Version 0.60 Discussion Thread

[MadmanAtW]

I saw that happen in a multiplayer game I'm hosting running 0.50f as well- one random tile of blockade in an area where I had just killed a bunch of barbarian galleys. It didn't occur to me to keep a saved game, but I can check and see if one of the autosaves has it- it was pretty recent.

 

[TopDog]

Awsome work.

Anyone got this plus an increased civs limit dll merge?

EDIT: dont worry, found it. Right there on the downloads page... woops

 

[Arlborn]

I also got this bug they are talking about in a game, but it was 4 tiles if I remember correctly. I forgot to report at the time, sorry.


I also see that they do take over more barbarian cities than what I remember. Very good for them indeed.

 

[Bursk]

I appreciate that changes have been made to encourage AI's to settle more coastal cities, but is there anything that can be done to stop this sort of thing from ever happening:

http://img205.imageshack.us/my.php?image=civ4screenshot0008vp2.jpg

 

[Roland Johansen]

- AI now has extra encouragement to put cities on rivers early in the game

What's so special about the start of the game to encourage the AI to place cities on a river? Trading, health? I don't think that trading using rivers is important during most games as the AI quickly builds roads and rivers rarely connect many cities and coastal areas are far more important for trade during the start of the game. The health bonus is usually not important during the start of the game as during the start of the game happiness is usually the bottleneck. Many of the health resources are available early in the game and and the health bonus from forests is also still available in the early game.

So why emphasise river sites extra (above and beyond normal emphasizing of river sites) during the start of the game?

The risk is of course that river locations are picked while a non-river position is better in the long run.

 

[weimingshi]

Is there any features that can show me in game that if I'm playing with better Ai and which version I have? Something like if you hold down ctrl and hover mouse over your name on the scoreboard and it will show "Better AI v6.0"? with all those mods installed, I'm not always sure if I set things up right. If you can implement an indicator in game, that would be great.

 

[MadmanAtW]

Hover over the flag- it'll tell you your Better AI version along with your Civ traits.

 

[Lone Wolf]

What's so special about the start of the game to encourage the AI to place cities on a river?

Future levees, perhaps?

 

[jdog5000]

I appreciate that changes have been made to encourage AI's to settle more coastal cities, but is there anything that can be done to stop this sort of thing from ever happening:

http://img205.imageshack.us/my.php?image=civ4screenshot0008vp2.jpg

Is it the general 1-tile from coast placement you're objecting to or specifically because there are clams?

The possible coastal spots for the city are:

1
2C
34

where C is the current location of the city. Both 1 and 2 lose access to the banana, so it's a least a logically-based decision. Sites 3 and 4 would be good choices, my bet though is that the AI lowers their value because they are 1 tile farther away from its existing cities (can't quite tell given the fog). There's a strong weight towards producing tightly spaced empires where cities actually overlap a little bit, without any outside forces (barb cities, other players) it takes a serious desert or something for the AI to put gaps between its cities.

I think that's what's going on there.

To answer you question directly though, yes it is possible to effectively block 1-tile from ocean cities ... you'll notice it never happens with start locations now, thanks to Blake.

What's so special about the start of the game to encourage the AI to place cities on a river? Trading, health? I don't think that trading using rivers is important during most games as the AI quickly builds roads and rivers rarely connect many cities and coastal areas are far more important for trade during the start of the game. The health bonus is usually not important during the start of the game as during the start of the game happiness is usually the bottleneck. Many of the health resources are available early in the game and and the health bonus from forests is also still available in the early game.

So why emphasise river sites extra (above and beyond normal emphasizing of river sites) during the start of the game?

The risk is of course that river locations are picked while a non-river position is better in the long run.

My thought in boosting it early (ie pre-bridge building) was to help trade and religion spread early in the game. It's true that the AI does a good job of road building so trade is not a large issue ... the river weighting is not a large piece though either. It was +40 and now is +60 early in the game and +40 still later. A typical city found value ranges from 1200 for artic tundra to 4000+ for prime sites. For this change to be decisive, both potential cites have to be about the same so it's not going to cripple the cities future or anything. The coastal bonus, for comparison, was +400 now is up to +700 in some circumstances.

Does a city on a river connect to the coast also once the right techs are known, or am I making that up?

The buildings/wonders which require the city to be on a river are:
- Levee (Dike also works with ocean)
- Hydroplan
- Three Gorges

The levee was introduced in BTS and is quite powerful in the late game ... perhaps a boost for rivers is in order then.

I just did some historical digging for fun, here's the numbers for these different factors (for non-starting plots) over time:

Vanilla and Warlords:
Coast +800
River +60
Fresh water +40 +20*health: change)

Warlords Better AI:
Coast +400
River +40
Fresh water +40 +30*health: change)

Vanilla BTS:
Coast +400
River +40
Fresh water +40 +30*health: change)

Better BTS AI 0.60:
Coast +500-700 (depending on # of existing coastal cities)
River +40-60
Fresh water +40 +30*health: change)

So, given that history I think I'll make coastal access +600-800 (would probably be the difference in Bursk's scenario above), and river access +60 always.

Thoughts?

 

[Arlborn]

Hum, but wasn't there a constant complain about the Vanilla AI making too many coastal cities? I think I remember that... Maybe I am wrong though.

 

[Bursk]

jdog, yes, it's the position of the city I'm not happy with. I can't for one second (unless I'm missing something) see how where the city currently is can be better than location 3 or 4 in your example.

 

[Yakk]

Last I looked at the city placement AI, it valued resources outside of the fat cross rather highly.

You will notice where it is, at radius 4 culture(!), it catches a marble and a wine. Moving to either 3 or 4 costs it the marble and wine until radius 5.

Now, as a human, I aim for cities based mainly off of the fat cross. I sometimes arrange a city so that it can pick up a resource at radius 3 in an otherwise crappy area that isn't being approached by another civilisation. I do not think I've ever taken into account radius 4 resources when placing a city, unless you are talking about "set up a city to discourage an AI from expanding towards that resource", which is about topology not distance.

I don't know if the AI currently pays attention to radius 4 resources when placing cities.

 

[Roland Johansen]

My thought in boosting it early (ie pre-bridge building) was to help trade and religion spread early in the game. It's true that the AI does a good job of road building so trade is not a large issue ... the river weighting is not a large piece though either. It was +40 and now is +60 early in the game and +40 still later. A typical city found value ranges from 1200 for artic tundra to 4000+ for prime sites. For this change to be decisive, both potential cites have to be about the same so it's not going to cripple the cities future or anything. The coastal bonus, for comparison, was +400 now is up to +700 in some circumstances.

Clear, the bonus doesn't add a lot. Thanks for the details.

Does a city on a river connect to the coast also once the right techs are known, or am I making that up?

After the development of sailing, a river placed city can trade along a city-river-coast-(river)-city path as long as the path is visible to the civilisation. This can in rare cases help. However, it won't help often as a civilisation developing sailing will typically already have a coastal city and road building and thus it will typically not need this river-coast trading path as it has a direct coastal trading path originating in a coastal city. But sometimes river trading is helpful to connect to foreign civilisations before you have coastal cities when you use map types that consist of lots of land and not a lot of sea.

The buildings/wonders which require the city to be on a river are:
- Levee (Dike also works with ocean)
- Hydroplan
- Three Gorges

The levee was introduced in BTS and is quite powerful in the late game ... perhaps a boost for rivers is in order then.

I just did some historical digging for fun, here's the numbers for these different factors (for non-starting plots) over time:

My first thought to avoid 1-tile-from-the-coast-cities would be to add a negative value to the city founding value for each coastal tile in such a city and a smaller negative value for each sea tile in such a city. (example values: -200 for each coastal tile, -100 for each sea tile). The reason is clear: one coastal tile in such a non-coastal city is bad but could in some rare cases be a better city than the coastal cities in nearby locations, but 5 such tiles in a non-coastal city is appalling. This in addition to the flat bonus for coastal placement.
But I guess such a valuation is already present.

If the value for river placement is enhanced because levees were added to the game, could the bonus then be related to the number of buildings buildable by the civilisation which benefit from river placement. In that case the valuation would be more mod-friendly. In my personal mod, levees don't need the strict river placement.

Last I looked at the city placement AI, it valued resources outside of the fat cross rather highly.

You will notice where it is, at radius 4 culture(!), it catches a marble and a wine. Moving to either 3 or 4 costs it the marble and wine until radius 5.

Now, as a human, I aim for cities based mainly off of the fat cross. I sometimes arrange a city so that it can pick up a resource at radius 3 in an otherwise crappy area that isn't being approached by another civilisation. I do not think I've ever taken into account radius 4 resources when placing a city, unless you are talking about "set up a city to discourage an AI from expanding towards that resource", which is about topology not distance.

I don't know if the AI currently pays attention to radius 4 resources when placing cities.

This sounds very weird and anyone reading this would probably immediately say that such valuation is not very wise. An explanation for this weird valuation could be that it is needed to avoid cities that greedily grab all resources, leaving nothing for competing cities that need to be placed in adjacent areas. I do remember reading something about cities greedily grabbing resources in the BetterAI mod for Warlords and that it was an issue.

Personally I think it is logical that resources are highly valued in city placement, but maybe they are valued so high (above and beyond their pure valuation for resources like food, hammers and commerce) that they force city placement for the AI too strictly.
If you give some reasonable value to resources and place cities not too far apart, then resources will always be claimed by one of your cities and they will not dictate city placement too strictly.
When cities are placed 1 tile of the coast, then I guess it will typically always be resources that cause this decision by the AI.

I also think that the AI should give some flat bonus to cities that claim resources that the AI doesn't have. I often see a fur/silver location that the AI won't claim because the city doesn't have enough food. But such a city could help the entire empire with happiness. The same for a desert incense city with no food nearby.

 

[Arlborn]

I also think that the AI should give some flat bonus to cities that claim resources that the AI doesn't have. I often see a fur/silver location that the AI won't claim because the city doesn't have enough food. But such a city could help the entire empire with happiness. The same for a desert incense city with no food nearby.

Well, I've seen many 1 pop cities in the middle of ice built on top of a silver or fur all right. But normally it is when the AI ran out of places to expand otherwise.

 

[Roland Johansen]

Well, I've seen many 1 pop cities in the middle of ice built on top of a silver or fur all right. But normally it is when the AI ran out of places to expand otherwise.

Yes, the AI will build cities anywhere when it's out of premium spots, but whether the AI will eventually settle that spot is not the issue that I was raising in that part of my post. The issue is that the AI doesn't recognise that one poor city when purely valued by it's own commerce, production and food output can still be great for the empire when judged by the fact that it allows a +1 or +2 happiness throughout the empire. A flat bonus for new resources could alter that.

 

[Yakk]

So..

"Shore/Ocean Square without Coastal" is a crappy square.
"Shore Square with Coastal" is a decent square.
"Lake Square without Coastal" is a decent square
"Lake Square with Coastal" is a good square

This still neglects the fact that a coastal city can build boats, and a coastal city tends to use up fewer squares you could use for other cities (or, alternatively, a close-to-coast city prevents the ocean squares from being used by anyone). That's a further reason to build along the coast.

The 2nd of these reasons (it blocks other decent sites/squares) is an additional reason to not build "close, but not on, the coast" cities -- especially given that the AI doesn't have the number-of-city maintenance (ie, the AI can afford some city spam) problems that a higher-difficulty human is saddled with.

 

[PieceOfMind]

I'm not sure if it's all that relevant but can I just add, sometimes I will found a city 1 tile off a coast when I would not want that city to be vulnerable to attack from the sea. Firstly, ships can bombard defenses down, trade routes can be blockaded by privateers (which are now a real thing to worry about in Better AI), sneak amphibious attacks etc.

Coastal tiles are not all that valuable even when the city is coastal, so I never see a problem if having 2 or 3 unusable tiles means I will gain 2 or 3 more grassland tiles or an extra resource.

Here is an example like I'm talking about...(see pic)

Access to the currently unused iron mine was important, and using up the grass tiles that didn't fall inside the other cities was important too.

Spoiler warning: Location of iron for the Dual Deity Challenge game from S&T forum.

Spoiler :

 

[Yakk]

Blocked trade routes only blocks if you don't have a land route, right?

In any case: Moving one square left is somewhat tempting.

Where you are: Hill Defense, 2 more grassland tiles, 2 nigh-useless water tiles.
One square left: Water access, 1 more grass hill tile, 1 more plains, 5 coast tiles
Shared: 3 grass woods, 2 plains woods, 1 grass hill, 3 plains, 1 plains water.

Pre-biology food budget, on hill: 5 grass (+5), 1 iron (-2), 1 windmill (+0), 6 plain farms (+0) = +3
Production = Iron + 8
Commerce = 2
Size = 13
Cost per size lost = 1 production
Windmill <-> mine: -1 food +2 production -1 commerce

Pre-biology food budget, on coast: 3 grass (+3), 1 iron (-2), 2 windmill (+0), 7 plain farms (+0), 5 coast (+0) = +1
Production = Iron + 10
Commerce = 14 (with windmills)
Size = 18
Cost per size lost: 2 commerce or 1 production
The "on coast" solution also becomes sick post-biology, but that is a long way off.

The "on hill" solution grows faster. And you are in the ancient times, so windmills aren't allowed. It also peaks better at lower population levels.

 

[PieceOfMind]

No apparently privateers will remove the trade routes from your city if you are not at war with the civ who is privateer'ing you. Oddly, if you are at war, the privateer's blockade becomes an ordinary blockade - just preventing use of the tiles and removing those water tiles from the trade network, but not removing the city's trade routes.

Oh, and 775AD.

In this situation the benefit of the defense from the hill was a big factor for me, as I was expecting to get attacked by massive stacks eventually. I was going a cottage heavy economy (as I usually do) so maximising the number of grass tiles was another high priority, and I never feel good settling on a grass tile because that way you aren't getting any free food from the city tile.

 

[Roland Johansen]

So..

"Shore/Ocean Square without Coastal" is a crappy square.
"Shore Square with Coastal" is a decent square.
"Lake Square without Coastal" is a decent square
"Lake Square with Coastal" is a good square

This still neglects the fact that a coastal city can build boats, and a coastal city tends to use up fewer squares you could use for other cities (or, alternatively, a close-to-coast city prevents the ocean squares from being used by anyone). That's a further reason to build along the coast.

The 2nd of these reasons (it blocks other decent sites/squares) is an additional reason to not build "close, but not on, the coast" cities -- especially given that the AI doesn't have the number-of-city maintenance (ie, the AI can afford some city spam) problems that a higher-difficulty human is saddled with.

A flat bonus for coastal cities should take care of the two benefits that you mention: boat building ability and ability to use more tiles in total. Such a flat bonus is already in place. The value of this bonus can of course be discussed.

Another important reason to build on the coast is the better trade routes.

Here is an example like I'm talking about...(see pic)

Hmm, in this case I'd agree with Yakk: 5 coastal tiles + 1 grassland hill + better trade routes vs 2 grassland tiles (and 2 worthless non-harbor coastal tiles). Note also that the bigger coastal city will also get better trade routes due to its size next to the bonus for being coastal. The bonus for trade routes alone will be comparable to 2 (free speech, universal suffrage) towns in the late game (that's a guesstimate). Early on, you don't get great towns, the value will be similar to coastal tiles for a while. So purely from a commerce yield point of view, I think that the coastal city is better.

Then next to that you get the benefits of health (harbor) and ship building. I do understand the military disadvantages.