SGOTM 12 - One Short Straw

[mystyfly]

Wow, quite some good talk and analysis, nice spreadsheet, bruce

Here's another riddle, guys:
F9 shows max AI production at 3h and minimum food at 2f. Since city centers give at least 2f1h, what tile is the 2f AI working?!?!? It has no food and max 2h, right? Furthermore, we know he has at least one water tile. Would an AI prefer a 2h1c plains hill to a 1f1c coastal tile?

The 2f AI might be the one with 57yrs life expectancy, suffering from unhealthiness. AIs (counting the city govenor as well) nearly always work high food tiles when possible and not at -cap.

Interesting theories, bcool and LC. bcool's looks viable (and very evil) but I doubt that as the AIs shouldn't know eachother now and I doubt that many trades would take place just yet. I think that thing with EPP to keep track on research is a good idea. Erkon definitely didn't pick techs for the AIs at random.

I'd tentatively rule out any techs that unlock expensive wonders or projects.
The AI isn't clever enough not to start a 200-turn Space Elevator with its only city.

 

[LowtherCastle]

I'd tentatively rule out any techs that unlock expensive wonders or projects.
The AI isn't clever enough not to start a 200-turn Space Elevator with its only city.

Erkon definitely didn't pick techs for the AIs at random.

Man, you guys way overestimate Erkon's ability as a mapmaker, his general intelligence, and his manliness. Most likely he picked a bunch of 200-turn wonders totally at random, and when we get to them with our infantry, their capitals will still be guarded by 2 warriors.

Btw, Erkon can't read this, can he?

 

[LowtherCastle]

As long as Gandhi is ranked 4th or worse, we'll have a hidden +2 with him, so all we need is +1 FairTrade to get him to Cautious. We should be able to trade Gandhi for some early techs, no problem.

 

[LowtherCastle]

Espionage should go to Gandhi, so we don't overlap research.

How much does "Can See Research" cost? We'll be 4"espionage" behind Gandhi until he finally meets someone else. Will we manage to see his research in the next 50 turns?

Edit: Okay, I see that the base cost is 75 EPPs. IF Gandhi only does 4/turn, then we gain about 1EP per turn on the overall cost. Then after he meets someone, presumably we'll gain a bit faster. I think it's going to take a while to see his research though. Maybe we should see if we can decide who we're most likely to want to steal a tech from and put them on that guy.

 

[mystyfly]

Do you guys want to put all EPs on gandhi and thus lose the ability to see the other civ's graphs?

 

[bcool]

So I was looking for a way that I could contribute... In general I haven't analyzed games in the level of detail that you guys clearly have. I'm trying to catch up as quickly as I can. I apologize for my rather poor reading of previous posts. For example, Mitchum clearly indicated he had already filtered the techs that contributed soldiers out.

Anyways I thought we might be making a test save that more accurately reflects the AI's starting situations. It will be difficult since there are so many unknowns, but I did figure out how to add the Manhattan project to the starting save.

ProjectType=PROJECT_MANHATTAN_PROJECT

is added to the worldbuilder file in the team info section.


What else would we have to do to create a more accurate test game?

-guess the extra techs each AI is given
---no tech (except maybe Stalin's) which give soldiers
---probably no tech (like rocketry) that gives the AI the option to build expensive projects (like the Apollo Project) or expensive wonders
---maybe something creative that might allow AI's to more quickly get to utilizing the manhattan project earlier than we might expect
---maybe techs to reward us for focusing a bit more on espionage

-figure out / guess how much fallout surrounds the AI's starting positions
--probably doesn't help much but the event log suggests we launched ~42 tactical nukes (if all the fallout was created by just these nukes then there are ~200 fallout tiles out of 1080)

-based on what I've read, we can even guess how many AI have coastal capitals -- could guess which ones and modify test map to reflect this.

-guess which continent various AI are on
-guess about the islands Erkon mentioned

-guess about the 230k soldiers Stalin has

Do you guys usually make multiple test games to test out different scenarios?

It would probably be useful to play out a game to close to the end again just to see what the AI do without Ecology. (I'm surprised they don't know it already)

I would imagine the game is going to play very differently, I would guess the other AI will be significantly crippled compared to Stalin and ourselves. Or at the very least confined to a smaller than normal area.

Would Erkon try to protect the AI and us from Stalin early on by putting Stalin on an island? It would take him quite a while to get astronomy presumably.

 

[babybluepants]

One note: randomly generated maps with these settings (including my original test) have circa 1250 land tiles. The actual game only has 1080. That probably confirms the many-islands bit.

 

[babybluepants]

Gandhi: Alphabet
De Gaulle: 1 Renaissance tech
Roosevelt: 1 Renaissance tech
Churchill: 1 Industrial tech
Mao: 1 Industrial tech
Stalin: 1 Modern tech
....
Any guesses on which techs each AI received?

Two random ideas:
1) Techs that give freebies (Lib & Econ in Renai, Comm & Phys in Ind..).
2) Fave civic techs (Nationalism, Communism, Constitution) - this one wouldn't entirely work, unless they were given further score some other way. It could also be confirmed very soon, if they actually switched.

Techs on our path that could be usefully stolen is a good guess. Unfortunately, I don't see us generating more than the base 4EP for a long time. GW is impossible here.

 

[ZPV]

How much does "Can See Research" cost? We'll be 4"espionage" behind Gandhi until he finally meets someone else. Will we manage to see his research in the next 50 turns?

Edit: Okay, I see that the base cost is 75 EPPs. IF Gandhi only does 4/turn, then we gain about 1EP per turn on the overall cost. Then after he meets someone, presumably we'll gain a bit faster. I think it's going to take a while to see his research though. Maybe we should see if we can decide who we're most likely to want to steal a tech from and put them on that guy.

It's 75 + 0.75*his espionage against us.
so it'll be ~75 turns, if he doesn't meet anyone.

 

[shyuhe]

Finally found some time to open the save. This is going to be one crazy game I don't do PBEM so I'm not as good as you guys at dissecting the start save, so I'll try to do some start testing instead.

Also, everyone please make sure your autosave interval did not get reset to 4 when you installed the new mod.

 

[bcool]

Ok new approach to guessing the techs
what techs would make the game significantly different than an average game -- assuming these SGOTM games force us to use units/strategies not usually seen. (like paratroopers in SGOTM9, and the requirement of nuking all our opponents in SGOTM10, and the strange requirements of SGOTM11)
And maybe there is some in game logic to the techs the various civilizations have

Industrial techs
Physics (reveals uranium, enables airships -- also seems linked to the nuclear aspect of the game and you can see some people surviving in airships)
Rare to have a game where an AI has air unit before the player.

and Electricity (allows bomb shelters -- seems like something someone would retain if the world was nuked back to the stone age and I don't think I've ever seen bomb shelter built)

Stalin
Superconductors (allows research institute)
--would be interesting to let a russian have this building early on

Renaissance techs are harder to guess
Roosevelt
Corporation? --obsoletes the great lighthouse which might make not getting the GLH less of a blow if we can get this a bit earlier by trade or espionage? Lets AI do a bit better even though they don't know ecology.
Economics? (similar to corporation, also follows banking -- which unlocks roosevelt's favorite civic so he might be motivated to tech the prerequisite)

De Gaulle
Printing Press?
Constitution? -- might let him do a bit better with representation (he would still be motivated to get its prerequisite of Nationalism (allowing his favorite civic)

Nationhood would allow the Taj Mahal (so that is probably not likely, since the AI has issues with expensive wonders / projects)

Having one person with Alphabet from the beginning is interesting (I wonder if he had to do this to give a us a chance to make peace before getting killed)

 

[ZPV]

I've continued playing around with settling on the plains-coast, starting AH-Fishing-Writing this time.
If we're really lucky, we could trade AH for peace; else Writing.
For a 2-city CS slingshot, and no trades, it's a little slower than settling in place, but if we can pick up BW from Gandhi, it's quite a bit faster, since we get the Academy quickly.

 

[LowtherCastle]

I've continued playing around with settling on the plains-coast, starting AH-Fishing-Writing this time.
If we're really lucky, we could trade AH for peace; else Writing.
For a 2-city CS slingshot, and no trades, it's a little slower than settling in place, but if we can pick up BW from Gandhi, it's quite a bit faster, since we get the Academy quickly.

What are your first builds? When do you need BW from G to make it useful enough?

 

[LowtherCastle]

New test save. Didn't try to do anything particluarly special, but Gandhi has Alpha, the others have something other than Ecology, except I gave Stalin Eco. SO all except Stalin have clean resources in their fat cross. Made most of the coastal too.

ZPV: Hang on a minute. How many warriors should I suicide to equalize the 1 nuke:14 warrios ratio?

EDIT: Never mind, I just nuked everyone and suicided 14 warriors...d'oh!

 

[ZPV]

ZPV: Hang on a minute. How many warriors should I suicide to equalize the 1 nuke:14 warrios ratio?

What do you mean? The nuke gives us 10 war success, the 14 warrior suicides gives them 42 war success each - they don't naturally subtract from one another. If you want to try, 11 warriors each would be close.

I'm thinking about trying to work out how much tile culture there has to be to give us the 97 war weariness.

 

[ZPV]

What are your first builds? When do you need BW from G to make it useful enough?

I started worker-WB-worker-warrior-Library. It does cost us a bit in terms of exploration, but we can get an exploring warrior out well before our second city is up (currently planned for 3S in my mind so it can share a gold and cow).

When would BW be useful? Anytime before the slingshot. The limit is the production in St. Pete. If we can get it by t50ish, we can chop the forest 3S before settling there as well. I'll do another run, paying a bit more attention, with your new save, to give a better estimate.

edit: on this last trial, with the latest test save, I got peace with AH, traded for BW on t62, and got the CS slingshot in 1080BC (t73). There's still a little room for optimization, though - if I'd gifted Fishing the turn I got it, I might have been able to trade earlier and speed things up a bit more. Of course, it depends on Gandhi meeting another AI. If he ends up next to Stalin it will be easier to trade with him, since we need not be his worst enemy at Annoyed.

 

[LowtherCastle]

Is that the one I updated with all the right stuff (it shows 0 Views)? Because I'm having trouble getting Gandhi to trade for peace in the first place. If you have the same trouble I have, Erkon has given us a serious mess by giving Gandhi Alpha.

 

[ZPV]

I think it was one of the old ones. I'll try this one.

 

[Mitchum]

I started worker-WB-worker-warrior-Library. It does cost us a bit in terms of exploration, but we can get an exploring warrior out well before our second city is up (currently planned for 3S in my mind so it can share a gold and cow).

When would BW be useful? Anytime before the slingshot. The limit is the production in St. Pete. If we can get it by t50ish, we can chop the forest 3S before settling there as well. I'll do another run, paying a bit more attention, with your new save, to give a better estimate.

edit: on this last trial, with the latest test save, I got peace with AH, traded for BW on t62, and got the CS slingshot in 1080BC (t73). There's still a little room for optimization, though - if I'd gifted Fishing the turn I got it, I might have been able to trade earlier and speed things up a bit more. Of course, it depends on Gandhi meeting another AI. If he ends up next to Stalin it will be easier to trade with him, since we need not be his worst enemy at Annoyed.

FYI, it's a bit risky adding edits like this. I would have missed it had I not re-read the last 2 pages of posts on my computer. Safer would be to put more important edits like this in a new post.

In any event, a CS sling in 1080 BC sounds relatively safe. Plus, your plan has us settling in two locations where we can see everything (I belive LC's test had us getting a Pig resource that will most likely not be in our game). Can you upload a save?

 

[bcool]

Okay I editted LC's latest test save a bit

added the manhattan project
gave
Stalin Ecology
Churchill Electricity
Mao Physics
De Gaulle Constitution
Roosevelt Economics
Gandhi Alphabet

sacrificed 18 warriors to each AI to get 97 war weariness I saw in real game, but this gave some of the AI a great general (which I deleted). Not sure if Erkon removed the great general points somehow from the AI to avoid this.

Edit: gave stalin 1 ICBM and 6 Tactical Nukes to get the 230k soldiers. He does hit us with the ICBM the turn after we settle. Not sure if this really would be what Erkon would want since this would randomly kill some forests for some teams.