gandhi could circumnavigate in as little as 5 turns. We can contact the Americans, Chinese, and English by T23. So things will be a lot clearer 4 turns from now.
 
"I feel like having an extra movement point for ships for the rest of the game will prove more useful than getting the extra worker now. Agreed, but at this point the worker is certain, the circum is not.The stolen worker won't help hook iron or our ivory faster so it doesn't have much strategic importance. We can whip an extra worker for 1 pop in Moscow whenever we need one so stealing a worker is like gaining 1 pop in Moscow." We also have forests to chop before they are expanded onto, and more quickly denying G use of the pastured cow, and plantation silk probably has a little strategic significance in preventing G's expansion, plus one pop does set research back a bit.
 
I think you are attributing too much intelligence to the AI. The truth is that the AI is extremely dumb. As you say the caravel must have been given the UNITAI_EXPLORE command and if you look in the world builder you will see that there doesn't exist any UNITAI_CIRCUMNAVIGATION command. Thus it is highly unlikely that the caravel(s) are doing anything but exploring.

It looks to me from the map that Gandhi's caravel that is now going east left Delhi and headed *north*, bumped into land and turned to follow the coast eastwards, meeting England, China and then America. I think where the caravel goes is determined by the shape of the map rather than any desire to circumnavigate. If the continent on which America is found has a coast that heads east for a while, we could be in trouble getting circumnavigation because the caravel will beeline that direction. The coast could easily bend south or north too, sending that caravel in a circle.

Hobbling Gandhi is clearly the more important objective, but declaring war now has more than just circumnavigation implications. SCT pointed out this:

ShannonCT said:
If we get peace/CF and then redeclare, we'll have -6 diplo with Gandhi for declaring war on him. That will be hard to recover from with religious difference thrown in soon. I don't think we should burn our bridges with such a willing tech trader

It is this which encourages me to be willing to delay a war declaration at least a few turns to see what more map turns up over the next four turns or so after Gandhi's caravel moves a little more, and we meet China/America. The later worker steal is not certain, but then again we've also seen that an extra worker is not worth a lot to us right now either. I think this discussion will be a good deal clearer in a few turns.

We should not forget though that Gandhi must be hobbled. The negative effect to us of him getting a roaring start is huge.
 
Just have no time to test, but glad Fred had a go at it -- useful info once war starts. Please be more specific re: "eventually start sending longbows towards our cities." Hope leif is able to have a go as well.

"In the test I declared and captured a worker on the ivory tile." How did you manage this? Was one ivory already hooked up, and then you attacked when the worker moved to a perimeter tile? Were there two workers at this stage?

"If we get peace/CF and then redeclare, we'll have -6 diplo with Gandhi for declaring war on him. That will be hard to recover from with religious difference thrown in soon. I don't think we should burn our bridges with such a willing tech trader." Concur.

Don't understand why getting G's map for only one more turn would be very useful. Indeed, I'm wondering how many turns before it would likely be useful.

If the power stays on throughout today (ie the next 8-10 hours) I'll probably have a chance to have a play with SCT's latest test to say more of what Fred told us, or possibly a different scenario depending on what happens. I have some work to do first but if the power stays good I'll get it done with plenty of time for play afterwards!
 
I've had a chance to play through a few turns a few times over. I don't tihnk I duplicated SCT's moves from T1 exactly, but it shouldn't make much difference to what Gandhi does. I was interested in seeing how likely it was we could achieve both goals of achieving circumnavigation AND then declaring on Gandhi to steal workers or otherwise hobble him.

I noticed that Gandhi works quickly to develop an ivory or two, and also moves to develop the wheat fairly regularly. I noticed that if we *don't* declare on him now, he generally does not protect his worker(s) for quite a while - at least until the circumnavigation issue is over and done with. Small maps are not very big!

So I think a later worker steal is very much a possibility if we have the LBs in the right place - generally hanging around the elephants and the wheat seem likely places. One in each perhaps?

I also noticed that it's not the same leaders in that test file as in the real game, and this is a big deal because some leaders swap maps more easily than others. This is a big deal because if we can get either the Chinese leader, American, leader or English leader to sell us a map which goes far enough west that IF Gandhi's caravel going east does not get sidetracked north or south and continues to heads east, that combining an American, English or Chinese map with a map we buy from Gandhi when his caravel has gone far enough, will get us circumnavigation as long as Gandhi does NOT buy the same English,American or Chinese map that we do!

My testing therefore doesn't really help us with the circumnavigation thing except to say that we can get it if (and probably only if) we get a map out of one of those three northern neighbours that Gandhi does not get, and we keep paying Gandhi 5 gold (or whatever) every turn or two to get regulr updates on his map so that we put two maps together to get circumnavigation before Gandhi gets his caravel around the world.

I think the map shape will strongly determine whether Gandhi's caravel gets around the world or gets sidetracked.

I think our caravel should go to meet China and America at least (England is a bit further away) and we should try buying a map from them if they will offer it. I think it's only about four turns until we can do that. We'll know much more then!
 
ShannonCT said:
I'm not a C++ expert, but I see something in the CvUnitAI.cpp file that suggests that the availability of the circumnavigation bonus affects how AI exploring units decide to move.

The way I interpret this is that the AI will explore further from home as long as the circumnavigation bonus is available. So it has no direct effect on how the unit moves but it will allow the unit to move further away (i.e. increased exploration range). This is also consistent with what I saw in my test - basically random movements of the caravel into the fog. In any case it's not the product of any "plan" laid by the AI but rather a common feature of the initial exploration of any AI.

Cactus Pete said:
Don't understand why getting G's map for only one more turn would be very useful. Indeed, I'm wondering how many turns before it would likely be useful.

The information is useful because we can compare with the map we bought this turn and determine from the difference in defogged tiles how many boats Gandhi has and where they are moving. We must buy his map again before we buy the map of anyone else to be able to do this.

Mad Professor said:
I think the map shape will strongly determine whether Gandhi's caravel gets around the world or gets sidetracked.

I think so too. It's particularly true if Gandhi has a galley going west because it's forced to follow the coast.
 
I've been messing around a bit with random renaissance starts plus Gandhi worldbuildered in and I give him a caravel and see what he does with it.

The results are very varied, to the point of bewildering. Sometimes the caravel wanders around aimlessly, and other times it seems to beeline either eastwards or westwards. I don't understand what directive is driving it other than a combination of random exploration plus fortunate or unfortunate land placement turning the caravel's nose.

That doesn't help us much other than to say that Gandhi *might* beeline eastwards and circumnavigate quickly. Or not.

Having said that, in the real game, we can still circumnavigate first if we are lucky enough to buy a map which gives us a look at territory further west than Gandhi has, and we keep buying his map every couple of turns so that even if his caravel flukes a direct eastward march we get his progress eastwards and add it to the other map we buy to get circumnavigation before Gandhi completes the circuit.

If we can't buy a map off England, China or America, we will only get circumnavigation if Gandhi's caravel wanders aimlessly. If we can buy a map which shows something westward, then we'll get it no matter what Gandhi's caravel does as long as we continue to have the option to buy a map from Gandhi, which means delaying a war declaration. A war declaration now will make circumnavigation a much smaller possibility,

Either way I think the first thing our caravel should do is go meet at least China and America (because they are closer) and we try buy a map and see what that gives us! If nothing then the next best bet is to sail the caravel to England and see if we can buy a map there. We need a map that goes further west than Gandhi's map.

Surely we can delay the declaration a few turns to see what map we can buy from the northerners while still keeping an eye on Gandhi's workers and other units to see what he's building? The instant we have circumnavigation we can look for the first opportunity to steal a worker (or more than one if he's built another by then). I suggest having a LB hang around the southern side of Gandhi near the silk and wheat and another hang around the north side near the elephants, and use the explorer to follow the workers around inside his territory where the LB's can't see far enough to see the workers.
 
Back
Top Bottom