SGOTM 02 - Memphis Blues

[ShannonCT]

A samurai with the free Combat 1 and 2 first strikes is a significant improvement over a maceman. It can certainly tip the scales. For example, a Combat 1 Samurai is 79% to win over a unupgraded maceman in the open. Versus a crossbow, our Sam is 42.3% to win rather than 24.9% for a maceman. The sooner we get Sams the better. The one catch is that they require iron, but we might need to hook up iron anyway if we cant get bronze.

After hunting-archery-mining-BW, I like going for writing via pottery with a detour to iron working if necessary. Besides the techs leading up to machinery and CS, monarchy would help Kyoto run a lot more scientists (with caste system) and construction will be pretty crucial for any middle ages warfare. Agriculture, animal husbandry, masonry, and sailing seem like "maybes". Meditation, monotheism, horseback riding, literature, and drama seem pretty useless (ok, drama is always useless). What about alphabet? This isn't going to be a very peaceful game but if we meet several civs early on, we might want to be the first to alphabet.

I'm fine with the play order. 10 turns would be a good time for us to stop and discuss. Kyoto will get its border expansion by then and being 10 tiles closer to Kyoto should reduce maintenance to 5 GPT.

 

[BSouder]

Does anyone have any objections to the playing order KingdomBrunel posted?

I am fine with the play order as long as I can assume I am tacked on to the end.

With the added twists of this game I would also suggest we just do 10 turn turnsets the first round with a minimum of 24 hours discussion between.

For short term research I agree wih Archery then Bronze as the first 2 beelines. After that we'll have more information to decide where to go next, although since conquest is the only valid winning option we do know where our focus will end up. My suspicion is that sailing type techs are also going to be key again also.

 

[Mad Professor]

Right - I think we have a settled playing order for the moment (which we will then change, swap, etc. as needed. It looks like this:

Mad Professor (up next)
Frederiksberg
Radiopill
KingdomBrunel
ShannonCT
BSouder

I will play ten turns to begin with, and post results for discussion. Do we want to play ten turns each through the first round, or should I then play a second ten turns before handing it on for everyone to play 20 turns each?

I'll start with hunting and go to Archery.

I'll move the settler west to see if the land allows us to settle closer to Kyoto and scouting for good city positions. If there's a place, or game situation that requires settling before the end the ten turns I'll stop right there and ask your opinions before actually settling.

 

[Mad Professor]

For medium/long term stragety I agree that samurais will be good, so CS/machinery should be a priority. And we'll need construction for cats, and if we have ivory and come accross a lot of horsemen in other civs, the elephants we get from construction will be very handy too.

I'd like alphabet in there somewhere fairly early if we can afford it, for the tech trading, but maybe we won't be able to afford it. Let's see and discuss that later. Let's get archery and BW and see where we're up to then and what's happening.

 

[radiopill]

I'd like alphabet in there somewhere fairly early if we can afford it, for the tech trading, but maybe we won't be able to afford it. Let's see and discuss that later. Let's get archery and BW and see where we're up to then and what's happening.

As we have to go on early war, and as we wouldn't have a lot of time to make a lot of science search, we absolutely need Alphabet for "trading" peace with our victims...

 

[Frederiksberg]

After hunting-archery-mining-BW, I like going for writing via pottery with a detour to iron working if necessary. Besides the techs leading up to machinery and CS, monarchy would help Kyoto run a lot more scientists (with caste system) and construction will be pretty crucial for any middle ages warfare. Agriculture, animal husbandry, masonry, and sailing seem like "maybes". Meditation, monotheism, horseback riding, literature, and drama seem pretty useless (ok, drama is always useless). What about alphabet? This isn't going to be a very peaceful game but if we meet several civs early on, we might want to be the first to alphabet.

I agree with the main ideas stated above. I think Writing is important because it will enable us to speed up research in Kyoto by building a library and hiring scientist for beakers and Great Scientist points. If copper is not available I also agree that IW takes priority over Writing. As I said in a previous post this way of developing Kyoto will give us some Great Scientists in Kyoto. First one to build an Academy and the second (or third) may be used to discover Astronomy if certain conditions are met. One of them could be that we don't know Meditation so this tech should be avoided at least for a while even if we can get it for free! Regarding Alphabet and Sailing I would say we need those unless conquest can be achieved early without crossing the ocean. Even if we don't have many friends to trade with techs can be extorted from enemies in exchange for peace.

Right - I think we have a settled playing order for the moment (which we will then change, swap, etc. as needed. It looks like this:

Mad Professor (up next)
Frederiksberg
Radiopill
KingdomBrunel
ShannonCT
BSouder

I will play ten turns to begin with, and post results for discussion. Do we want to play ten turns each through the first round, or should I then play a second ten turns before handing it on for everyone to play 20 turns each?

I'll start with hunting and go to Archery.

I'll move the settler west to see if the land allows us to settle closer to Kyoto and scouting for good city positions. If there's a place, or game situation that requires settling before the end the ten turns I'll stop right there and ask your opinions before actually settling.

Sounds good to me! We could probably let the settler move unescorted the first 6 turns. After that animals start to appear and it's probably required to wait for the warrior to catch up before moving further. Another - but riskier - way of advancing would be to use only a single step at a time so that there is always a spare move allowing us to either settle on the spot if we see a double move animal like panther or wolf or move out of reach if we see a single move animal like lion or bear.

How do we identify a good spot for settling? Of course we want to move as close as possible to Kyoto. Settling on a plains hill has the merit of an extra hammer in the city tile thus doubling speed of production of the first military unit. The city would also be more difficult to capture. A hill inside the square 3x3 cultural borders would allow for an easily defensible improvement since archers can get hills defense promotions. Gold or silver would be a very nice bonus and a river is also nice for extra commerce. Settling at the coast would make it easier to set up a defensíve perimeter around the city to defend improvements. Any improved sea ressources would also be easier to defend.

 

[ShannonCT]

I agree with the main ideas stated above. I think Writing is important because it will enable us to speed up research in Kyoto by building a library and hiring scientist for beakers and Great Scientist points. If copper is not available I also agree that IW takes priority over Writing. As I said in a previous post this way of developing Kyoto will give us some Great Scientists in Kyoto. First one to build an Academy and the second (or third) may be used to discover Astronomy if certain conditions are met. One of them could be that we don't know Meditation so this tech should be avoided at least for a while even if we can get it for free! Regarding Alphabet and Sailing I would say we need those unless conquest can be achieved early without crossing the ocean. Even if we don't have many friends to trade with techs can be extorted from enemies in exchange for peace.

As far as sailing goes, I'm operating under the assumption that we will have several rivals to conquer on the original continent and that we will probably need the samurais to do so. In that case, we will already have the most expensive prerequisites to astronomy (metal casting, machinery) and can start on the astronomy path while our samurai are avenging Tokugawa's brother. The "maybes" I can see for getting sailing earlier are 1) Kyoto is within reach of coast, 2) barbarian galleys are pillaging valuable seafood, 3) we want to connect two coastal cities on the continent and roads are too slow or vulnerable. If we just want to explore, work boats are cheaper.

Sounds good to me! We could probably let the settler move unescorted the first 6 turns. After that animals start to appear and it's probably required to wait for the warrior to catch up before moving further. Another - but riskier - way of advancing would be to use only a single step at a time so that there is always a spare move allowing us to either settle on the spot if we see a double move animal like panther or wolf or move out of reach if we see a single move animal like lion or bear.

I think the Professor should work it out so that the settler is joined by the warrior by the end of the 6th turn (if that is indeed the magic number). If the settler were going to move one step at a time, he might as well be joined by the warrior. Even the single steps are no guarantee that a wolf or panther that is out of his field of vision wont get him. We can combine the single steps with the warrior escort for greater protection though (i.e., the settler makes one move onto plains/grass/desert, then backs off if he sees something bad, then the warrior rejoins the settler).

About the magic number 6, I ran the setup 10 times with my settler and warrior wandering around and recorded the number of the turn where I saw the first animal. The numbers were: 13,9,7,13,8,8,9,12,11,11. Has anyone actually seen an animal after 6 turns? Does anyone know how to find it in the game code?

I ran the same test for the appearance of barbarians. I settled near the coast as early as possible and then waited for the barbs. Number of turns to barbarians were: 45,50,47,61,68,54,43,47,45,47.

How do we identify a good spot for settling? Of course we want to move as close as possible to Kyoto. Settling on a plains hill has the merit of an extra hammer in the city tile thus doubling speed of production of the first military unit. The city would also be more difficult to capture. A hill inside the square 3x3 cultural borders would allow for an easily defensible improvement since archers can get hills defense promotions. Gold or silver would be a very nice bonus and a river is also nice for extra commerce. Settling at the coast would make it easier to set up a defensíve perimeter around the city to defend improvements. Any improved sea ressources would also be easier to defend.

Yes, I'll take all of those, thank you!

 

[Frederiksberg]

As far as sailing goes, I'm operating under the assumption that we will have several rivals to conquer on the original continent and that we will probably need the samurais to do so. In that case, we will already have the most expensive prerequisites to astronomy (metal casting, machinery) and can start on the astronomy path while our samurai are avenging Tokugawa's brother. The "maybes" I can see for getting sailing earlier are 1) Kyoto is within reach of coast, 2) barbarian galleys are pillaging valuable seafood, 3) we want to connect two coastal cities on the continent and roads are too slow or vulnerable. If we just want to explore, work boats are cheaper.

I see your point. I think we can decide about sailing later. Light houses may come in handy, but it's hard to tell with the info we have now.

About the magic number 6, I ran the setup 10 times with my settler and warrior wandering around and recorded the number of the turn where I saw the first animal. The numbers were: 13,9,7,13,8,8,9,12,11,11. Has anyone actually seen an animal after 6 turns? Does anyone know how to find it in the game code?

I ran the same test for the appearance of barbarians. I settled near the coast as early as possible and then waited for the barbs. Number of turns to barbarians were: 45,50,47,61,68,54,43,47,45,47.

In my test game I entered the world builder after each turn and what I saw was that animals appeared after 6 turns simultaneously in many different places. What I didn't do was to create different test games so the number could of course be varying from game to game.

I think the Professor should work it out so that the settler is joined by the warrior by the end of the 6th turn (if that is indeed the magic number). If the settler were going to move one step at a time, he might as well be joined by the warrior. And even the single steps are no guarantee that a wolf or panther that is out of his field of vision wont get him. We can combine the single steps with the warrior escort for greater protection though (i.e., the settler makes one move onto plains/grass/desert, then backs off if he sees something bad, then the warrior rejoins the settler).

If we are willing to take a small risk we could rush the settler west for the first 6 turns and then place him on a hill where he could wait for the warrior. From the hill he should be able to see any approaching animal. Of course it would be annoying to be forced to settle the city but we would never loose the settler this way. The idea with the westward rush would be that we migth encounter a coastline that stops further progress west. In that case we might be able to find a spot for settling before 6 turns. Or we might find another civ blocking the west of our continent. As soon as we have found a spot that seems optimal on this continent I think we should found our 2nd city.

Regarding the number of turns I think 20 is fine but with a "breakpoint" in the first turnset so that we can discuss the placement of our 2nd city. Like Mad Professor already suggested.

 

[ShannonCT]

In my test game I entered the world builder after each turn and what I saw was that animals appeared after 6 turns simultaneously in many different places. What I didn't do was to create different test games so the number could of course be varying from game to game.

Ah, good idea. I wish I had thought of it. So animals do appear on turn 6. I guess my numbers still give an idea of when to expect to actually see a beast. Using your idea, I checked that barbs appear on turn 42. I think these numbers are in the code.

If we are willing to take a small risk we could rush the settler west for the first 6 turns and then place him on a hill where he could wait for the warrior. From the hill he should be able to see any approaching animal. Of course it would be annoying to be forced to settle the city but we would never loose the settler this way. The idea with the westward rush would be that we migth encounter a coastline that stops further progress west. In that case we might be able to find a spot for settling before 6 turns. Or we might find another civ blocking the west of our continent. As soon as we have found a spot that seems optimal on this continent I think we should found our 2nd city.

We would have to hope there was a hill to stand on at turn 6 though and even then it might force us to settle somewhere pretty bad. I would still vote for using the warrior escort after turn 6 and finding a prime spot. I'm not sure we want to settle as early as turn 6 even if we did find a good spot from the settler's westward rush. Every turn we wait is 5-6 more beakers, which means getting archers earlier, being able to pop rush in Kyoto earlier, getting a library and scientists in Kyoto earlier, etc...

I was working out the specifics of the distance penalty in the worldbuilder. Every step to the north and west has an effect on the distance penalty. The distance to the capital seems to be calculated as the north/south distance + the east/west distance, rather than the straight line pythagorean distance. Our settler is starting at a total distance of around 42 tiles (6 south and 36 east). One step to the NW would reduce the distance by 2, while one step to the SW would have 0 effect (I'm not suggesting that the settler should never move south). I started playing around with cities at different distances from the capital at different populations (yes, maintenance does increase with population). Here's the results:

Maintenance of 2nd city as a function of distance to capital and population of 2nd city:

Distance......Pop1.....Pop2......Pop3.....Pop4......Pop5
27..............-5.........-6.........-7.........-7.........-8
28..............-5.........-6.........-7.........-8.........-8
29..............-5.........-7.........-7.........-8.........-9
30..............-5.........-7.........-8.........-8.........-9
31..............-5.........-7.........-8.........-8.........-9
32..............-5.........-7.........-8.........-8.........-9
33..............-6.........-7.........-8.........-9.........-10
34..............-6.........-7.........-8.........-9.........-10
35..............-6.........-8.........-9.........-9.........-10
36..............-6.........-8.........-9.........-10........-10

So for the most part, when population is factored in, every step closer to Kyoto will reduce maintenance. It's not as if crossing some magic line will reduce maintenance a lot. But the difference between being 35 tiles away and 32 tiles away is 1 beaker per turn for every turn until we move our palace.

 

[radiopill]

If we are willing to take a small risk we could rush the settler west for the first 6 turns and then place him on a hill where he could wait for the warrior. From the hill he should be able to see any approaching animal. Of course it would be annoying to be forced to settle the city but we would never loose the settler this way. The idea with the westward rush would be that we migth encounter a coastline that stops further progress west. In that case we might be able to find a spot for settling before 6 turns. Or we might find another civ blocking the west of our continent. As soon as we have found a spot that seems optimal on this continent I think we should found our 2nd city.

You pointed out a really interessant fact, what is the behavior of an Aggressive AI's archer in front of a lonely settler or a defenceless city???

 

[ShannonCT]

You pointed out a really interessant fact, what is the behavior of an Aggressive AI's archer in front of a lonely settler or a defenceless city???

I've played several tests, including ones with defenseless settlers and defenseless cities and the AI has never attacked. As far as I can tell, aggressive AI affects their attitude toward you. In a normal game, if an AI's attitude points toward you added up to 0, they would be Cautious, but with aggressive AIs, 0 = Annoyed. So that's going to hurt trading, but I dont see it having much affect on their likelihood to declare war.

 

[radiopill]

Ah, good idea. I wish I had thought of it. So animals do appear on turn 6. I guess my numbers still give an idea of when to expect to actually see a beast. Using your idea, I checked that barbs appear on turn 42. I think these numbers are in the code.



We would have to hope there was a hill to stand on at turn 6 though and even then it might force us to settle somewhere pretty bad. I would still vote for using the warrior escort after turn 6 and finding a prime spot. I'm not sure we want to settle as early as turn 6 even if we did find a good spot from the settler's westward rush. Every turn we wait is 5-6 more beakers, which means getting archers earlier, being able to pop rush in Kyoto earlier, getting a library and scientists in Kyoto earlier, etc...

I was working out the specifics of the distance penalty in the worldbuilder. Every step to the north and west has an effect on the distance penalty. The distance to the capital seems to be calculated as the north/south distance + the east/west distance, rather than the straight line pythagorean distance. Our settler is starting at a total distance of around 42 tiles (6 south and 36 east). One step to the NW would reduce the distance by 2, while one step to the SW would have 0 effect (I'm not suggesting that the settler should never move south). I started playing around with cities at different distances from the capital at different populations (yes, maintenance does increase with population). Here's the results:

Maintenance of 2nd city as a function of distance to capital and population of 2nd city:

Distance......Pop1.....Pop2......Pop3.....Pop4......Pop5
27..............-5.........-6.........-7.........-7.........-8
28..............-5.........-6.........-7.........-8.........-8
29..............-5.........-7.........-7.........-8.........-9
30..............-5.........-7.........-8.........-8.........-9
31..............-5.........-7.........-8.........-8.........-9
32..............-5.........-7.........-8.........-8.........-9
33..............-6.........-7.........-8.........-9.........-10
34..............-6.........-7.........-8.........-9.........-10
35..............-6.........-8.........-9.........-9.........-10
36..............-6.........-8.........-9.........-10........-10

So for the most part, when population is factored in, every step closer to Kyoto will reduce maintenance. It's not as if crossing some magic line will reduce maintenance a lot. But the difference between being 35 tiles away and 32 tiles away is 1 beaker per turn for every turn until we move our palace.

That's a great job tou have done here...

So, the aim of our settler is to find a good place to settle in the NW... but the more I think about it, the more I doubt we can go NW. With the small screenshot we have, it looks like we are almost in the north-west coast of our continent... and as I told earlier Gyathaar is a vicious... I'm pretty sure, he won't let us decrease our maintenance cost as easily...

 

[ShannonCT]

That's a great job tou have done here...

So, the aim of our settler is to find a good place to settle in the NW... but the more I think about it, the more I doubt we can go NW. With the small screenshot we have, it looks like we are almost in the north-west coast of our continent... and as I told earlier Gyathaar is a vicious... I'm pretty sure, he won't let us decrease our maintenance cost as easily...

Thanks. You may be right about the continent. Our settler can start the game by moving westward twice and verify if there is anywhere to go. If not, we will have to decide whether to start looking east. Kyoto is 48 tiles to the east of the settler. He would have to move 12 spaces north/east just to get back to the starting distance (14 including the initial 2 moves). This would require a commitment to waiting longer to settle (which I have no problem with). It might still be worth the risk to move. I will reiterate my point about the advantage of settling close to an AI's capital (fewer barbs, quick attack). I have to think Gyathaar has it set up to punish people who settle near the starting place.

 

[Mad Professor]

There's some good analysis done here. It's good to have some real numbers to look at.

Thanks. You may be right about the continent. Our settler can start the game by moving westward twice and verify if there is anywhere to go. If not, we will have to decide whether to start looking east. Kyoto is 48 tiles to the east of the settler. He would have to move 12 spaces north/east just to get back to the starting distance (14 including the initial 2 moves). This would require a commitment to waiting longer to settle (which I have no problem with). It might still be worth the risk to move. I will reiterate my point about the advantage of settling close to an AI's capital (fewer barbs, quick attack). I have to think Gyathaar has it set up to punish people who settle near the starting place.

<chuckle> It's probably not wise to base decisions on guesses about what Gyathaar might or might not do, but you just might be right.

When I play, I'll try moving west, and if it becomes bvious there's nowhere to go, I can stop right there, even after one or two turns and check back in with you guys to discuss it if necessary. This thought doesn't bother me.

Since we absolutely cannot afford to lose the settler though, I'm nervous about it wandering around after animals are there. Is there some way of analysing the cost/benefit of delaying the settler? It's clear that our science will be better off with a delay, but how is that offset by delay in units produced in the second city, when we'll need them absolutely against barbarians which will certainly appear sooner than we want, and in numbers more than we want? That's a bit hard to measure I suspect, as it really depends on the exact location of the city for what you're sacrificing. I mean - whether that first unit would be produced with one hammer, two, three, etc.

 

[ShannonCT]

When I play, I'll try moving west, and if it becomes obvious there's nowhere to go, I can stop right there, even after one or two turns and check back in with you guys to discuss it if necessary. This thought doesn't bother me.

Since we absolutely cannot afford to lose the settler though, I'm nervous about it wandering around after animals are there. Is there some way of analysing the cost/benefit of delaying the settler? It's clear that our science will be better off with a delay, but how is that offset by delay in units produced in the second city, when we'll need them absolutely against barbarians which will certainly appear sooner than we want, and in numbers more than we want? That's a bit hard to measure I suspect, as it really depends on the exact location of the city for what you're sacrificing. I mean - whether that first unit would be produced with one hammer, two, three, etc.

Well if it becomes clear from the settler's first move that there's no way west, the warrior can start moving NE. Actually, if we can move 6 tiles directly NE, we're still at 42 tiles away from Kyoto (42 east/west and 0 north/south). By the time the warrior moved 6 NE, the settler should be able to catch up and have an escort for any further movement. If the settler finds the end of the continent on the first turn, settling on that coast will mean 7 GPT maintenance and will also mean being farther away from the nearest AI.

If we (you) decide we need to go NE and settle after 6-10 turns, we will be completing our first archer earlier than if we had settled in the first couple turns. If you look at the probability of actually seeing an animal before turn 10 and then the probability of said animal being able to kill a smartly maneuvered combat 1 warrior, the danger is quite small.

 

[KingdomBrunel]

I will play ten turns to begin with, and post results for discussion. Do we want to play ten turns each through the first round, or should I then play a second ten turns before handing it on for everyone to play 20 turns each?

Sounds good to me.

 

[KingdomBrunel]

If we are willing to take a small risk we could rush the settler west for the first 6 turns and then place him on a hill where he could wait for the warrior. From the hill he should be able to see any approaching animal. Of course it would be annoying to be forced to settle the city but we would never loose the settler this way. The idea with the westward rush would be that we migth encounter a coastline that stops further progress west. In that case we might be able to find a spot for settling before 6 turns. Or we might find another civ blocking the west of our continent. As soon as we have found a spot that seems optimal on this continent I think we should found our 2nd city.

I agree wholeheartedly with this - if we can move quickly west, and then settle after those 6 or 7 turns, getting us a dozen squares closer to Kyoto, then that sounds very good. Although not moving that distance may not give the optimal site, there should still be a pretty good site somewhere close.

It's clear that our science will be better off with a delay, but how is that offset by delay in units produced in the second city, when we'll need them absolutely against barbarians which will certainly appear sooner than we want, and in numbers more than we want?

Our science is only better off in the short term - once the city is producing more commerce as its maintenance costs we're better off researchwise. In terms of dealing with the barbs - if we settle and can get two hammers straight off the bat (if I'm not mixing my metaphors), then a warrior will be out in 12 turns (15base cost * 1.5 for epic = 23). So settling on say, turn 7 means we could have two extra warriors by turn 31, useful for exploration, fog busting and defence. They might even have a chance to get a promotion or two on some animals.

Oh, and thanks ShannonCT for the work on the maintenance re distance and population - very useful.

 

[KingdomBrunel]

In terms of research, I think we're all agreed on archery, and BW not long afterwards. I'd like to say I agree with Frederiksberg approach re writing - this would be a very useful tech for Kyoto, and the sooner we can get specialists in here the better, I think the idea (also Frederiksberg's I think) to have this as a GP farm is absolutely right.

 

[Frederiksberg]

Our science is only better off in the short term - once the city is producing more commerce as its maintenance costs we're better off researchwise. In terms of dealing with the barbs - if we settle and can get two hammers straight off the bat (if I'm not mixing my metaphors), then a warrior will be out in 12 turns (15base cost * 1.5 for epic = 23). So settling on say, turn 7 means we could have two extra warriors by turn 31, useful for exploration, fog busting and defence. They might even have a chance to get a promotion or two on some animals.

I think KingdomBrunel does have a point here. Postponing settling after we have found a good city site closer to Kyoto is to some extent just postponing the pain of the maintenance for a new city with little income. It's a phase we have to go through sooner or later. The main question here should be if we can use the extra food and hammers we get by settling early for any usefull purpose. In most of my testgames warriors were not all that usefull and most of them were eventually killed by barbarian archers or axemen. Barracks on the other hand might be useful and we can build them at reduced cost. Our 2nd city will only be able to pay for itself if we have some gold producing tiles available. Gold or silver bonus ressources are optimal but we may not be able to locate those and in that case coast tiles are probably the best since they produce 2gpt unimproved. One more reason for settling at the coast! And maybe a reason for researching sailing somewhere along the line and building lighthouses (Reduced cost). Building cottages is also an option, but not in the early phases before we have created a defensive perimeter around the city.
Assuming that Kyoto remains isolated until we learn astronomy I would say that an important goal is to build 3 cities on the continent our settler is on and build a Palace in one of them to cut down maintenance. After that we have a "normal" game and Kyoto will be the only city with an unusually high maintenance. In order to get there as fast as possible I would prefer not to delay founding our 2nd city for too long.

 

[Frederiksberg]

By the way - here's a thread discussing city maintenance:

City Maintenance