SGOTM 07 - The Real Ms. Beyond

[chmoosi]

Yeah, why not explore? I can't see it having any negative effects. A few points:

-Didn't someone mention that building a warrior didn't contribute much at all to power, so building one would be a (at least relativily) a waste of hammers because we can use them to barracks - wb - archer - archer line which is more solid imo (if ultra-solidness is desired depending onthe neighbours the order can be switched to bar- archer-wb-archer). Actually, reading back. What's the decay effect? I'm not aware of such thing.

-Despite having those extra food resources, would it be useful to build a worker earlier? I don't see any use of the extra food until we get Monarchy (otoh we get there pretty soon and it would be nice to have all that food available right away)

-Not researching BW is imo a viable option as we can trade it (and masonry and sailing) when we get alphabet. On BW I'm not so sure as I'm not a slavery expert, but at least if there are enough other CIVs around we can certainly trade the alphabet for all those beginning techs and save many turns of research (as is the plan I presume, but at least in one game I played when I had researched BW, only about 50% of the other civs didn't have alpha when I got it, so its trade value was diminished).

-If we wanna chop it probably should be those hills as those tiles produce +1 commerce when chopped and wait at least until IW (which we can trade when we get alpha early enough)

 

[TheArchduke]

Hmm, a nice start though the surrounding lands are not that hospitable.

In my opinion a 2nd warrior exploring is always nice, but seeing how we are pressed against the sea and the tundra the only viable option to me would be exploring south.

So my vote goes for a barracks.

 

[Jabah]

Extra food :
1. Could whip every now and again
2. Get 2x Scientist (need writing very soon) for GPP (Academy, settled or bulb)

If Warriors count for 1/2 archer in the AI mind, they are not that not bad in cost/effect since they are only slightly more expensive than 1/2 archer. Plus they can be there very soon and will be exploring later (We don't want to send expensive archers exploring when we have OB do we ? ) -> so for me I still vote to build a warrior while researching fishing (if we still want that first).

Power : I think BW count for more than an archer when checking power, so if we skip BW (that seems not a good idea with the food and the suspect tile), we should build extra archers.

Toku : he didn't attack me in either test games (I had barrack+3archers early)

Jabah



Jabah

 

[Kodii]

The problem with being in school is that you return home to find an extra two pages of discussion which you have missed.

Anyways, I'm somewhat surprised to see that our starting position is even more buffed up than first imagined. I have no idea why the system thinks we should have moved from our starting position, but I am glad we settled in place.

(Aside: I thought we were following the 1984 theme for the demogame, not here! ... and I agree that Airstrip One is a more obvious (and perhaps less foreboding) name than the Malabar Front)

The discussion has not convinced me against the original route of Fishing and barracks. I think ruff is good to go to press Enter and then move the warrior once more. We can figure out the specifics of micromanaging after that. Work the grassland forest.

 

[Compromise]

Okay, it's a sweet site, but that's to be expected from an OCC SGOTM I suppose.

My biggest question is: Is there a date before which the AI is hardcoded never to attack the human? I think we want to be positive that we have a barracks and a couple troops before any such date.

I think we're okay to save the game *before* hitting Enter. Then, let's nail down the opening. I could see building a warrior timed so that it finishes on the turn that Fishing comes in with maximum overflow into a workboat. (Might knock a turn off the WB build time.)

In fact, tonight or early tomorrow I'll try to generate a save with the most aggressive of the aggressive AIs--suggestions welcome--nearby so that we can attempt to draw that super-early AI declaration. (In fact, for testing the early declaration time, we can probably just delete the initial worker, research Mysticism and the like, and build workers. Power ranking ~= Zero!)

Edit: My god, I thought I'd caught up, but I'd only read to the bottom of page 5, so I effectively crossposted with the last 4 posts.

@Ruff: Save it. We'll hash this out in the next 20 hours or so.

 

[ruff_hi]

I guess I sit on the current turn for a day or two. I'm going to wait for comp to hack up a suitable save and then run some tests. Actually - forget that - I'll see if I can generate my own. The options I want to look at are:

G1
Build: Warrior, Warrior, FBoat, Barracks, Archer
Tech: Fishing, Hunting, Archery
Work: Max hammers until crabs netted, then max food to size 2, then 3H tile plus netted crabs

G2
Build: Barracks, FBoat, Barracks, Archer
Tech: Fishing, Hunting, Archery
Work: Max hammers until crabs netted, then max food to size 2, then 3H tile plus netted crabs

G3
Build: Barracks, FBoat, Barracks, Archer
Tech: Fishing, Hunting, Archery
Work: 2f1h tile until FBoat available, then max hammers crabs netted, then max food to size 2, then 3H tile plus netted crabs

 

[Swiss Pauli]

My biggest question is: Is there a date before which the AI is hardcoded never to attack the human? I think we want to be positive that we have a barracks and a couple troops before any such date.

I don't know about this, but what I do know is that the AI designates roles for its units as they're built and that this role does not change. The roles are scout, defender, and (potential) attacker IIRC. As far as I know, the Vanilla AI will never designate Archers or Warriors (even Quechua) as attackers, which means that danger enters our horizon when AIs learn BW. For me, knowing where the copper is makes BW a priority; we could have plains river copper ourselves, and, if the worst comes, having slavery to whip might just save us if the AI comes underprepared.

 

[Swiss Pauli]

In fact, tonight or early tomorrow I'll try to generate a save with the most aggressive of the aggressive AIs--suggestions welcome--nearby so that we can attempt to draw that super-early AI declaration. (In fact, for testing the early declaration time, we can probably just delete the initial worker, research Mysticism and the like, and build workers. Power ranking ~= Zero!)

Take Alex and put him to our south, on the coast with seafood. He's the most aggressive competent AI in Vanilla (40% units, 20% war chance at Pleased, PHI for tech), but with HR as fave civic he can be befriended.

As it's an OCC, we can always remodel based on later info as it won't take long to run 20 turns (e.g. if we find we're blocked off by Toku, isolated start, etc.).

 

[Compromise]

@Swiss: Good suggestion. I'll put Alex near us when I generate a game tomorrow morning.

I wanted to address the question of "Is there an date before which the AI will not attack the human?"

I just finished a quick test to answer this. I used Test Game #2 and edited my LeadersInfo.xml file so that Cyrus' "declare at cautious" rating was 0, meaning (I think) that he'll almost certainly want to attack if he can at cautious.

Then, I deleted the starting warrior and built a worker (disbanded) and started Stonehenge while researching toward Monotheism. I think Masonry actually increased my Power score, but otherwise it was zero. Cyrus declared in 2640BC so that in 2600BC, I started the turn at war with him (turn 34).

Generalizing from the single example--thus subject to the massive caveats this implies--I think we have 32T of guaranteed peace.

Also, Cyrus attacked with one stack of 3 archers and a second stack of a single archer, thus proving that the vanilla AI does not need BW before attacking. :gulp:

 

[Jabah]

I guess I sit on the current turn for a day or two. I'm going to wait for comp to hack up a suitable save and then run some tests. Actually - forget that - I'll see if I can generate my own. The options I want to look at are:

G1
Build: Warrior, Warrior, FBoat, Barracks, Archer
Tech: Fishing, Hunting, Archery
Work: Max hammers until crabs netted, then max food to size 2, then 3H tile plus netted crabs

G2
Build: Barracks, FBoat, Barracks, Archer
Tech: Fishing, Hunting, Archery
Work: Max hammers until crabs netted, then max food to size 2, then 3H tile plus netted crabs

G3
Build: Barracks, FBoat, Barracks, Archer
Tech: Fishing, Hunting, Archery
Work: 2f1h tile until FBoat available, then max hammers crabs netted, then max food to size 2, then 3H tile plus netted crabs

I have corrected my XLS file (attached here) and test various strategies with our OLD start (need to add the special tile, and take into account worker).
IF we don't want to build a worker in the first 25t-30t, the best solution are:
1. "Test5" : max food until size 2 ->max prod until WB -> max food until size4
2. "Test6" : max food until size 2 ->semi-max prod until WBx2 -> max food until size4.
Depending on how much you value Commerce compare to Prod+Food, you could consider your G3 opening, but I think it is slightly less optimum.

--> besically I think the best opening is max food (until size2), even while building a WB.

Jabah

 

[Swiss Pauli]

Then, I deleted the starting warrior and built a worker (disbanded) and started Stonehenge while researching toward Monotheism. I think Masonry actually increased my Power score, but otherwise it was zero. Cyrus declared in 2640BC so that in 2600BC, I started the turn at war with him (turn 34).

Generalizing from the single example--thus subject to the massive caveats this implies--I think we have 32T of guaranteed peace.

Also, Cyrus attacked with one stack of 3 archers and a second stack of a single archer, thus proving that the vanilla AI does not need BW before attacking. :gulp:

Not good news about attack archers, though it's slightly flawed as we won't delete the initial warrior which may buy a turn or two. However, more importantly, we'll have our barracks by then which will give us a big power surge (I saved screenies of the power graph).

 

[Jabah]

Quick link to 'Power' explaination:

http://www.civfanatics.com/civ4/strategy/demographics.php

So basically in WU (Warrior Unit !) for the early things.
archer = axe = 2
Wall = 2
Barrack = 4
Sailing, hunting, mining & animal = 2
Wheel, Alpha = 4
Math, Archery = 6
BW = 8
IW, Horse =10

That is to explain the 'power' in powergraph or 'soldier' in Demo (or both?), but now I am not sure the AI is using this info to declare (since most of our power comes from tech).
But that could explain why I was (almost) never attacked with warrior since I had lots of early power from tech (mining, BW, hunting, archery...)

Jabah

 

[Swiss Pauli]

Good find, Jabah (Another reason to research BW, btw)

 

[Jabah]

Also a question regarding the Map.

I have trouble reading properly the map form the last SShot.
What exactely is the deer tile : it looks like a plain but deer should be (I thought) only on toundra, also is there a forrest on top or not ?
(And is it a 2f2h or a 2f1h, can't see clearly).

[if this is a 2/2, we should consider worker and hunting first ?]

 

[ruff_hi]

I have trouble reading properly the map form the last SShot.
What exactely is the deer tile : it looks like a plain but deer should be (I thought) only on toundra, also is there a forrest on top or not ?
(And is it a 2f2h or a 2f1h, can't see clearly).

Deer in South is plains forest deer ... 2F2H. Beaver is also forest (1F2H1C). Sheep is not, just plains (2F1H).

 

[ruff_hi]

Cyrus declared in 2640BC so that in 2600BC, I started the turn at war with him (turn 34).

I was declared on by Japan in your Test3. He discovered BW in 3280BC and declared on me in 2160BC (log doesn't give me the turns). I had 1 warrior (the original one). Gave myself 3 archers (WB) and survived.

 

[Swiss Pauli]

I was declared on by Japan in your Test3. He discovered BW in 3280BC and declared on me in 2160BC (log doesn't give me the turns). I had 1 warrior (the original one). Gave myself 3 archers (WB) and survived.

1 warrior and no barracks (I assume). We won't be trying this opening in the real game

@Jabah - It would be prudent to improve that deer sooner rather than later: 4/2/0 is a pretty good yield. Worker & Hunting would be quite risky as it would we would get our power boost later and therefore have more chance of being attacked.

My current thinking is:

Barracks-WB-Barracks-Archer-Archer- whip worker at size 4 - Archer/WB with overflow (neighbour dependent) - WB/Archer (whichever we didn't just train).

Fishing - Hunting - Archery - BW - AH (Writing - Alpha, or Wheel* - Pottery - Writing - Alpha).*If we're lucky and can trade a crab for a happy, we won't need to hook up the furs so much (though we may choose to).

My inclination is to be 'progressive-defensive': i.e. to target GLib then CS (progressive) but to make sure we're safe (barracks and 3 early archers). Fully defensive would be to tech through the Feudalism path, fully progressive would be to try go forward to CS with as little power as possible.

 

[Jabah]

Deer in South is plains forest deer ... 2F2H. Beaver is also forest (1F2H1C). Sheep is not, just plains (2F1H).

So worker first is an option to consider (it will take 11t : 5t before culture and 6 after [5x5 + 6x6]>60) since he will be able to 'work' lots of tiles that have a big gap between improved and not. (the only downside is that the carbs were giving lots of commerce).

 

[Jabah]

My current thinking is:

Barracks-WB-Barracks-Archer-Archer- whip worker at size 4 - Archer/WB with overflow (neighbour dependent) - WB/Archer (whichever we didn't just train).

Fishing - Hunting - Archery - BW - AH (Writing - Alpha, or Wheel* - Pottery - Writing - Alpha).*If we're lucky and can trade a crab for a happy, we won't need to hook up the furs so much (though we may choose to).

If we go that way, I would change BW to be 2nd instead of 4th. We will have archery (4th) anyway before finishing the barracks if we grow to size4 fast.

 

[ruff_hi]

I looked at this on the train again ... I've put together my own version of the production spreadsheet (see attached) because I found the other a little confusing regarding timing of turn production.

Anyway, based on this spreadsheet, I think we have the following options ...

First 18 turns - middle ground

Tech: Fishing (6), Hunting (6), Archery (8) <- later turns might be less due to 'known tech' discount
Build: Barracks (6), WB (8), Barracks (6)
Size: 1 (11), 2 (7)
Tiles: 2F1H (5), 2F2H (6), 2F1H+2F2H (1), 2F2H+3H (5), 2F2H+4F3C (1) <- this might be diff depending on tech rate - aim is to finish barracks and archery tech as same time

The observant reader will notice that I actually haven't worked the netted crabs (ok, maybe 1 turn). If that is the case - why tech fishing and build a workboat? As such, I think our best STR opening is ...

First 18 turns - highest STR index

Tech: Hunting (6), Archery (8), Fishing (6) <- later turns might be less due to 'known tech' discount
Build: Barracks (13), Archers (5)
Size: 1 (11), 2 (7)
Tiles: 2F1H (5), 2F2H (6), 2F2H+3H (7) <- this might be diff depending on tech rate - aim is to finish barracks and archery tech as same time

If we want to go a high commerce route, then fishing, workboat x 2 and then some defense while working the 4F3C tiles.