The ultimate challenge deity part II, Liz

[Rusten]

I played a little into the first start (cows/silver) earlier today and.... (~500 BC)

Spoiler :

I'm fine (although barbarian galleys are truly annoying). The start didn't feel that difficult for me, I'm in a very good position.

- - - - - - - - - - - - - - - - -

I've never seen a cat rush on deity, Rusten was considering it in his first aztec deity game.He had a few things going for him of which i remember:

-He had altars
-Pacal was the intended victim.

The plan was out when Pacal built the soz.

There was more to it than that.

Spoiler :

Why I believe this is an exception to a usual approach:

- Aggressive trait helps (easy cover promotion and free combat I for axes, longbows are less of a problem).
- Jaguars (I can heal my troops quickly even before I get a GG medic saving a lot of time)
- Commerce (I have 2 sources of gold, a pretty decent river and a good capital so I will be able to research the required techs quite quickly which means an earlier war and an earlier recovery of my economy.
- Sacrificial Altar (Will be able to whip out an army faster and suffer less from unhappy citizens (can afford war weariness))
- My neighbours (especially Pacal) are peaceful and/or wonderhogs and bound to have much less units than a regular deity AI.
- I have a "second UU" (ivory)
- Early marble, forested and production-heavy capital and early commerce (means I'll get the GL/NE up quickly so there's a big chance I can keep some tech trading going on with GSs while warring)

I decided against it when Pacal made the SoZ and because Peter lightbulbed philosophy meaning I wouldn't be able to make any trades (war with Pacal and Peter had philosophy already). I also found relations going to friendly quickly a likely (due to how the religion spread and fav. civics) so tech trading was more beneficial.

 

[Dirk1302]

I'd be interested to see a save or a short report of the cow/silver start while i found it playable it really didn't seem easy.

spoiler about the new start until ~800 bc

Spoiler :

But this one's probably even more difficult. Alone on an island with Gil. I succeeded in keeping power at 0.3 until ~1000 bc by building a barrack an archer and some warriors, i teched to alpha to give it to him right away. Then i teched to monarchy to get HR going and pfff i've got him to pleased while my power has dropped to 0.2. Might be he has intended to declare already otherwise i get to live a few days. If i succeed in bulbing Tao i'll be totally be save. At least with currency i get 2 good trade routes but overall this scenario is worse than isolated unless i'm able to take him out later.

 

[Dirk1302]

@Oyzar until 600 BC

Spoiler :

oh well....

Maybe building barracks in 2 cities, connecting copper immediately by placing a city in a bad position and churning out some axes, not building a third city was the way to go. Not to rush him as that'll be to difficult against a protective ai but for power.I prefer the cow, silver start after all. I'll give the others some time to see how they are doing this time but i feel another re-roll coming up. Maybe i'll ask silver to put up an immortal student start for a change so we can kick some ass. Immortal is just so much easier than deity.

 

[Rusten]

Quick report up until 25 BC. I didn't mean that the map is easy, but I wouldn’t rate it as one of the worst maps I’ve played.

Spoiler :

Settle in place which reveals deer. Hunting -> AH is great now.

I come across an injured warrior on turn 25. I have not been in battle with it so this means I’m not alone—I decide to get the GLH (we have multiple forests to chop).

Settle my second city.

This was pretty painful the coming turns. I hadn’t been able to explore due to several barbarians so I missed out on the much better spot to the south where I’d claim gold. At least I’ll pick up copper early, but that gold mine would've been great for happiness and commerce.

I meet Frederick 1760 BC (wb).
1600 BC – Frederick’s GA has ended. Wow, that’s really early for a GA—Fred’s pretty stupid.

Pyramids went 1400 BC but I didn’t want it anyway. I succeed in my plans of grabbing the GLH with some chops. Unfortunately a barbarian city is blocking trade routes….

My work boat searching for Frederick (finally) finds something worthwhile.

I didn’t get any foreign trade routes up until now but it was definitely worth the wait. Frederick is overseas giving me 3 commerce for each trade route.

Suleiman is going to be dangerous this game—he’s almost settled 25% of the land but the worst part is that his land is really good as well, he has lots of rivers and flood plains it seems.

I settled my GM (GLH) as London needs food and it’s very early in the game still making the other alternatives sub-par.

Capture my first barbarian city 550 BC.

I make it in time for a northern one too. I don’t like the location but I have no choice but to take it. I also unlock the HE (10 XP swordsman), that’s going to be very useful.

I usually opt for currency at this point, but I decide to grab compass instead for harbors and because it’s a safer trading chip. Hopefully I’ll be able to trade for currency instead (still just Brennus having it 25 BC though).

Frederick declares on Louis 175 BC and completes the GL. I guess it’s a good thing he made it rather than the Hindu block—that would speed up liberalism a lot. As things stand right now I’m quite safe diplomatically (Fred is worst enemy) and I’ve started running the favourite civics of my Hindu allies—I should be able to bump them to friendly early on.

Brennus has founded Taoism—that’s a rare sight.

Settle a city for the gold I missed earlier 150 BC. It’s not going to be able to work the gold most of the time given the lack of food but the city is still worth it as it’s coastal (GLH), has some grassland and because it has forests to speed up the most important infrastructure.

Currently getting CS and harbors. London has 39 commerce but that’s going to increase with the harbour (ETA 1), cottage growth and the wine tile so bureaucracy is going to do a lot for me here science-wise (I have an academy up and running). Unfortunately barbarian galleys are swarming me--I need MC for coastal safety. 2 AIs have it already so I’m hoping to trade for that (+currency) asap.

Suleiman is going to be troublesome, but having the HE unlocked in addition to redcoats helps a lot. I'll put a coastal city south of London picking up the incense, stone and grassland.

I can't find my 25 BC save atm so the 475 BC one will have to suffice.

 

[Dirk1302]

@Rusten,well played

Spoiler :

You didn't play it so much different from me, i built one city more where sulei's Diyarbakir is in your game. It's a very good one but it hurt on maintainance and i'll lose the cow there one day. I also think you played a bit tighter than i did resulting in slighty higher research 25 bc. (i had 128 with slider at 100% with a similar decifit 100 bc)

The problem also in my game is Sulei because there's not much wrong with my position otherwise. I thought the over expansion might cost me lib but thinking about it philo will help to get some scientists together and Sulei doesn't even have col atm. It's not like he's Darius as in the previous game.

So we have 2 games here now, maybe i transport one of them to a new thread, copying the relevant posts, alternatively we can put above the spoiler which game we're talking about.

 

[SnowlyWhite]

@Rusten

Spoiler :

how weren't you beat to the sheep site?

I did york 1se, but I expected to lose the sheep site to sully, and he beat me by a mile; otherwise, yeah, I also started the glh - or more, the lighthouse, where I stopped. It's good that freddie is oversea, but still there's no water... so post pp things will probably be rough.

anyway, I'll keep going with 1st start for the time being... though I agree with snaaty

 

[SnowlyWhite]

old game, up to -560 bc where it ends

Spoiler :

thinking about it, I decide to settle in the south asap, while chaning research to agriculture

I previously sent a wb from london to york, but borders didn't pop yet

york switches from lh to 2 archers which I'll need to fogbost for barb. galleys

galleys already there... man, who decided this change in 3.17...

I miss glh -1330 bc gone; was sloopy on it anyway, not that the capitol helped in any way.

having to build 2 galleys obviously kinda screws my econ... I needed a lib., not stupid galleys...

obviously, 1st 1 dies 1st combat... definitelly nothing goes on this game...

I become judaist(learnt a long time ago to rather fear sully); at least, atm, there's religios diversity

610 bc freddie adopts bureau... ok...

560 bc game over

I'm glad I decided to oil up sully only to get kicked by brennus...

endgame land

 

[SnowlyWhite]

new game up to -2460bc where it ends

not too much land revealed

Spoiler :

getting my ass beat, I bravely start the new challenge with... phants to the east... I love phants

worker while techin' agriculture... how surprising; then planning hunting(how original) for archery(2 gems, rice, phants - I don't dare hoping for bronze too... normally I would've go for bw and hope, but here I feel I'd ask too much from the rng )

. .. .. .. .

and close... and with a close to nil peace rating... bad, bad, bad

after warrior, I go 2nd warrior while sending the 1st to fogbust; I improve phants after rice and 1 gem - need production now! If I'm to block.

gilgamesh is friggin' fast...

and... game over

-2460 bc

unfortunatelly, my 2 archers were out to fogbust, not that the vulture ain't op anyway...

 

[Dirk1302]

@Myself, SnowlyWhite,Oyzar and all,

I think we're wasting time on the gems start, this one really seems totally unplayable to me unless you get lucky.

I'll also put a link here to an article from Niklas detailing ai war decisions.
It's from the SGOTM 07 thread (a deity occ game).This Niklas has a very scientific approach to the game, there's a wealth of valuable info by him and others in this thread.

Don't open this article if you're currently playing SGOTM 07 as it gives info about that game.

http://forums.civfanatics.com/showpost.php?p=6806299&postcount=391

The main conclusion in this article, our power rating in the early game doesn't mean a thing, declaration is only averted if our power rating is close to the rating of the ai or higher depending on an ai dependent modifier n XML.

I have to add that this article is about vanilla. I looked up the info in BTS (it's in cvteamai.cpp from the sdk for those interested), and compared it to the vanilla one, on first glance it hasn't changed much since but the function is very long and calls other functions that i'd have to read into. One thing has changed, there's a time modifier, as far as i can see this precludes war for the first ?? turns, this could be dependend on the level you're playing.

While the info in this article is probably correct (i've seen parts of it stated before), there's one thing that still puzzles me: if power rating isn't important up to a threshold where you won't get attacked i'd expect similar declarations on immortal and even emperor and this isn't the case as far as i've seen.

 

[Rusten]

That was an OCC so naturally they had less power than normal, especially seeing as they were competing for fastest victory. In normal games you can usually keep enough power for it to make a difference, but it's true that it doesn't matter whether you have 0.00005 or 0.005 (random numbers out of thin air) power of the opposing AI. You are above this threshold most of the time in casual games.

However, if you only guard your cities with warriors (like Snaaty tried here) you're going to be below it for sure. Sometimes you get lucky and survive long enough to get diplomacy going but other times you get that early DoW on you--it's a big risk to take.

@Rusten

Spoiler :

how weren't you beat to the sheep site?

I did york 1se, but I expected to lose the sheep site to sully, and he beat me by a mile; otherwise, yeah, I also started the glh - or more, the lighthouse, where I stopped. It's good that freddie is oversea, but still there's no water... so post pp things will probably be rough.

anyway, I'll keep going with 1st start for the time being... though I agree with snaaty

Spoiler :

It was a barbarian city--I took it after the southern one. Believe me, I want York south too, but you can't have everything your way--there are always some set-backs.

 

[Dirk1302]

Yes it's an occ game but afaik the article is based on the function ai_dowar in cpteamai.cpp. These treshold values (they're percentages) are very high as Niklas says, you can look them up in CIV4LeaderHeadInfos.xml. For Montezuma for instance iMaxWarNearbyPowerRatio = 130 meaning that,

- if you're near him
- and he has decided to go to war
- and his power multiplied by 1.3 > your power

His decision to declare on you stands.

The result is indeed very anti intuitive, it doesn't explain why there are so much more early declarations on deity.The ai_dowar function is rather complicated calling other functions and xml vars. I'd like to dive into that myself sometime but not today .

 

[SnowlyWhite]

yeah, knew that article...

well, usually when building troops I build them for defence, not for power graph; screw power graph... especially if marathon(x3 turns, even at a small random, means it'll happen and that's that).

problem is -2460 bc decl.(I knew, can't say it took me by surprise, but even if building 4 archers... 4 cities vs 1 while praying for the peace event... too little) and pre alpha declarations where you have nothing to give for peace, even if you beat his sod, because matchin' diety power pre rene is... yeah, right... On the other hand, you can last while hoping the peace event pops., but it sets you back so much all the fun is spoiled. And gilgamesh till hr is the worst neighbor, period.

@Rusten

Spoiler :

ah, then I'm happy I did a good fogbusting for him

 

[silverbullet]

Yes it's an occ game but afaik the article is based on the function ai_dowar in cpteamai.cpp. These treshold values (they're percentages) are very high as Niklas says, you can look them up in CIV4LeaderHeadInfos.xml. For Montezuma for instance iMaxWarNearbyPowerRatio = 130 meaning that,

- if you're near him
- and he has decided to go to war
- and his power multiplied by 1.3 > your power

His decision to declare on you stands.

The result is indeed very anti intuitive, it doesn't explain why there are so much more early declarations on deity.The ai_dowar function is rather complicated calling other functions and xml vars. I'd like to dive into that myself sometime but not today .

According to that it really doesn't matter whether you power ratio to the AI is 0.1 or 0.6. As long as you are not 1.3 times stronger (never the case on Deity early game) your power ratio will not help prevent a war declaration.
So according to this the decision to go to war is pretty random.
We should have seen more early declaration in this case on Monarch even.
If AIs like gilgamesh have roughly 1.6% chance per turn to consider either a war or a limited war, it means that they are 73% likely to declare war sometime in the first 80 turns. If power ratio of 0.5 is not what deters them, what is it then? Why do I not get declared as often on Monarch?

 

[silverbullet]

I have a theory why power rating matters. It's not based on code or anything, just a guess.
THE BARBARIAN NATION.
Even if the AI decides to go to war, they will choose the weakest nearby AI available. In low level games that is usually the barbarian nation.
On Deity however, barbs are so strong that the human player becomes the prefered target.
If that is true, the important threshold of power is your army compared to the nearby barbarians (and it also means you should not do a total fogbust and let some barbarian cities live).

Anyway, I would like to dig into the code sometime to figure this out. If power ration really doesn't matter and there are no other factors you can control, than all that remains is to restart Deity games until you are lucky not to get declared early.

 

[Dirk1302]

That's an interesting thought, somehow i don't quite believe it as i assume it's coded that every player is at war already with the Barb state. But i keep an eye out for this.

- Some other possibilities

1. Bts or even warlords changed this, bts ais are somewhat more prone to declare on every level compared to warlords, vanilla. They spam more units also so power sure seems to have an influence.

2. There might be a level param that wasn't interesting for Niklas because he played deity

3. Niklas might have overlooked something in the rather long function that deals with these things.

One thing is imo almost sure, either power difference or the level we're playing on must make a difference in the declaration probability.

 

[Snaaty]

@ Dirk, another REALLY challenging game

This time, I was doing better, having just built the internet and now I´m sure I win the space race

With only space for 5-7 cities, I played some weird mixture of OCC with helper cities and late game wonderwhoring (later I maged to grab 2 marginal island cities, one having only 2 landtiles and the other deadlocked in a culture fight from the turn I built it). I did surprisingly well, keeping up in tech all along the gamewith AI´s having the double ammount of cities and not doing a single trade up to 50 AD (when I finally researched Alpha, Gilga was friendly sind 800-700 BC or so

Up to 17something:

Spoiler :

Round 3, the first shock:
I find Gilga´s city only 6 tile away from my capital

If I dont find a way to bring him up to pleased up to arround 2000-1500 BC I´m dead.

I pop bronze from a hut... ...bronze is between Gilga and me... ...anothe objective: settle near the bronze bevor Gilga and build at least 1 axe to show him

I come up with a plan:
Thanks to 2 diamonds and lots of forest, I should be able to get the Oracle build BEFORE 2000 BC and get CoL (Confu).

My plan works, I get Confu 2040 BC and convert Gilga 3 turns later. In parallel I gift him 1 diamond and make a useless trade to get the trade bonus also.

Then I chop two blocking settlers (1 claiming the copper, the other claiming fish/cow) and start to lean back.

Gilga gets pleased around 1500 already

Beeline Monarchy to run Gilga´s favorite civic and get him to friendly before 500 BC.

Build the Konfu-Shrine in capital, settle my backlands (squeeze in 7 cities with lots of overlapping).

First to Phil around 100 AD (my only bulb...) and build the AnkorW in capital. Now I´m settling some GPs for the hammers and the money. DeGaule finds my around 200 AD (pretty early).

Get Lib around 800 AD (havent expected it, becaues DeGaule had edu befor I had it, pick Astro.

Manage to settle 2 rather crappy islands, but well, better then nothing.

Beeline Medicine, get Sushi.

Beeline Radio and co. get most of the modern happy wonder + christo

Beeline Computer, get internet

I havent really thought I could make it, but with an early priest and several religions and a friendly neighbour sharing my reli (could have bulbed Theo also, making it 3 relis), I decided against cultural and diplo, because that felt too easy... ...now I´m really surpised how strong my research is compared to my tiny empire with 12% land compared to 35% of the biggest AI

Was a quick game, played a little sloppy (allowed Gilga to switch religion, was forced out of Bureau, didnt check tech and ressource trades often enough), because I didnt think I would have a real chance, but now this will be my first completely peacefull deity

 

[Dirk1302]

@Snaaty, well done .

Spoiler :

I didn't think of chopping oracle. I keep it in mind for next time if a situation like this arises. It might well be the only thing that works here.

That's a lot of research and you even have some room to get more switching to representation. Was it possible btw to smash Gilg in the renaissance if you'd played for it? That was my plan at least if i could survive the first turns.

 

[Snaaty]

Now I have some time, so here are some pics (apparently I havent taken many...):

Spoiler :

DeG showing up 275 AD and offering a deal:
View attachment 186818

Me grabbing lib (was completely surprised by this)
View attachment 186819

Meheheheheh... ...worldwar...have to love it, when the superpowers (France + vassals against Holland + vassals) try to play rough (each having about 25 cities, including the vassals)
View attachment 186820

Overview over our tiny empire of 7 cities and the two islands (one city I placed on an Island holland already had settled to grab some suggar... ....The island became later Amerika...)
View attachment 186821

 

[Snaaty]

EDIT:

doublepost

 

[SnowlyWhite]

nicely done