Test Game attached at start of t46. Includes a new Chariot that will kill our Warrior t47, if it takes the Worker t46.
EDIT: Attached save has Capitalist unhealthy t44 and t45 and 2F less. Also, made gift of Gold and Fish to Joao II to make Capitalist unhealthy in same way it was unhealthy in the real game. It is important to make the test game match the real game as much as possible as the issue with Capitalist remaining unhealthy unnoticed for two turns illustrates.
Suppose we lost the Warrior and captured Worker and we're still at war with Genghis Khan. Would we better off than we are now? Sorry, I missed where it was suggested that the Wheat trade should be canceled if Capitalist became unhealthy. I suppose the test game is now off because Capitalist in it did not become unhealthy.
The library whip will still be nice although now we spend a turn at size 2 instead of size 3, so we may have to work the Gold or we'll end up with Currency T52.
The library whip will still be nice although now we spend a turn at size 2 instead of size 3, so we may have to work the Gold or we'll end up with Currency T52.
The library whip will still be nice although now we spend a turn at size 2 instead of size 3, so we may have to work the Gold or we'll end up with Currency T52.
You can't be serious about taking the Worker now. The just built Chariot will kill both Warrior and captured Worker on t47 no matter which city built the Chariot. Or has the Chariot already been moved out of range?
We talked about it a while ago when the wheat trade was made. At the time, we noted that our cities would not need it for at least 10 turns. This is not a major error as long as we make sure we get Currency T51.
But it brings up a point:
It's important that we all have our alert messages set. You should have gotten a big warning that Capitalist was unhealthy (along with it grew to size 4). It's in the log, but I'm not sure that means you have the warnings set to appear at the top of the screen each turn.
You can't be serious about taking the Worker now. The just built Chariot will kill both Warrior and captured Worker on t47 no matter which city built the Chariot. Or has the Chariot already been moved out of range?
1) Wait 2t turns and try poaching a Worker t48. If a Chariot is in either of Genghis Khan's cities now (t46), it will kill both our Warrior and captured Worker in the AI's part of t46.
2) I'd suggest whipping the Capital 1t later, so it will be 3P after the whip.
Ok, what outcome do you expect after poaching the Worker? Genghis Khan will counter attack, if either any of his ATTACK Archers or his Chariot is in range. Is your plan to sacrifice our Warrior and be a war with Genghis Khan again to just kill his 2nd Worker? Or do you have good reason to believe that Genghis Khan's attack units will be out of range and thus the Worker steal will completely succeed?
After much facepalmery I finally decided to read the rest of this line
..... and realised that its the game speed tech cost multiplier, as in 67 for quick, 100 for normal speed, 150 for Epic, and 300 for marathon, explains where the /100 vanished to
Pretty sure i've got this figured out now.
The amount of gold they are willing to trade is the lowest value of either the iMaxGold in the spoilered code below
which is AIs total gold - (12*defecit * Speed modifier)/100.
The actual slider percentage is irrelevant
The iMaxGold I mentioned earlier acts as a cap defining the absolute max the AI is willing to trade away in a turn, even if they had more. Its calculated by (AI Total Population * (Turns since you met the AI + 10) * iMaxGoldTradePercent)/100
Turns since meeting is found by (current turn number - number of turn you met), so on the turn you actually meet them this is zero.
iMaxGoldTradePercent is an AI personality number found in the XML, each AI has either a 5 or 10 value for it and if theres any interest i'll table them up and post it.
The final max gold the AI is willing to trade is the lowest result of these two formuae rounded down to the nearest 10.
Unfortunately screenshotting isn't working for some reason so heres an example thats easy to calculate, and easy to test in WB.
If at turn 0 you have met an AI with a iMaxGoldTradePercent of 10 (Churchill is one example) and 2000 gold with a 1 pop city then we get
Gold available to trade = (Pop (1) * (turns known (0) + 10) * MaxGoldTradePercent (10))/100
Simplified 1*10*10/100 = 1
Which when rounded down leaves 0....
We can also find the population required to allow him to trade 50 gold on turn 1
50 = Pop * 10 * 10/100
Pop = 50/1 = 50
Some of what this means
Contrary to popular belief the AI doesn't always keep a certain percentage of its gold stashed away, its in fact capable of trading its whole stockpile provided certain conditions are met (Ais total gold is below the cap and AIs GPT is 0 or higher).
The biggest factors affecting how much gold an AI is willing to give are the AIs population, and especially how long you've known them. The latter explaining why overseas AIs have so little available when you first meet them.
How much they like you has no impact at all.
iMaxGoldTrade cap can be found using the iMaxGoldTradePercent in XML and information that can be aquired ingame (Pop, turns since meeting) so it may be useful for HoF style games
When you give or trade gold to an AI if the gold available for trade doesn't go up by the amount you give, then they have hit their iMaxGoldTrade cap and you can use that to calculate their total population.
Some AIs have maximum trades twice as high as others (though available gold often stops this making as big an impact as it sounds)
If it isn't restricted by the iMaxGoldTrade cap, the AI will only hold back enough gold to cover their current defecit for 12 turns (36 turns on marathon). So if your seeing zeros on the diplo screen long after meeting them then the AIs gold is either 0, or between 0 and 12 * their defecit.
iResearchBuffer = -calculateGoldRate() * 12 can be used to your advantage provided the AI has a negative GPT. Reducing the defecit by 1GPT frees up 12 gold for trading on that turn (again 36 on marathon), and the GPT you 'give' doesn't even start to get payed till next turn
His deficit can't be more than say -4 gold per turn.
That means according to the 2nd cap, he will only offer us 700-(4*12)= 652
or 650 rounded to the nearest 10.
According to the 1st cap, we've known Mansa for 52 turns.
He has an iMaxTradeGoldPercent of 20.
If Mansa has 8 population, then he will trade us at most [8*(52+10)*20]/100 = 99.2 or 100 rounded to the nearest 10.
The game picks the lowest cap between the two caps when deciding how much gold to let an AI offer us from its stash.
So Mansa on T52 should be willing to trade 100 to us if his population is 8.
If his population is 9? 110 from his stash of 700
If his population is 10? 130
In theory, if we could gift Mansa enough cities so he'd had 51 population, he'd give us 650. We'd have to start gifting him gpt at that point and a few more pop to get the full 700 gold.
Now if the game has gone on 200 turns, and Mansa has 100 population, the main limiter becomes Mansa's deficit spending. With just pop and turns he'd be up for trading away up to 4200
The best way to erase a civs deficit spending midgame to get a huge chunk of gold available for trade is to gift them a tech and switch their civic/religion so they experience Anarchy.
(I don't think flipping them with a spy results in Anarchy? )
However, Mansa is spiritual, so this isn't an option.
The other is to set the slider to 0%, gift Mansa a Fish+200gpt to completely eliminate his gold defecit spending
Make a tech trade for his huge pile of gold, then pillage the fish to cancel the gpt subsidy
Or DOW to cancel the gpt subsidy
Being able to judge if an AI has a ton of gold that could be unlocked by an Anarchy bribe or gpt gift needs testing to figure out how best to determine going about it.
Sometimes a small gpt gift changes the slider to make the AI's deficit bigger.
Or canceling a resource for gpt trade changes the slider and increases the AI's deficit.
Can a spy determine how much gold a civ has in its treasury? That would be a big piece of the puzzle if it could!
**Edit**
It seems passive espionage can tell you exactly how much gold an AI has in its treasury!
Spoiler:
How to convert that number into gold? Really need LowtherCastle
He/she could effortlessly interpret passive espionage data
There does seem to be a limit to this idea besides the 2 caps that are used to limit what gold an AI will part with. I have seen this behavior while testing this idea and came up with it as an explanation: (not 100% sure it is accurate)
Every time you succeed in getting gold from an AI, it gets subtracted from the (pop*turns) cap. That means when we take the 100 from Mansa on T52, we have to wait for additional population growth from Mansa, or a bunch more game turns to pass before it will start inching up again.
This is the main reason that civs who have experienced huge population losses (vassals) never have any gold available to trade for essentially the rest of the game.
Also, some civs don't have iMaxTradeGoldPercent=20 like Mansa. The warmongers tend to have it equal 5.
These civs will only be able to part with 1/4th as much gold as Mansa usually does on average.
If you are looking for a monster payday from a civ, they need many things:
1) A huge pile of gold
2) No defecit spending
3) A huge population and you've known them a good number of turns (newly met civs on another continent will never have a huge payday waiting for you)
4) Haven't been relentlessly harvested 100 every few turns
In conclusion, no there is no way to get Mansa's 700 right now.
But midgame when he has way more population (thank you floodplains and Immortal difficulty!), it could certainly be possible if he still has it or built on it.
Contrary to popular belief the AI doesn't always keep a certain percentage of its gold stashed away, its in fact capable of trading its whole stockpile provided certain conditions are met
On t45, I talked to Genghis Khan and he bragged about an archer. On t46, he bragged about having built a chariot. Does this mean it was built turn 45 and can be moved turn 46? Or was it build turn 44 and was moved turn 45? I believe it means the former which means our warrior taking the worker on t46 is suicide, since the chariot must be in either city ready to move on the AI's part of t46.
I'm sorry for my sloppy micro that lead us to believe we could get a chariot T45.
However, I believe this means we are going to get 4 gold because we are going to exceed our max OF from the chariot with 2 chops.
This will make getting currency easier even though the capital will not grow. The capital growth error came about in part because the test game is not accurately representing the resource trades that we have in the real game I believe.
I unfortunately don't have much time, so I'm sorry I can't test what is best for the capital.
With a little math, we should be able to predict if getting currency is difficult if we don't whip the library to bring the capital to 2 pop. I hope we can do a bit of testing or math to see if whipping to 2 pop is better than waiting and whipping it 3 pop.
I think we should make an attempt to improve our communication and commit to looking at the fine details if we hope to avoid misunderstandings/mistakes in the future.
One way to improve our communication would be to repost the PPP each time we make edits to it rather than just edit the prior post. Reposting is simple, copying the entire post while in the edit window preserves all the formatting and it increases our post count to boot!
That way everyone is clear on what the plan is without having to go back to a prior post that they might not realized was edited.
I'm satisfied that the risk of taking the worker was too high. At the end of the game we can go back and use STW's save game to take the worker and we will see what would have happened (well with new random seed maybe it will take a few tries to see what would have likely happened)
Let's look forward and look for ways to communicate and play better in the future.
I believe the latest test game is not reasonably accurate. I made the missing Gold & Fish gift to Joao II on turn 44 which caused Capitalist unhealthiness in t44 & t45, causing the loss of 2 food. I should have made those gifts to Gandhi. I don't recall when those Gold & Fish gifts should have been made to Joao II.
I'm concerned that at one City, Gandhi may not survive to vote for us.
If the test game is accurately representing the real game (at least the access our empire has to resources) then I'd like to propose that we
make it a policy to play out the test game using our PPP so we avoid similar problems in the future.
I'm guilty of this since I didn't play out the test game when making my micro suggestion for Pigs and made a mistake with the chop timing.
bcool, you were right to point out that the mistake in Capitalist's unhealthiness was definitely a flaw in the test game. Playing the PPP on the test game would not have helped, because the flaw was in the test game itself as you had previously pointed out.
This site uses cookies to help personalise content, tailor your experience and to keep you logged in if you register.
By continuing to use this site, you are consenting to our use of cookies.
from his trade mission to Joao because he can't trade the gold away yet or spend a big deficit with only 2 cities.
)
