SGOTM 21 - Plastic Ducks

Doshin · Jan 24, 2015

@Yamps

Decided on a new post after all.

I've marked some changes:

Yamps said:
T30

Warrior 1N, worker 1 chop, worker 2 moves to the same spot (gems), scout stays put.
Sailing at 0% research.

T31

Warrior 1N, workers chop forest, scout stays put.
Sparta at size 2, working the cows and forest hill to get the wbt on the next turn.
Sailing 100%.

T32

Move wbt to crabs, warrior 1N, scout put, workers mine gems.
Start warrior in Sparta, working the lake and cows which doesn't matter since we also
revolt to Slavery.

T33

Warrior 1N, workers finish mining, scout to forest 5N from city spot 3, wbt nets crabs.
Sparta has to choose between cows and gems at size 2. ~~Per Doshin's input, Sparta works cows instead of gems for one turn only.~~

Work Gems + Crab on this turn.

T34

Warrior 1N, workers 2S1E, fortify scout.
~~Sparta switches to gems and works gems and crabs from now on.~~

Stick with Gems + Crabs.

T35

Worker 1 1SE, partial farm and cancel. Worker 2 1SE, 1NE (Doshin's micro). Warrior 1NW.

T36

Warrior 1NW, worker 1 to city spot 3. Worker 2 to gems north of city spot 3.
Athens 1 pop whips settler (mine gets whipped).
~~Sailing to 0%.~~

Sailing @80%

T37

Warrior 1NW, workers chop. Settler moves towards city spot 3.
Athens starts on wbt, working all improved tiles except the mined grassland hill.

The Athens build might change to a Warrior, depending on the preference of the team (as in my last post). This won't change much, our 30H will just shift to Warriors.

~~Sailing back to 100% research.~~

Sailing from 80% ---> 100%

T38

Sailing done.
Animal Husbandry @ 0%

Sparta: Warrior completes (15/15H).
Begin Lighthouse.
Change Gems to Cows.

Warrior 1NW, settler on spot, workers keep chopping.

T39

~~Sailing done.~~
Worker 1 chops for Athens, settler founds Corinth. Worker 2 chops for Corinth (galley). Exploring warrior fortifies for spawn busting duties.
Sparta finishes warrior (moves 1W), starts on LH at size 3.
Athens about to get exploring wbt on T40.

Pause for input.

I attach my save file, in case I've forgotten to list a change.

Doshin · Jan 24, 2015

By the way, I'm sorry that it took me so long to provide feedback, but I had no idea why your test file didn't allow us to build the Oracle. The result *should* be the same as my own.

Take a look at this:

City 2 @ size 3 (Doshin ---> Doshin)

Spoiler :

City 2 @ size 3 (Yamps ---> Doshin)

Spoiler :

WTH? Why on earth does the city in the second picture have 1F less per turn? This is not isolated to a single turn and compounds throughout the turns I tested.

It is a bug, right?

I attach the saves, in case it is just me. :dunno:

Duckweed · Jan 24, 2015

That is something wrong about the test file. I remembered that in SGOTM14, a wheat tile showed 1F less from time to time. Creating test file with BUFFY was very inconvenient.

Edit: I just downloaded the save and took a look, I saw the same as you. There must be something wrong about Yamps's save.

Doshin · Jan 24, 2015

Yamps said:
@Doshin

I'm having trouble reproducing your play. We're talking about more than 20 turns of micro after Sailing with some decision point on almost every turn. In the pre Sailing part it was enough to work one wrong tile for only one turn for you not being able to get the Oracle on T61. This is also true after Sailing. Could you post the automatic BUFFY log of your play?

Do you know how to access it? I can post it if you tell me.

In my current run I got to Writing one turn later. Probably because of Corinth (third city) working the forest with the intention of getting the galley earlier. I'll try again. What were your builds in that city?

The two Workers, after mining the first Gems, must move 1NE. Both chop the Gems (2/3).

On the next turn, the borders expand. Worker A chops (3/3). Worker B mines (1/4).

On the next turn, the Galley completes (50/50H).
Worker A moves into city 3. Load the Worker onto the Galley. Move the Galley towards the island.
Worker B chops (2/4).

On the next turn, start a Warrior.

I'll also note that we could have some trouble chopping the second forest in Cow city. Your test save is different than the one I'm using, with protecting mountains that are not there in the real game. Corinth could also have some trouble with the most northern gem.

Mountains were for lazy testing. In the real game, Sparta's Forest will need to be spawnbusted.

The Corinth Gem is 100% safe with your PPP.

EDIT: XPost

Please use the updated save when replaying, there could be subtle differences in the one you're using.

EDIT: okay, I'm getting T49 Writing with my save which is one key landmark. Some dilemmas: what does Sparta work with the 4th pop? Lake or forest grassland hill? For now, I'm working the grassland hill since it doesn't speed up Writing even more. Still, it's not clear what's the right way.

This took a long time for me to understand too. In general, at size 4, the Lake > Forest. I think that is consistent.

Size 5 presents more options (riverside Grassland vs. Forest Grass Hill).

I would say that, on T54, you will need the following hammers invested:

Lighthouse: 29/60H
Library: 26/60H
Warrior: 14/15H.

On this turn, the pasture completes. Build a Warrior (14H + 7H = 21/15H).
On T55, whip a Lighthouse. Regrow to size 4 is instant.
On T56, put all OF (= 40H) into the Oracle.[/b][/color]

I also have a problem of getting the fur settler from the capital too early. It's ready to finish on T46 and I need to 1 pop whip it to Library on T49. I'll put something else in the queue and see how it plays out. Probably another work boat for the fur spot.

As soon as the Gems are improved, work this in City 3.
Athens should change the Grassland hill to the Wheat and build a Warrior.

On the next turn, Athens should change the unimproved Furs to the Grassland Hill.

On the next turn, keep the same set up.

Athens should then grow and the Warrior will complete with 1H of OF.

There's also that idea Kossin mentioned about chopping the fur forest to consider. I'm also boarding one worker on the galley immediately when it finishes, that should be okay.

Another edit...how are you getting so much production to build the library in one turn?

The Settler needs to be 1-pop whipped on T48. This will produce 114/100H. The OF and chop will amount to 90/90H.

Hmm...I'm getting the first work boat one turn earlier than you, T40 vs T41.

I get it on T40 also.

Did you apply Corinth city spot forest chop to settler instead of work boat? This is really convoluted business to reproduce.

I know. The glitch/bug in some Sparta tests doesn't help much either.

I've attached a number of benchmark saves, in case it's of any help.

Doshin · Jan 24, 2015

Duckweed said:
WB for me. There's also the possibility of easy horse. After Oracle, the bottleneck should be the newest city, in the worst case, we have room to suffer in mainland cities.

Edit: I saw that you have not produced the 5th settler in your last test, I'm not sure whether it's important.

What I did in my test was to chop a forest + delayed settler in capital so that capital will be size 5 when the lib is done (T50), the cost is that I have to chop the forest 1N of the Fur on the island to be able to improve the Fur, but gain 1 pop in capital.

The Settler is ready to whip on T61, so 1 turn behind the test you posted:

Spoiler :

Earlier, the capital is at size 5 when the Library completes and runs 1 Specialist:

Spoiler :

On the following turn it runs two Specialists and City 4 is settled:

Spoiler :

Note that there is no WB on the second crab. The Furs are constantly worked by the capital or the island, so a minimum of 15 turns commerce would be lost (45 :commerce:

, with +3CPT) if we don't improve the tile. 10H are also lost with the pre-Math chop.

I think this is too much to save 1 turn on a Settler, although if we reveal Horses with Animal Husbandry, that might change my mind.

Kaitzilla · Jan 24, 2015

Doshin said:
By the way, I'm sorry that it took me so long to provide feedback, but I had no idea why your test file didn't allow us to build the Oracle. The result *should* be the same as my own.

Take a look at this:

City 2 @ size 3 (Doshin ---> Doshin)

Spoiler :

+4F from Cow
+4F from Crabs
+2F from city center
+1 from Gems

-6 from population (size 3)

= +5F net total

City 2 @ size 3 (Yamps ---> Doshin)

Spoiler :

+4F from Cow
+4F from Crabs
+2F from city center
+1 from Gems

-6 from population (size 3)

= +4F net total?!?!?!????

WTH? Why on earth does the city in the second picture have 1F less per turn? This is not isolated to a single turn and compounds throughout the turns I tested.

It is a bug, right?

I attach the saves, in case it is just me.

I cannot explain this bug in Yamp's test save.
The farmed cows are definitely giving +3 :food:

instead of the correct +4. :hmm:

I like the choice of 1 Workboat to explore instead of 2 Warriors.

I'd also like it to explore Northeast past the fish, not South and West towards the icy tundra.
We are near the bottom of the world a bit east of the map's center (55,8), so I'd like us exploring north sooner rather than later.

Spoiler :

The Buffy Autolog can be found in Windows 7 at least by clicking the START button in the bottom left, then clicking on Computer/My Computer -> C : Drive -> Program Files (x86) -> Firaxis Games -> Sid Meier's Civilization 4 -> Beyond the Sword -> Mods -> BUFFY-3.19.003 -> Test Game SGOTM21 (Text Document).

For steam the path will be different, but once you locate the Civ4 files click on BTS->Mods->Buffy->Name of the test game you are using.

Then copy the info and paste it into a forum post to show us a game log.

If you play the same game over and over it will continue the log each time you play, so go to the bottom of the log and only record the info for the game session you want to post about.

Keep in mind that Buffy only keeps a log if you turn on Logging by BUFFY using the in game controls ALT+CNTRL+O and making sure the file path is correct.

If you do everything right, you get a nice game log that looks like this:

Spoiler :

I'm sorry I haven't been contributing much lately.
I'll be leaving town soon and won't be back until Feb. 5th, so my posting will be limited.

Don't forget we have over 3 months left to play, so there's no huge rush in getting the next 30 turns done.

Yamps · Jan 25, 2015

Wow, this is so crazy! I've spent a couple of hours yesterday trying to get it work and there's some odd bug involved. I picked the starting T30 test save from here, does it also cause trouble for anybody else? What are your starting T30 saves, I'll try those when I get back home.

shakabrade · Jan 25, 2015

Crazy bug indeed.

Finally, I finished a huge project and got some time to familiarize with the game.
I have been trying to skip Maths bulb and got Currency from Oracle 2T later than Dosh and Duckweed.

I have also done some testing regarding Oracle and GLH dates (only 5 maps so large variance is expected) until 1000BC. It seems the most frequent Oracle date is around 1360BC (2 times), and then 1880BC (2 times), and one time it was 1560BC. GLH dates are 1520BC and 1360BC, and other times it wasn't built. Mids were never built, even with stone.

My attempt loses Oracle 3 times in respect to my test games, and Dosh's and Duckweed's lose 2 times. Also, I remember from my experience that Oracle is often built around 1560BC so those 2T gained from Maths are critical. Also, other techs, like TW and Pottery will be gotten faster with Maths bulb and I find it wothwhile.

Duckweed · Jan 25, 2015

Doshin said:
Note that there is no WB on the second crab. The Furs are constantly worked by the capital or the island, so a minimum of 15 turns commerce would be lost (45 , with +3CPT) if we don't improve the tile. 10H are also lost with the pre-Math chop.

I think this is too much to save 1 turn on a Settler, although if we reveal Horses with Animal Husbandry, that might change my mind.

Forgot that the 2nd crab was delayed if sending the WB to scout, we lose quite a few food doing this.

Now I prefer to do 3 whips in Cow city, we lost a few turns of yield on a resource tile before the 1st whipping anger wears off or the fur is connected, but gain from the capital and fur city.

Yamps said:
Wow, this is so crazy! I've spent a couple of hours yesterday trying to get it work and there's some odd bug involved. I picked the starting T30 test save from here, does it also cause trouble for anybody else? What are your starting T30 saves, I'll try those when I get back home.

I also ran my test from Doshin's T30 save here, no problem for me.

Kaitzilla · Jan 25, 2015

Yamps said:
Wow, this is so crazy! I've spent a couple of hours yesterday trying to get it work and there's some odd bug involved. I picked the starting T30 test save from here, does it also cause trouble for anybody else? What are your starting T30 saves, I'll try those when I get back home.

I'm pretty sure we're all using Doshin's Turn 30 Test Game save.
I've had no troubles with it so far. :dunno:

I looked at your save again.
It looks like 4 :food:

on the cows, but once you open world builder it shows 3 :food:

.

Spoiler :

Try deleting the farm improvement and adding it back again on turn 30, that should fix it.
To do so, press Escape, click Enter World Builder, then the Edit Map button, then the improvements tab, then farm, then right click to delete the old farm and left click to place the new one.

Duckweed · Jan 25, 2015

shakabrade said:
I have also done some testing regarding Oracle and GLH dates (only 5 maps so large variance is expected) until 1000BC. It seems the most frequent Oracle date is around 1360BC (2 times), and then 1880BC (2 times), and one time it was 1560BC. GLH dates are 1520BC and 1360BC, and other times it wasn't built. Mids were never built, even with stone.

My attempt loses Oracle 3 times in respect to my test games, and Dosh's and Duckweed's lose 2 times. Also, I remember from my experience that Oracle is often built around 1560BC so those 2T gained from Maths are critical. Also, other techs, like TW and Pottery will be gotten faster with Maths bulb and I find it wothwhile.

5 tests are a bit less. especially if run tests from the same map. Moreover, random tests rather offers a rough idea than a special date of Oracle could be gone. Our test game in SGOTM14 always failed the Oracle because a AI was industrial AI having Marble. As I replied to Yamps before, it depends on whether there are AIs who like PH having Marble, if so, then no matter how hard we try, we are going to fail, otherwise, 90% we are going to accomplish.

Yamps · Jan 25, 2015

I've had another quick run with the original T30 save and there was no glitch this time. No idea why or how that happened. :dunno:

I don't have much time left today for civ, but I'll make another test run till Sailing just to make sure it all works okay and post it here. After that, there will also be a lot of work since there are many possible pitfalls till T61 at every step.

kossin · Jan 25, 2015

It's a known bug to me at least; it happens with WB edits and such. Destroy the farm and remake it using WB'd workers. The first time we saw it was SGOTM14.

shakabrade · Jan 25, 2015

Duckweed said:
5 tests are a bit less. especially if run tests from the same map. Moreover, random tests rather offers a rough idea than a special date of Oracle could be gone. Our test game in SGOTM14 always failed the Oracle because a AI was industrial AI having Marble. As I replied to Yamps before, it depends on whether there are AIs who like PH having Marble, if so, then no matter how hard we try, we are going to fail, otherwise, 90% we are going to accomplish.

If everyone did 5 tests we'd have very good idea about the dates.

My method was lacking only in too small number of tests. Settings were: known AIs + random AIs, we are Pericles, map is always random with high sea level.

What I have noticed in my civ experience is the fact that wonders go BIAFAL in narrow bands of turns (I know for sure for GLH since it interested me the most that it goes BIAFAL around 1920BC, 1520BC, 1320BC). It is very low chance that it'll go BIAFAL outside of the band. That is logical since techs take some time to be teched, time which is roughly the same every time. Gems, gold, marble can make wonders go outside of the bands, but that is rare since not all AIs will go for wonders even with those resources.

From my experience, Oracle dates (bands) I remembered were exactly the ones I got from testing, and I stopped testing when my hunch got confirmed (and also went to sleep

).

I also know about 2800BC Oracle from Izzy and 2400BC Oracle from Izzy and Pacal.
If we lose Oracle, it'll be best to lose it like that.

I hope you have understood what I wanted to say.

Yamps · Jan 25, 2015

Ok, everything should be fine now. The path to Sailing is clear and I think we're good to go by plan corrections in Doshin's post #241.

Here's also the BUFFY log of my last test. We should use these more because it is often very hard to reproduce tests where the slightest micro detail matters. There's also a nice feature ALT+E for player comments at key moments that automatically get logged with other data.

I'll leave it as it is and play tomorrow if the team agrees.

Spoiler :

Duckweed · Jan 25, 2015

shakabrade said:
If everyone did 5 tests we'd have very good idea about the dates.

My method was lacking only in too small number of tests. Settings were: known AIs + random AIs, we are Pericles, map is always random with high sea level.

What I have noticed in my civ experience is the fact that wonders go BIAFAL in narrow bands of turns (I know for sure for GLH since it interested me the most that it goes BIAFAL around 1920BC, 1520BC, 1320BC). It is very low chance that it'll go BIAFAL outside of the band. That is logical since techs take some time to be teched, time which is roughly the same every time. Gems, gold, marble can make wonders go outside of the bands, but that is rare since not all AIs will go for wonders even with those resources.

From my experience, Oracle dates (bands) I remembered were exactly the ones I got from testing, and I stopped testing when my hunch got confirmed (and also went to sleep ).

I also know about 2800BC Oracle from Izzy and 2400BC Oracle from Izzy and Pacal.
If we lose Oracle, it'll be best to lose it like that.

I hope you have understood what I wanted to say.

If you look at the tests that kossin did in SGOTM11, you'll see that the wonder dates are totally random, not clustered. That result is reasonable.

The wonder date is affected by

1. The time when AI learns the tech, which is map dependent.
2. The food(able to whip) and production(general high production resource tiles) of their capital, which is also map dependent.
3. Other AIs competition of earlier wonders like SH and TGW since AIs tend to pick a cheaper wonder. This is also the reason why stone and marble are the major factor that affect the wonder date. This determinative factor is not random in SGOTMs, but is controlled in designer's hand. That's the real reason why I said we have 50% chance of getting Oracle. ~T60 Oracle is a safe date if designers don't intentionally prevent it.

The key thought of going for a wonder is that the gain is bigger enough than the loss if failure. Moreover, ourselves should try to reduce the loss, in which we try to make the chop and whip in the last minute.

shakabrade · Jan 25, 2015

Duckweed said:
If you look at the tests that kossin did in SGOTM11, you'll see that the wonder dates are totally random, not clustered. That result is reasonable.

The wonder date is affected by

1. The time when AI learns the tech, which is map dependent.
2. The food(able to whip) and production(general high production resource tiles) of their capital, which is also map dependent.
3. Other AIs competition of earlier wonders like SH and TGW since AIs tend to pick a cheaper wonder. This is also the reason why stone and marble are the major factor that affect the wonder date. This determinative factor is not random in SGOTMs, but is controlled in designer's hand. That's the real reason why I said we have 50% chance of getting Oracle. ~T60 Oracle is a safe date if designers don't intentionally prevent it.

The key thought of going for a wonder is that the gain is bigger enough than the loss if failure. Moreover, ourselves should try to reduce the loss, in which we try to make the chop and whip in the last minute.

I didn't follow civfanatics back then, but I guess kossin did it right, with statistically relevant specimen unlike me.
Since I don't rely on wonders in my games, I didn't think I was biased to get an impression. It is possible that my wonder dates got mixed with AI's. :dunno:

But, also, I do think that bands of elevated probability must exist. They could be lower than I thought, though and, therefore, next to irrelevant.

Oracle was disabled a SGOTM or two or three ago (you couldn't have beaten an AI) and people weren't happy since it reduced the diversity of the game. I don't think Oracle was made mission impossible on this map.

Pardon me for not having an accurate memory to give you the exact reference but someone must remember that.

Doshin · Jan 25, 2015

Duckweed said:
WB for me. There's also the possibility of easy horse. After Oracle, the bottleneck should be the newest city, in the worst case, we have room to suffer in mainland cities.

Kaitzilla said:
I like the choice of 1 Workboat to explore instead of 2 Warriors.
I'd also like it to explore Northeast past the fish, not South and West towards the icy tundra.
We are near the bottom of the world a bit east of the map's center (55,8), so I'd like us exploring north sooner rather than later.

Two for exploration > early spawnbusting, which sounds fine to me.

Horses would be a big help, but, we also have enough commerce and time to go to Archery after Priesthood if barbarians are being difficult.

Kaitzilla said:
The Buffy Autolog can be found in Windows 7 at least by clicking the START button in the bottom left, then clicking on Computer/My Computer -> C : Drive -> Program Files (x86) -> Firaxis Games -> Sid Meier's Civilization 4 -> Beyond the Sword -> Mods -> BUFFY-3.19.003 -> Test Game SGOTM21 (Text Document).

For steam the path will be different, but once you locate the Civ4 files click on BTS->Mods->Buffy->Name of the test game you are using.

Then copy the info and paste it into a forum post to show us a game log.

Thanks.

Will track it down later.

Duckweed said:
Forgot that the 2nd crab was delayed if sending the WB to scout, we lose quite a few food doing this.

Now I prefer to do 3 whips in Cow city, we lost a few turns of yield on a resource tile before the 1st whipping anger wears off or the fur is connected, but gain from the capital and fur city.

There are 4 turns in which the Clams can no longer be worked, due to the capital producing an exploring WB. The difference is –10F +6H +2C, which seems significant, but an additional whip in Sparta costs 1-population point (–24F) and causes the city to drop to size 2, costing us one of the Crabs/Cows/Gems for 3–6 turns (depending on the tiles worked). The city also cannot whip at size 3 without losing another resource tile or suffering bad unhappiness. We'd lose 18 turns of Workboat exploration.

I would prefer not to whip Sparta too hard because this will be our main Worker/Settler/unit pump after the Oracle is built.

Yamps said:
Ok, everything should be fine now. The path to Sailing is clear and I think we're good to go by plan corrections in Doshin's post #241.

For exploration: move the Scout 1NE and back each turn. This grants visibility across the water, and we can might see a barb or AI unit.

If Shaka's Scout moves in the right direction, we might be able to explore a little sooner or shadow.

Otherwise, green light from me.

Duckweed · Jan 25, 2015

Doshin said:
There are 4 turns in which the Clams can no longer be worked, due to the capital producing an exploring WB. The difference is 10F +6H +2C, which seems significant, but an additional whip in Sparta costs 1-population point (24F) and causes the city to drop to size 2, costing us one of the Crabs/Cows/Gems for 36 turns (depending on the tiles worked). The city also cannot whip at size 3 without losing another resource tile or suffering bad unhappiness. We'd lose 18 turns of Workboat exploration.

I would prefer not to whip Sparta too hard because this will be our main Worker/Settler/unit pump after the Oracle is built.

Sparta's 1 pop can't be counted as loss since they are used to translated into hammers, the real loss is from 1 resource tile. The loss from WB exploration is indeed a problem. Fair points for staying with your last tests. :goodjob:

@Yamps

Good luck then, either a gold from the hill or a forest grows in Cow city would be very welcome.

Yamps · Jan 27, 2015

Sorry for delays, I was too exhausted yesterday. I'll play it out later today (~8h).

SGOTM 21 - Plastic Ducks

jolly yellow giant

Attachments

jolly yellow giant

Attachments

Deity

jolly yellow giant

Attachments

jolly yellow giant

Lord Croissant

Deity

Praise Vivec!

Deity

Lord Croissant

Deity

Deity

Deity

Praise Vivec!

Deity

Attachments

Deity

Praise Vivec!

jolly yellow giant

Deity

Deity

Similar threads