Ring City Placement

[Smirk]

Originally posted by DaviddesJ
I don't really think that RCP does remove much variability. Several people have expressed an intention to try some form of RCP in GOTM21. It will be interesting to see what those games look like, but my suspicion is that they will all look quite different. There are still a lot of choices to make.

I don't expect we'll have to wait long, I assume at least 10 people are going to try it this gotm.

So unless Firaxis decides to change the formulas in a patch,...

That is if its a design error. If not say hello to rings!

Many powerful strategies, valid or not have been cut back, some severly enough that they remained useless for years like pop-rushing. I like the strategy and it seems obvious enough that it can be a generic skill for anyone, but at the same time it seems to fill in a spot of the corruption algorithm so much that it then looks like a flaw in the corruption algorithm.

 

[DaviddesJ]

Originally posted by Smirk
That is if its a design error. If not say hello to rings!

I'm not sure what you mean by this. I'm sure no one at Firaxis thought about it, one way or the other, when they wrote the code. Whether it's an "error" is rather subjective.

Many powerful strategies, valid or not have been cut back, some severly enough that they remained useless for years like pop-rushing.

I'm not sure I understand what you mean, but I don't know all of the history. Pop rushing certainly isn't useless in the current version (PTW 1.21f).

My impression is that Firaxis considers themselves pretty much done with fixing game balance issues. Sure, there were some changes in early patches, but nothing in a long time now. If they are open to some changes, there are some other things than this that I'd like to see changed (i.e., bigger exploits for the human players).

 

[Elandra]

Wow, powerful idea, great research. I'll try it on the next GOTM with a favorable map.

You gotta love a game discussion forum that includes references to engineering paper, spreadsheets, and homemade graphical overlays! Where do you all find the time??? The culture here is as enjoyable as the strategic analysis!

 

[Smirk]

Originally posted by DaviddesJ
I'm sure no one at Firaxis thought about it, one way or the other, when they wrote the code.

I'm not sure I understand what you mean, but I don't know all of the history. Pop rushing certainly isn't useless in the current version (PTW 1.21f).

Thats my point. Many of the game design and balance issues were the same thing. Pop-rushing as a way of life was obviously not considered in early game testing since it was very powerful. (BTW This is circa 1.17 back when Civ3 first came out.) The initial knee-jerk reaction was to cut the production down to half in addition to some other changes. It wasn't truelly fixed until around civ1.25: back was the 20s production, no bottom to the unhappiness and the unhappiness stays with the civ when abandoning a city. (This isn't the case when the AI does it, only the human.)

I'm not so sure about the game balance issues, why would they just stop when there is still another expansion planned? They may not be updating Civ, but PTW and their expansion may get the changes.

I don't want to sidetrack the thread, so only time will tell weither firaxis cares or not.

 

[DaviddesJ]

Perhaps a not-crazy idea is to try to time the palace jump to coincide with a change in government (for non-religious civ). This would give several turns to relocate cities while not worrying about increased corruption from the overlap. Unfortunately, you can't rush production in anarchy, so you can't get the people out of your old cities, which is really the time-consuming part of the move. So there's no great way to do it, except accepting several turns of overlap, or abandoning many citizens. Unless there's something I'm missing.

 

[DaveMcW]

I agree with Smirk and the others who value strategic placement over RCP in the first ring of cities.

In the short term, an extra bonus shield or freshwater can outweigh the corruption benefits of first ring RCP. In the long term, these cities will be the first to benefit from roads, republic, courthouses, and WLTKD; eventually there won't be any corruption for RCP to eliminate.

The second ring is where RCP really shines. RCP gives increasing returns as the number of cities increases, and you can fit a lot of cities into the second ring. There isn't as much pressure to have immediately productive second-ring cities; you can afford to place them correctly and watch them grow into powerhouses.

For a palace jump, you should ignore first-ring RCP and concentrate on getting a good second ring around the FP.

 

[Smirk]

Originally posted by DaviddesJ
Unfortunately, you can't rush production in anarchy, so you can't get the people out of your old cities, which is really the time-consuming part of the move.

Shield adding processes still work, like cutting a forest or disbanding a unit. Just get the settler almost finished, revolt then disband or cut the forest and the settler should pop. I don't know if I've ever done this myself, so there may be some food micromanagement, not sure.

 

[DaviddesJ]

Originally posted by Smirk
Shield adding processes still work, like cutting a forest or disbanding a unit. Just get the settler almost finished, revolt then disband or cut the forest and the settler should pop.

I guess you could do this by disbanding units. But if you have several cities with 6+ pop to get out, that's a lot of shields of units to disband (i.e., you could use 6 turns of anarchy to get 6 workers out of a size-6 city, but you'd have to have 240 shields of units to disband, which seems hardly worthwhile).

 

[rabies]

Originally posted by Smirk
I don't expect we'll have to wait long, I assume at least 10 people are going to try it this gotm.

1 of the 10 getting caught up here for the first time...i had not noticed this thread before. i would like to thank the guys for posting the graphics for determining distance, because much like Moonsinger, I was having a hard time figuring it out exactly from the text.

I can definately see a difference applying this technique in gotm21...but there certainly are enough variables (how far to make each ring mostly) in there that i think comaparing GOTM games of people who used the strategy should be educational.

 

[whstaff]

this is a powerful idea and I appreciate the work that went into it, might be mandatory at higher levels. But for most who play at lower levels, I would suggest you use the more pragmatic less esoteric approach. Expand toward strategic resources, the trade resources, and the natural boundaries of your civ, fill in the empty space later. Let's face if you get all the 8 trade goods for your civ and the right wonders your citys will spend a large portion of the game in WLTKD and corruption is not an issue then. If you have steel and saltpeter and your neighbor dosen't he's basically dead meat.

 

[Yndy]

Question for DaviddesJ (or anybody else that tested the RCP and the corruption formula):

Say I have a 3.5 ring around the Capitol and I build the FP on the other side of the map. Will the new ring around the FP be influenced by the one around the Capitol?

Should I also build the ring around the FP at 3.5 to minimize corruption?

In my game I built rings of 3.5, 7 and 10 around the FP and rings of 5.5, 8 and 14 around the Palace. Is this less than optimal?

 

[DaviddesJ]

The two sets of rings are entirely independent (unless they overlap). It makes no difference whether you choose the same distances or different distances. In other words, you did fine.

 

[Txurce]

Yndy, would the palace be in Berlin, by any chance?

DaviddesJ, what confuses me about the "overlap" question is that, in theory, two rings will eventually overlap. I guess the ideal would be that the overalpping area is so far from either core that it is totally corrupt. But if a city could theoretically fall under the "gravitational pull" of both palace cities, is it safe to say that the palace it's geographically closest to will dominate... or would it be where it ranks in terms of city number for that series of rings?

 

[DaviddesJ]

For each city, figure out which core it's closer to (Palace or FP), and then compute the corruption based on its distance to, and rank from, that core.

If the Palace and FP are relatively close together, then the rings will "overlap". Essentially, that means that some of the locations are wasted, and you aren't getting the full benefits of RCP.

For example, in Yndy's example (rings at 3.5/7/10 from the Palace and 5.5/8/14 from the FP), suppose that a city at distance 10 from the Palace is also distance 7 from the FP. Then the Palace distance is probably irrelevant (the corruption based on the FP will be lower), so there's no point in putting the city at that particular location, as there's no benefit to fitting into the distance-10 ring, and it's not getting the benefit of RCP.

Now, suppose the city at distance 10 from the Palace is distance 11 from the FP. Then the ring placement does help that city get low corruption in terms of its distance from the Palace. But the city is a misaligned city from the point of view of the FP: it increases the corruption of all of the ring cities at distance 14 from the FP, even though that city itself isn't getting any benefit from the FP.

Hope this helps.

 

[nimbus99ca]

Just out of curiosity have you compared this rather complex formula with a slightly simpler method where you just count the number of squares in the NE,NW,SE,SW directions ONLY to come up with a total distance from the origin?

I noticed that by going exactly 5 squares (and building on the "6th" square that my corruption stayed extremely low in all the cities and I was able to expand my next ring using the exact same formula and the 2nd ring was also extremely efficient. I also noticed that I was able to very efficiently use the map using this strategy (mind you 5 square in the same direction would waste 3 useable squares in the middle and I tended to go 4 and 1 or 3 and 2 to space the cities out).

I was also surprised at how easy it was to then find squares for which the same formula would apply to all adjacent cities as well so that just about all my cities obeyed this rule for distance between them???

Any comments?

 

[Txurce]

DaviidesJ, thanks for the thorough explanation. You answered questions I had yet to formulate.

One of RCP's drawing cards is that while not as simple as 2+2, it's fairly easy to grasp its concept and, having done so, doesn't require high game-playing skill to employ it. RCP helps everyone, but it should supercharge the games of mid-level players.

 

[Qitai]

This is working out better than I thought. I had no intention to restrict myself to use this in GOTM21. But after unintentionally place 3 cities in RCP position, I kept to it and liked the reduce corruption very much.

 

[Yndy]

Originally posted by Txurce
Yndy, would the palace be in Berlin, by any chance?

No, even if I considered it.

I'm contemplating whether the AI tries to build a 5.5 RCP ring or it just happens to do so. I spotted several positions to build the palace on former AI lands and to get 2-4 cities at 5.5 RCP.

5.5 is decent distance for minimal overlap actually.

 

[DaviddesJ]

The AI placement algorithm seems to aim for covering all territory with minimal overlap (OCP style), which often puts their cities 5.5 apart. So (at least in my limited experience), when I try to modify their existing city placement to an RCP style, I often end up with a 5.5 ring.

IMHO this is ok if the FP is well inland, so that you get a complete or nearly complete 5.5 ring, and also if you don't happen to hit many mountains or lakes or other tiles you really don't want to build on. It may be too far out for a good inner ring if the FP is near the coast. Of course, unlike the initial place, you can generally choose where your FP goes; it doesn't have to be the AI's capital, so it's usually easier to get a good location, as opposed to the potluck you get when you start.

 

[Smirk]

Originally posted by nimbus99ca
Just out of curiosity have you compared this rather complex formula with a slightly simpler method where you just count the number of squares in the NE,NW,SE,SW directions ONLY to come up with a total distance from the origin?

Effectively what you are saying is that you have two oscillating rings in each level, 3 squares going along the tiles is a distance of 3, whereas going diagonal (across the tiles) its 3.5. A simple case of this would have 4 cities at 3 which will all have the same corruption as a second city, and then 4 at 3.5 will have all the same corruption as a 6th city.


This was talked about briefly in the first page, and the difference in corruption will be small for most maps, and you could stand to benefit by more flexibility in placement. But for the second ring and beyond this wouldn't really help out all that much unless you stay under the optimum number of cities for a map.