SGOTM 08 - Fifth Element

[WastinTime]

Are we planning on milking wonders with a domination victory? That seems to be the case if we're planning on corporations.

If I have time I want to try a quick domination game where I let the AI build the wonders and I just take them, to see if it would yield a better score than a prolonged milked domination game.

Please do, and thanks in advance! It will give us a better picture.

BTW, what exactly are the map settings? sea level? etc.

 

[MacBalt]

Checking in.

Work is kicking my butt right now, so keeping up with this is going to be a bit of a challenge. I suggest we get a variety of test maps going to check out our strategy ideas under a range of conditions.

Oh yeah, and welcome to our new members!

 

[BLubmuz]

Test games can help, but without knowledge of at least a couple AI (who and where) and a decent portion of the map they will not help much.
A test game with some 20% of the map reproducing the actual one and the AI we'll know will surely help a lot.
What if we're on the small?
What if in the other continent there's a bunch of Industrious AIs (maybe with stone, marble and everything)?
What if Astro is needed for a domination?

I'll try a tech/wonders count, to try to figure where the break is (or are).
This will be much theorical, but it can give some idea.

 

[Sweetacshon]

BTW, what exactly are the map settings? sea level? etc.

All we know is:
# Rivals - 7 AI civs
# World size - Standard
# Difficulty - Monarch
# Landform - Big and Small
# Environment - Temperate climate, medium sea level
# Game Speed - Normal
# Everything else - Default
(dude - it's in the OP!)

I've been pressing enter through a pretty much automated watchathon to see what the AI will do for us and analyse the w-t curve. Sure, it depends on the AI involved, etc, but I just want to get my head around it. I was going to post the results, but I'll wait 'til I finish.

 

[Sweetacshon]

Ahh, what the hell - here are some figures up to about turn 300. See that the graph flattens at around 200. Note that this is the AI building wonders only, and doesn't include NWs. I'll finish it off and put the rest in soon, and give some idea of how concentrated the wonders are.

edit: See post #29 for attachment.

 

[BLubmuz]

Great job, Sweeta
it's what i meant in my previous post.
If we can build/conquest al the listed wonders, and maybe add 2 corps to them before turn 300 we'll maximize our WS/T (wonders' score/turn).

I guess the laurels will be decided by the 4th decimal.

 

[culdeus]

Yeah without the other AI's known to us it will be hard. I'm sure Murky is looking through the python script to find out the willingness of AI to build wonders and what not.

ETA: looked at that curve. That's really dramatic. Should be something to work around. Might world build myself some mechanized infantry and hit enter a bunch of times myself.

So a target of a T180 dom/conquest and we'll be able to GE rush some of these post t200 wonders you have listed there. AP in t180 seems awfully late as does some of the others just by my memory of things. So this could be a worst case curve in terms of AI performance. Though it is Monarch so I'm not all that used to that, maybe it's normal for them to be this slow at Monarch.

Just a proposal for the captain. On the first page perhaps one of the posts should be edited to list something like this:

Current Turn: X
Wonder Points: X
Current Wonder/Turn :
Previous TS W/T:
n-1 TS W/T:
n-2 TS W/T:
And so on



Wonder Name : Tech Needed: Hammer Needed : Simulation Build : Actual Build : Who Build : When? : Captured When?

Oracle : Priest : 150 : t95 : ? : ? : ? : ?

Or perhaps there's a better way to handle spreadsheet data here or online in some way. This will be fairly math intensive.

ETA2: If we take that sheet as the bible there's just no way we do corporations. The break point comes in the Kremlin/Taj timeframe. Well before getting Corporations and the other advanced techs like railroad etc. needed for a corp. That is unless we can somehow continue the .1wpt rate prior to t200 shown in the graph beyond t200. There will be some serious spinning of wheels between Education and all the super advanced techs.

 

[WastinTime]

How many wonders do you see us getting by t200?

If you say 1/2, then we would score better if we get them all before t400.
If you say 2/3, then we would score better if we get them all before t300.

I'm glad you're doing the analysis, but it just feels like the answer will most likely be ALL wonders by t250. That's about the same # turns we used on the space race in SG7 and we'll need less tech this time.

 

[Sweetacshon]

OK, I finished it off... I'm not sure why, considering the trend was obvious, but I suppose for completeness' sake.

Notes:

• I didn't build any wonders, so we'd still have to try to build some or capture most to get close to that curve. At first this concept seems easy, but if we beating on AI, they won't build wonders, and if we eliminate AI, the world may tech slower, so it'll be a trade-off.
• National Wonders weren't included, so we'll have to work out how to fit them in efficiently.
• The curve is fairly flat in the 2nd half, so we have some leeway - but a thought hits me that it might be good to finish the game with a flourish of corporations if we can combine a win condition with a beeline to that tech, as well as save the required GMen in time.
• I didn't have chance to check the distribution of the wonders yet, but you'd imagine the AI continues to build wonders in their best couple of cities, and not spread them out too much. Knowing who is around and what resources they have will (again) be all important. Let's not neglect scouting in the early game.
• Were any wonders missing?

Too tired - bedtime. I might actually have a look at the real save tomorrow.

 

[culdeus]

[*]The curve is fairly flat in the 2nd half, so we have some leeway - but a thought hits me that it might be good to finish the game with a flourish of corporations if we can combine a win condition with a beeline to that tech, as well as save the required GMen in time.

[/LIST]

Too tired - bedtime. I might actually have a look at the real save tomorrow.

Keep in mind corporations take 2 techs. Corporation And some other one like railroad, rocketry, etc. , plus the resource. Each corporation has it's own tech associated with it no doubles. What's worse is the early corporations are single tech:wonder. Later ones like Radio, medicine, MM have multiple wonders per tech.

 

[BLubmuz]

Yes WT, we need to discuss this: of course a great tech lead grants us the chance to build the wonders oueselves and to go for cultural or anything we like.

Culdeus, no need to go in the SDK: all the leader infos are in this file: CIV4LeaderHeadInfos.xml

I'll draw a list in order of wonders-orinted.

 

[Mesix]

I really think this game is going to break my calculator's 5 key. I don't think we'll really have a good feel for the VC till we get well into the game. The builds will almost certainly center on military rushes and then transition to a wonder build-a-thon. There's certainly a chance we won't need/want to research corporations.

It all depends on the math.

I think that Wasting Time's idea to hold onto great people to found corporations is a good idea. With a sound research strategy we should be able to found a lot of them. Most of the corporations are in the mid game portion and can be completed in the relatively early. There is no need to research into the future era, and Mass Media seems to be the furtherest that we would tech for wonders.

We still have to consider the timing of our finish. It may be infeasible to get the least few wonders as the research and build time may actual drag down our score. We need the most wonders in the fastest way possible.

 

[Mesix]

How many wonders do you see us getting by t200?

If you say 1/2, then we would score better if we get them all before t400.
If you say 2/3, then we would score better if we get them all before t300.

I'm glad you're doing the analysis, but it just feels like the answer will most likely be ALL wonders by t250. That's about the same # turns we used on the space race in SG7 and we'll need less tech this time.

According to the game setup, there are a possible 310 wonder points. That is 62 wonders. If we can get all 62 wonders by turn 310 (to make the math simple) that would be a score of 1.0. To achieve the same score we would need 20 wonders by turn 100 or 40 wonders by turn 200. Assuming that we are shooting for a score of 1.0 or above, we need to have one or more wonders for every five turns played.

Allowing the AI to build some of the wonders and capture them seems to be necessary. We need to decide which early wonders to build ourself and which ones to let the AI build for us. As you mentioned earlier, letting the AI build SH and us build the Oracle for the slingshot (and GP points) seems like a good idea. Pyramids are not rushed by the AI very often and make for a good GE city (not to mention the research bonus from running Representation early with a SE strategy).

We also need to decide what VC to shoot for. Early wins can be had through a Religious victory. Domination and Culture seem to be appropriate VC for a mid game fiish. Domination seems best if we are going to let the AI build wonders for us and Culture if we are going to self build them.

 

[Mesix]

The starting save is up and available to play. Who is going to play the critical first set of turns?

 

[Sweetacshon]

How about you, Mesix? It's a 5th Element tradition that the new teammates go first. I suppose the first task is city placement, which will require moving the scout, taking a screenshot to post here, and stopping.

 

[BLubmuz]

How about you, Mesix? It's a 5th Element tradition that the new teammates go first. I suppose the first task is city placement, which will require moving the scout, taking a screenshot to post here, and stopping.

Agreed!
So, Mesix volunteers to play the first TS.

Remember: after move the scout, take a SS, save and exit, then when you'll play the entire TS start from that save. We need to do this to accomplish to the GotM/SGotM rules.

Now
we have to decide where to move the scout: gold-hill, hill 2S of Settler, or forest 1SW of Settler?
I'm in favour of the Gold hill.

First research fishing, then ?
First build warrior, then WB.

but the above are my ideas, we need to discuss them.

 

[dutchfire]

Semi-lurker checking in

Welcome new team members!

Scout to gold hill, I'm fairly sure we'll settle in place either way, so it's best to get our scout out and exploring.

I think getting all/most wonders is the best idea. How long do you estimate it takes to get all necessary techs?

 

[BLubmuz]

Another idea.
Since we won nothing but a wooden spoon with this tradition, and despite a wonderful Deity game we still won nothing, what about let Sweeta start?

OK, i'm not superstitious, just knocking wood (or touch iron, to use the italian saying).

Please remember to take a look on the first page from time to time, i update the TS played fairly regularly, and it works also as a roster for the first round of TSs.

Can be OK 12 turns each? reduced to 9 in case of war?

Let me know, please.

 

[Mesix]

Here is a screen shot after moving the scout to the gold/hill tile.

Spoiler :

 

[Mesix]

Another idea.
Since we won nothing but a wooden spoon with this tradition, and despite a wonderful Deity game we still won nothing, what about let Sweeta start?

OK, i'm not superstitious, just knocking wood (or touch iron, to use the italian saying).

Please remember to take a look on the first page from time to time, i update the TS played fairly regularly, and it works also as a roster for the first round of TSs.

Can be OK 12 turns each? reduced to 9 in case of war?

Let me know, please.

Sorry...too late. I moved the scout and took the screen shot as requested. I'll wait for a consensus on the initial strategy before I start playing the turns.

How many turns?