[BTS] BOTM 237: Bismarck the Bully, Emperor, starts 17th June

Joined
May 18, 2005
Messages
23,514
What was the mechanics about aggressive AIs...?
Something about easier worker stealing, but other than that I don't know. :D
Unfortunately I cannot find the post where this was explained in detail, I think by @ZPV, but I can’t remember for sure. In any case, the gist as I recall, hopefully correctly, is twofold:

1. It is harder to get the AI to cease-fire. (which I consider more problematic in terms of stealing workers)
Edit: See WT’s post below.

2. When declaring peace, you get three times as much stuff from the AI as normal.

It’s entirely possible my memory is failing me on number one. But not on number two.

There are other consequences of aggressive AI, of course, such as more likely to declare war and so forth but I never paid much attention to those details.
 
Last edited:

WastinTime

Deity
Supporter
Joined
Jan 27, 2006
Messages
17,607
Location
California
1. It is harder to get the AI to cease-fire.
exact opposite. It's super easy. It hinges on putting a military unit in one of the their BFCs. DeGaulle is famous for talking in 1 turn.
@LowtherCastle : you were always the go-to guy for knowing exactly how quickly they will talk, and how a loss in war negatively affects that, so it's odd you're rusty on this point.
Keep in mind, if their archer wanders into your BFC, that works against you when trying to get a cease fire.
I can't recall if it's just 2 tiles away (so diagonal 2 tiles works or if it actually has to be in the BFC.) Anyone know off hand?
 
Joined
May 18, 2005
Messages
23,514
exact opposite. It's super easy. It hinges on putting a military unit in one of the their BFCs. DeGaulle is famous for talking in 1 turn.
@LowtherCastle : you were always the go-to guy for knowing exactly how quickly they will talk, and how a loss in war negatively affects that, so it's odd you're rusty on this point.
Keep in mind, if their archer wanders into your BFC, that works against you when trying to get a cease fire.
I can't recall if it's just 2 tiles away (so diagonal 2 tiles works or if it actually has to be in the BFC.) Anyone know off hand?
Ahhhhhhhh yes, now I remember. Thank you so much.


But when the AI will talk is a separate matter to whether they will cease-fire. I have not forgotten that calculation. I have it in my spreadsheet and use it frequently.
 
Joined
May 18, 2005
Messages
23,514
When an AI will talk during war:

iRefuseToTalkWarThreshold = {5,6,8,10,12}
DeGaulle = 5...

Refuse Duration = iRefuseToTalk * {20+(80*AIWarSuccess*2/HumanWarSuccess}/100

From that equation, to get Degaulle to talk in just one turn our success versus his success has to be 8:0. (Or 32:3 and so on.)

The tricky part is calculating the respective war successes. If we attack a unit or worker and win we get +4. If we attack and lose he gets +3. If we attack two workers on the same tile we get +4+1=+5. So the simplest solution for dG is to attack workers on separate tiles, giving us +8. Capturing a city is worth 10, and if you killed a unit innthe process +4. When he attacks and wins he gets something but I forget what, probably 4. Similar with when he attacks and we win. I’m rusty on this part.

The actual war success is 1+(total war success calculation) to make it non-zero in the Equation.

If you’re stealing one worker then the equation works out like this:

Code:
RTT   Duration
 5         2t DeGaulle
 6         3t Asoka, Catherine, Fred, Gandhi, Hatty, Lincoln, Louis, Mansa, Pacal, Wang
 8         4t many
10         5t many
12         6t Sitting Bull

RTT = RefuseToTalk
 
Last edited:

WastinTime

Deity
Supporter
Joined
Jan 27, 2006
Messages
17,607
Location
California
Thx. Bookmarked. Stating the obvious... 'Aggressive AI' has no effect on 'turns to talk', only if they will cease fire and how much they'll give. And IIRC, threatening an additional city gives you even more payment for peace.
 
Joined
May 18, 2005
Messages
23,514
Thx. Bookmarked. Stating the obvious... 'Aggressive AI' has no effect on 'turns to talk', only if they will cease fire and how much they'll give. And IIRC, threatening an additional city gives you even more payment for peace.
In my spreadsheet I found this that I believe I cut and pasted from ZPV:

Code:
CF/DoP equation
half-ceasefire: We pay (100+(ourCities+theirCities)*3+ourpop+theirpop+ theirwarsuccess*20) *(Theirpower+10)/(OurPower+TheirPower+10)
 forDaggerAIs: *9*Theirpower/10*OurPower
 forAIsw/morelandthanus: *Theirland/Ourland*.9
 forWarPlanAIs: *35*TheirPower/10*(OurPower+4*TheirPower)
 forTotalWarAI: *2
eitherDogpileAI: *1.5
if theirendangeredcities>ourendangeredcities  /3 (AggAI only)
round down to nearest 10

ZPV somehow used this to calculate our value and the AI value and the difference in value determined what they were willing to give in a DOP. For aggressive AI the key is to have more of their endangered cities than ours to divide our pay value by 3.

If I’m reading this right, the aggressive AI divide by three factor is only significant if we have endured a fair amount of losses so that’s what makes the AI trade value better for us.
 
Last edited:
Joined
May 18, 2005
Messages
23,514
OK so I found these posts from Dan F. The first one gives a pretty good explanation of the situation. Of course he goes far beyond my abilities in terms of determining all that stuff but it doesn’t really matter because the basic calculation is somewhat clear and when you try to get a good peace deal either he is willing to do it or you power up some more or beat the crap out of him some more or a combination thereof.

Incidentally, siege weapons are especially valuable in terms of creating this Endwar value difference because you can greatly increase your chance of war success and decrease his chance for war success. Furthermore if you attack with a catapult and it withdraws, I’m pretty sure that does not give him any war success. Combining bombardment with red lining his units can drastically minimize his war success. War success is by far the largest factor in the equation.


It's a little bit confusing, both parties determine EndWarValues and they represent their benefit of not continuing the war.

I.e. if a weak player is under a lot of pressure and his cities are threatened plus his powerful opponent was able to accumulate a lot of war success, then this player will calculate a very large EndWarValue. On the other hand the strong unthreatened player will determine a very low EndWarValue.

During peace negotiations both teams put their determined EndWarValues in the balance -- but on the opposite side of the balance! Thus the weak player needs to add other stuff (techs, gold, cities, ...) on his side to achieve a 'fair' balanced trade.

Therefore our goal is to lower our EndWarValue and increase Toku's EndWarValue as much as possible so that he has space to add techs on his side for compensation.

If we endanger more cities than Toku, our EndWarValue gets reduced via the presented division by 3. We should further avoid any war success for him. Even though Toku clearly has the higher power this division and our better war success so far will lead to a difference of the EndWarValues that should allow us to get Archery from him.

In addition to this reduction of our EndWarValue, there is also the dubious doubling of Toku's EndWarValue I mentioned for the case that Toku started a total war. AFAIK we can't tell whether that is true or whether he chose a limited war, but if it is true we might even get about 3/4 of Math from him (rough estimate!) -> we need ~100:science: invested in Math, less if Hammu gets Math too (e.g. in a trade from Toku for IW).

Further knowledge about war stuff and peace deal bargains from more thorough investigations:

oh well, it appears easy Math is currently out of reach. Although I can now confirm that Toku does wage a total war against us (it was impossible for him to plot a limited war), unfortunately the mentioned doubling of the EndWarValue is likewise implemented for us as well. :(

Furthermore I made a slight mistake when mapping the displayed latitude (thanks to Stonehenge) to the in-game x-coordinate of Toku's capital. This x-coordinate affects the calculation of the players' AI_STRATEGY. So now that he has Alpha (entered classical era) and is currently using AREAAI_OFFENSIVE, the dreaded AI_STRATEGY_DAGGER is also possible for him.

This is bad for us as it enforces a second scaling of Toku's EndWarValue by the power ratio, decreasing it to our disadvantage and thus in turn reducing the difference of EndWarValues.
We could prevent this second scaling by climbing over the hard coded power threshold, so that the check Toku's power >= 120% our power ? fails (equals power ratio of 0.833) ... which is probably too much effort right now (current ratio 0.580).

At least it should still be possible to get Archery from him as long as we put the 2 cities in danger (very important!).

Or let's wait a bit and see whether our lucky streak continues and he loses more units so that we gain more war success while we build more units and maybe manage to jump over the threshold at one moment. Then use this lucky moment to negotiate peace.

Wow, pretty interesting!

Everybody please don't get panicked by Toku! :)

At the moment he is threatening SheepTown while we don't endanger any of his cities -> it's no wonder that peace negotiations don't look good and he demands SheepTown.

But we can turn this around next turn by putting 1 Axe on our Sheep and move another via the new Eiffel-road to the tile 2N2W of Tokyo -- if we agree that we really want peace.
Toku is amazingly strong (power ratio 0.527 :eek:) so in consequence of the second scaling with respect to the powers due to the darn AI_STRATEGY_DAGGER even Archery is out of reach now (difference of EndWarValues ~ 50 -> 1/3 Archery).

But we won't have to pay anything for peace!

Oooh Mama! :D ;)
Very nice, with 106:science: invested in Math we need a difference of EndWarValues of 470 and would currently have ~440 (incl. hypothetical Tokyo threat) thanks to superior war success. :goodjob:
We will get Math next turn in a peace deal once we endanger Tokyo, we probably don't even need a 100% slider T87, but better make sure for the case that Toku whips more units and/or finishes Construction.

Yes. I believe it's best to get peace T88 for the rest of Math. Peace deals will be a lot better if we can pass the power ratio threshold of 0.833 but I would redommend that for the next war (Toku will likely start plotting again soon ...).

Edit: tried some test calculations -- current power ratio is 0.7260 (=106/146) -> we could whip an Axe in SheepTown, use the overflow in CowTown to build another Axe there so that the regularly built Axe from DeerGold would bring us to power = 123 on T88.

If Toku doesn't increase his power in the IBT we'd pass the threshold (0.8424) bumping the difference of EndWarValues from ~440 up to ~800 to easily get Alpha from him (needs just 700 so there'd even be space for 1/2 Archery).

Should we try? :dunno:

We also have the option to upgrade Cook (Warrior = 2000 Soldiers -> Axe = 6000 Soldiers) in case Toku builds another unit... :think:
IIRC DanF5771 calculated how we could get free Alphabet for peace from Toku in the multi-gawa SGotM.
 
Last edited:

Fippy

Mycro Junkie Queen
Joined
Mar 17, 2013
Messages
12,690
Other than that, they will plot war constantly ;)
From AI survivor where they always played with Agg AI: I cannot believe Roosevelt walked all across the map to declare on Lizzy, two very peaceful AIs.
There are no peaceful AIs anymore with this option on.
 

Dhoomstriker

Girlie Builder
Joined
Aug 12, 2006
Messages
13,418
get your chosen starting save
Never trust a template maker.
Yes, where are the Challenger and Adventurer saved game "choices"? :shifty:


Unless there were map-maker shenanigans of renaming another Leader to the name of Bismarck, we should be able to use the following template data from a previous game as a reference:
Bismarck is Expansive and Industrious, and you start with Hunting and Mining.

The Expansive trait gives +2 health/city and +25% hammers for worker production and double production speed of Granary and Harbor

The Industrious trait gives +50% wonder production and double production speed of Forge

Unique unit: Panzer (replaces Tank)
The Panzer benefits from: 50% combat bonus against armor units. This compares with the Tank which has: no directly comparable benefits.

Unique Building: Assembly Plant (replaces Factory)
Not a huge benefit over the factory... the assembly plant lets you assign 4 citizens as engineers (only 2 for the factory).


With all of this talk of getting techs for peace, has anyone tested whether a tech which an AI received in trade is eligible to be given to us via war? Is a peace treaty a way around of the no-tech-brokering rule?

Actually, it is also worth checking the opposite... if we're losing in a war and we manage to get a tech in trade from another AI, can we "broker" that tech for a peace treaty? Given that there may be Aggressive AIs possibly declaring war when THEY feel ready for war, this situation may actually come up.


How about if you have a Vassal? Can you gift a brokered tech to your Vassal? More importantly, can the AIs do so?
 
Joined
Sep 3, 2002
Messages
500
Location
Dor-lómin
Spoiler threads to share our bitterness?
 
Top Bottom