AU mod: Balancing the Governments

[alexman]

Originally posted by nbarclay
It actually makes Republic better than even under the standard rules (which have Republic 1/3/4) until a civ grows six cities to size seven or higher, and better than 1/2/2 until a civ grows twelve cities to size seven or higher.

This does not concern me very much, because the Republic has been weakened quite a bit as an initial government in C3C. The main reason for the AU mod changes to the Republic has been to weaken this government's position as the best late-game all-purpose government in C3C.

Having said that, I do like the 0/1/1+18 proposal a bit better as well. It seems like the smaller of the two changes, compared to the current version of the mod.

I suggest we put that proposal (0/1/1+18) under consideration now, so that if it is voted in, we may package it with the Cavalry change for version 1.07.

[nbarclay]

Originally posted by alexman

This does not concern me very much, because the Republic has been weakened quite a bit as an initial government in C3C.

That depends on the strategy being used. For example, a civ with one and a half workers and half a military unit per city breaks even with earlier versions of Civ 3 under the default C3C rules even with all its cities size 6 and under. Granted, the heavily builder-oriented opening game I generally favor is almost certainly a bit atypical, but it and C3C's new version of Republic can fit together very well.

In CivFanatics' C3C GOTM 4 (the September game), I completed the transition to Republic in 1125 BC. (Since the game was on Monarch, with an Agricultural civ and a relatively good starting position, I had no difficulty researching Code of Laws before Philosophy and then getting Republic as my free tech.) At the time, I had 18 units: two warriors, 14 workers, and two settlers. I also had eleven towns, providing free support for eleven units. So even with an earlier transition to Republic than would have been possible prior to C3C, at that point, I was four gpt ahead of where I would have been in the older versions of Republic.

Jumping ahead to 630 BC, which is more like the traditional time to switch to Republic (and the next save I made after 1175 BC), I had 36 units: 25 workers, 8 warriors, 2 Javelin Throwers, and a galley. For those 36 units, I was paying a grand total of eight gold per turn in upkeep. That's a savings of 28 gpt compared with what I would have paid prior to C3C.

So while the C3C changes to Republic can make Republic more expensive as an initial government than in previous versions, they can also provide a significant advantage for a player who is willing and able to cut the number of units to the bone to take advantage of the new support cost model.

By the way, looking at how I would have done in free support in that game in 630 BC with the cost models currently under discussion, I had six cities size seven or higher and fourteen size six or under.

1/3/4: 32 units (Stock rules)
1/2/2: 26 units (Original AU Mod change)
0/1/1 + 12: 18 units (Current AU Mod rules)
0/1/1 + 18: 24 units (My current proposal)
1/1/1 + 12: 32 units (pvzh's proposal)

[Theseus]

/me shuns Republic (but 0/1/1+18 would be a little better).

[nbarclay]

Okay, 0/1/1 + 18 free support (for both Republic and, by extension, Democracy) is now officially under consideration. Voting in a week (or we may shave a little off that on a provisional basis if we're otherwise ready to start the next AU game before then).

[player1]

My personal opinion, is that every government with fixed free unit support (like AU Repuiblic), is not well balanced on all types of maps.

Fixed suppost cost favors smaller maps, but are inadequate for bigger maps.


Anyway, I see some inflation of power in govenments.
First, you make Feudalism better, then you make Repuiblic better.

And what to do poor Monarchy?




IMHO, if I look at original govenments, in pre-modern era I used Feudalism as often as Republic as government of choice.
Lots of towns --> Feudalism rules, lots of cities & lux --> Republic rules.

Only balance problem was late Republic (and Democracy inferiority if not Religious).


Anyway, if you ditch free fixed support cost and have original 1/3/4, only way to town it down for later era is to lower support.

But, there is a small problemt.
1/3 town/city support helps in transition era when entering Republic. 1/2 town/city usually makes enetering really harsh.

But on the other hand when you get just cities and metros it's way to easy.

Possible option would be adding 2/2/2 support, but that would favor Feudal lots of towns aprouch and would relativly weaken Fedualism as chocie.

Anyway, I don't have solution for this question.

[nbarclay]

Originally posted by player1
My personal opinion, is that every government with fixed free unit support (like AU Repuiblic), is not well balanced on all types of maps.

Fixed suppost cost favors smaller maps, but are inadequate for bigger maps.

That's true to some extent, but keep in mind that the number of civs on a map scales with map size. As long as players play with a normal number of civs for their map size, the amount of territory a player can claim REXing doesn't vary as much with map size as one might think just from looking at map sizes themselves.

Also, I hadn't thought about it before, but Republic's advantage from its commerce bonus is itself something that is more potent on larger maps. For any given number of cities, corruption is not as big a problem on big maps as on small ones. Thus, on bigger maps, corruption does not eat away at the commerce bonus as much as it would on smaller ones. That's in addition to the possibility of having more cities to get the commerce bonus from on a bigger map. So while free support might be more of a problem on a big map, the greater advantage from the commerce bonus means players can afford to pay to make up the difference. Similarly, on a big map with relatively few civs, the combination of greater advantage from the commerce bonus per city and more cities to get the commerce bonus in makes the relative lack of extra support per city not such a big deal.

I wouldn't be surprised if 0/1/1 + 18 does make Republic a little better than we want it to be on smaller-than-standard maps, at least in the short term. (In the longer term, a civ only needs nine cities size seven or higher before 0/1/1 + 18 becomes no better than the 1/3/4 of the standard rules.) But I don't remember our ever having an AU game on a smaller than standard map, and my impression is that most people around here don't play on the smaller map sizes much except maybe in PBEM games. (By the way, for that plus other reasons, we might want to go with 1/2/2 or even back to 1/3/4 free unit support for Republic in the PBEM version of the Mod. In PBEM, it's a lot harder to get away with keeping one's military small and cheap.)

I certainly wouldn't want to switch part of the free support over to a fixed model for a government without a commerce bonus. But for Republic, with its commerce bonus, I don't think scaling to larger map sizes is a problem.

Anyway, I see some inflation of power in govenments.
First, you make Feudalism better, then you make Repuiblic better.

And what to do poor Monarchy?

Keep in mind that what we're making Republic "better" relative to is itself a LOT worse for Republic than the standard support model. In the standard rules, Republic's free support is 1/3/4. Our current AU Mod rules have it at 0/1/1 + 12, which a lot of us think has weakened it too much. Taking the free support up to 0/1/1 + 18 is really just intended to give back a small part of what we've taken away. Monarchy will still be significantly stronger relative to Republic than it is under the standard rules.

IMHO, if I look at original govenments, in pre-modern era I used Feudalism as often as Republic as government of choice.
Lots of towns --> Feudalism rules, lots of cities & lux --> Republic rules.

If you used Feudalism a lot under the standard government designs, you're the only person I remember hearing from who did so. Most of us view Feudalism under the standard rules as having only a small niche where it's worth using, with either Monarchy or Republic (if not both) being better in the vast majority of games.

Only balance problem was late Republic (and Democracy inferiority if not Religious).

In my view (and that of most of the other long-time AU regulars), there is also a serious imbalance between Monarchy and Republic under the standard rules. Republic's commerce bonus can easily pay to make up the difference in unit support plus the cost of an extra notch on the luxury slider to make up for the lack of military police. And with the power of bombardment units in C3C, a good player can do a lot of fighting in Republic without war weariness becoming a serious problem. The end result was that for a lot of us, using Republic instead of Monarchy became essentially a no-brainer. So we weakened Republic's free unit support to make Monarchy (and Feudalism) more interesting.

Anyway, if you ditch free fixed support cost and have original 1/3/4, only way to town it down for later era is to lower support.

But, there is a small problemt.
1/3 town/city support helps in transition era when entering Republic. 1/2 town/city usually makes enetering really harsh.

If you have so many units that it's that serious a problem, why are you choosing Republic over Monarchy or (depending on city sizes) Feudalism?

We actually tried a 1/2/2 free support model for Republic for a while (I think AU 601 was the only AU game we played under that model) but found that it still left Republic stronger as a warmongering government than we wanted it to be (thereby cutting into Monarchy's proper role). The problem is, every town conquered means free support for one additional unit, growing to two free additional units when the town grows into a city. So a large empire can have a fairly large military while paying very little in support costs.

The reason for the 0/1/1 + 12 (or + 18) support model is that it cuts very heavily into the advantage Republic can provide in extra free support when a civ conquers a lot of territory. Once the conquered areas get over (or back over) size 6, they can provide support for a small garrison to defend them, but that's about it. Which, in turn, gives players a lot more reason to consider using Monarchy if they want to do a lot of fighting.

[pvzh]

BTW, why not improve monarchy somewhat in terms of unit support: switch it to 4/4/6 instead of 2/4/8 like in stock.

[alexman]

Monarchy is already well-balanced as an early war-time government, IMO. Improve it and you will have to improve Feudalism, and we start the cycle all over again!

[player1]

Actually, my game play experience shows that Feudalism is superior to Monarchy in every game that doesn't depend from long war wearines.

Why?

Beacuse if you have smaller cities it leads to great gold gain.
On the other hand if you have large cities, then why aren't you in Republic already?

And pop-rushing is good when war-mongering.

[alexman]

Beacuse if you have smaller cities it leads to great gold gain.

The problem is that in most games people do have large cities by the time they get Feudalism.

On the other hand if you have large cities, then why aren't you in Republic already?

You might not have enough luxuries, so you need the extra MP, or you might have too many units, so you want to avoid paying 2gpt for each of them above the limit. You might also be suffering from war weariness.

[player1]

Well, it's pretty much a thing of style.

In some games, when I have room for peace development, I don't use Feudalism.

If on the other hand I'm pressed early, which means no aquaducts and lots of units, Feudalism just rules.


Anyway, i think that Monarchy lost some appel in C3C since Republic got stronger with larger cities.

[pvzh]

Since we gonna try Fascism for next AU game, why not attempt to fix it up before the game?

I will suggesst to remove forced resettlement flag, and reduce slightly unit support (to 4/6/8 from 4/7/10), so AI will not be lured in by free unit support. Lowered unit suport will not be a problem for humans because they do not build that many of defensive units. Plus, to make sure AI stays in democracy unless they experience heavy war weariness why not up free units for Democracy to stock republic 1/3/4 or thereabouts (1 gpt for excess units). Humans unlikely will benefit much of that anyway: if they peaceful it is slight increase in revenue, if not they will not be a democracy in the first place.

[nbarclay]

I strongly oppose the idea of making changes to the rules solely for one particular game. Doing so would significantly undercut our ability to apply lessons learned in that game to other games. And in any case, I see no reason why we should rig "Power of Fascism" to make AIs more likely than normal to choose or stay in Democracy.

Regarding the "Forced Resettlement" flag, it looks like removing it was included in the proposal Alexman placed under consideration but not in what we ended up voting on. A quick search for references to that flag reveals no trace of a reason why it was excluded from the voting, but it's possible that I missed something. Therefore, I suggest that we go ahead and vote on that change unless someone can come up with a reason not to in the next day or two. (Alexman, do you remember anything about why that flag ended up being excluded from the vote?)

[Theseus]

I don't even remember what "Forced Resettlement" is...

[nbarclay]

The Forced Resettlement flag is what causes civs to lose pop points when they switch to Fascism. Eliminating that flag would make Fascism a little less costly for AIs and a little less unattractive to human players.

[alexman]

Originally posted by nbarclay
Regarding the "Forced Resettlement" flag, it looks like removing it was included in the proposal Alexman placed under consideration but not in what we ended up voting on. A quick search for references to that flag reveals no trace of a reason why it was excluded from the voting, but it's possible that I missed something. Therefore, I suggest that we go ahead and vote on that change unless someone can come up with a reason not to in the next day or two. (Alexman, do you remember anything about why that flag ended up being excluded from the vote?)

I withdrew the initial proposal because I thought it would be better to leave Fascism as a underpowered government, and just try to discourage the AI from using it.

Currently we have a minimum change, which preserves the stock flavor of Fascism, and it's better for the AI, which would often make the wrong choice between the two late-game war-time governments, if those were balanced.

[Theseus]

Originally posted by nbarclay
The Forced Resettlement flag is what causes civs to lose pop points when they switch to Fascism. Eliminating that flag would make Fascism a little less costly for AIs and a little less unattractive to human players.

Thanks... damn, I forgot about that... the few times I've gone Fascist, I remember being shocked at the pop penalty.

If this is in fact going to be the next game, I think part of the opening post needs to be a clear spelling out of what Fascism entails.

[nbarclay]

Originally posted by alexman


I withdrew the initial proposal because I thought it would be better to leave Fascism as a underpowered government, and just try to discourage the AI from using it.

Currently we have a minimum change, which preserves the stock flavor of Fascism, and it's better for the AI, which would often make the wrong choice between the two late-game war-time governments, if those were balanced.

If eliminating the Forced Resettlement flag would make AIs noticeably more likely to choose Fascism, I agree with that logic. But earlier in the thread, you'd written

...Instead, I propose to remove the 'forced resettlement' from Fascism (which the AI does not take into consideration at all)...

If that's true, removing the flag, and hence making the government more worthwhile when AIs do end up using it, would not cause that problem. Thus, I would view eliminating the flag as entirely beneficial, both in helpiing AIs when they do use Fascism and in making Fascism less unattractive to human players.

[alexman]

We increased the corruption level for Fascism, and told all AIs to shun that government. According to my testing, those two changes should have been enough to prevent the AIs from choosing Fascism over Communism when at war.

I was just saying that if you improve a government that is almost never chosen by the AI, you give an advantage to the human player. Of course, the AI still might go Fascist if it has not discovered Communism, but that situation is only temporary.

As it is now, Fascism can be a good government in very limited situations, which only the human player can recognize. If we make it a good government in more situations, the AI will more often pick the wrong government, because it will always pick Communism.

[nbarclay]

The way I look at it, making Fascism more powerful when AIs choose it is beneficial from a "helping the AIs" perspective, while making it useful a little more often for human players (who very rarely find it useful from what I've seen) is beneficial from an "interesting strategic choices" perspective.

In regard to the idea that Communism would be the wrong government more often if we make Fascism more powerful, I view that as a purely aesthetic matter that makes no practical difference. It might not look pretty to have AIs choose Communism in situations where Fascism would be better, but the reality is that when AIs are in Communism, how good or bad Fascism is is completely irrelevant to the quality of game the AIs play. The only times when making Fascism better would actually undermine the AIs' effectiveness as opponents would be when human players choose Fascism and AIs don't.

Thus, from a "helping the AIs" perspective, the real question is whether AIs would benefit more often in situations where they choose Fascism because they don't know about Communism yet or humans would benefit more from finding situations where Fascism is the best government for them. Given that I've never used Fascism even once in all the C3C games I've played, my inclination is that aside from the "Power of Fascism" game we're planning, the change would tend to benefit AIs more than it does humans.

[alexman]

Originally posted by nbarclay
Given that I've never used Fascism even once in all the C3C games I've played, my inclination is that aside from the "Power of Fascism" game we're planning, the change would tend to benefit AIs more than it does humans.

But how can it benefit the AI when they don't choose Fascism? Has anyone ever seen the AI choose Fascism in an AU mod game? It's extremely rare.

Also, just because you personally haven't used Fascism it doesn't mean that Fascism is not useful to others. Ultimate Power-style games (like you prefer to play) do not lend themselves to Fascism, which is better for smaller empires.

My feeling is that if we remove the forced resettlement flag from fascism, we will see much more widespread use of this government from human players. Remember, we have already given Fascism an extra FP.

[nbarclay]

[QUOTE] Originally posted by alexman

Also, just because you personally haven't used Fascism it doesn't mean that Fascism is not useful to others. Ultimate Power-style games (like you prefer to play) do not lend themselves to Fascism, which is better for smaller empires.

It's not just the "ultimate power" angle. There is also the fact that even when Fascism might provide a short-term advantage, it doesn't take a lot of conquest for the advantage to shift to Communism. Thus, even with religious civs, it generally makes sense to go ahead and switch to Communism. The two main exceptions I can think of offhand are (1) if a player is planning limited warfare followed by a switch to Democracy and (2) in really tough games where making any significant military headway at all is expected to take quite a while.

My feeling is that if we remove the forced resettlement flag from fascism, we will see much more widespread use of this government from human players. Remember, we have already given Fascism an extra FP.

The question to me is not how often players would use Fascism, but rather how often it would give them a significant advantage over strategies involving other governments or combinations of governments. If players choose Fascism without getting a significant net advantage from it, that makes the game more interesting without upsetting its balance. Such an outcome would be very much within the spirit and purpose of the AU Mod.

The real danger is if we would make Fascism powerful enough that players not only use it more, but get a significant net advantage from doing so. (I say "net advantage" because of situations where Fascism's short-term advantages are offset by longer-term disadvantages.) My expectation is that even with Forced Resettlement removed, situations where that would be an issue would be extremely rare. Further, the times when the risk would be greatest would be times when the human player is weakest - and therefore (in my opinion) when a boost for the human player would be least objectionable.

[nbarclay]

Since the the removal of the Forced Relocation flag has, in a sense, been left in limbo with an "under consideration" status for months, I think it would be safe to go ahead and vote on it at the same time we do the Republic unit support changes.

By the way, I don't think I thought to bring it up before, but since Democracy mirrors Republic's unit support model in the AU Mod other than being cheaper per additional unit, if we increase support for Republic, we would also presumably do so for Democracy.

[pvzh]

The reason I proposed to remove Forced resettlement, is that I doubt that AI "knows" about the effect of the flag, more likely it is pure loss/gain calculation on productivity and unit support cost. Thus, I proposed to increase free support for democracy and lowered it slightly for Fascism, so AI will be more likely to stay in Democracy during *peace*.

For the human player it is more of psychological barrier, population loss is one time affair, so, if you manage population/food production well, it should not be much of the problem in the time of railroads/electricity.

Fascism should work well if you combine it with mobilization: if you xenophobic you do not accumulate culture until you have majority of population anyway, so you may just not build any cultural improvements in conquered towns at all.

[nbarclay]

Originally posted by pvzh
The reason I proposed to remove Forced resettlement, is that I doubt that AI "knows" about the effect of the flag, more likely it is pure loss/gain calculation on productivity and unit support cost. Thus, I proposed to increase free support for democracy and lowered it slightly for Fascism, so AI will be more likely to stay in Democracy during *peace*.

Even in the stock game, I've never noticed a problem of AIs choosing Fascism over Democracy in peacetime. With changes already in the AU Mod, such an occurrence is even more unlikely.

The bigger problem was that AIs tended to choose Fascism over Communism practically all the time. Alexman made some changes with the goal that the AIs will essentially always choose Communism over Fascism if it has both (a change which is especially good for maximizing the power of killer AIs. So I don't view the AI's lack of knowledge about the effects of the Forced Resettlement flag as a significant problem in the current version of the Mod.

For the human player it is more of psychological barrier, population loss is one time affair, so, if you manage population/food production well, it should not be much of the problem in the time of railroads/electricity.

There's also another angle I haven't thought to bring up before. By the time a player will want to change governments to Fascism, chances are that a lot of cities will have reached a steady state of maximum desired size and little or no food surplus. With Forced Resettlement, a player has to either wait for slow growth to replace the lost population in those cities or go through some micromanagement hassle to speed up the cities' growth and then slow it back down once the lost population is replaced. Getting rid of Forced Resettlement would eliminate that micromanagement hassle.

I suppose a sufficiently serious zealot might argue that the micromanagement issue is a "strategic option" that we shouldn't get rid of. But I seriously doubt that many players would consider the micromanagement issue more interesting than it is annoying.

[nbarclay]

Since Theseus seems to have the next game essentially ready, let's go ahead and vote. Voting will last 24 hours, assuming sufficient votes are cast in that time. Note that we have two separate issues to vote on.

1) Yes/No: Increase free unit support for Republic and Democracy from 0/1/1 + 12 to 0/1/1 + 18. (That compares with 1/3/4 under the standard rules.)

2) Yes/No: Eliminate the Forced Resettlement flag from Fascism. (If we vote yes, civs will no longer lose pop points when switching to Fascism.)

[nbarclay]

My votes:

1) Yes
2) Yes

[alexman]

1) Yes.
2) No. (There isn't enough benefit to justify a change IMO).

[Theseus]

1) Sorry for being difficult: Republic no (leave at 0/1/1+12), Demo yes.

2) No.

[nbarclay]

I would be against boosting Democracy's free unit support above Republic's. Keep in mind that giving Democracy any free unit support at all is an improvement compared with stock rules. (In the stock rules, Republic has free unit support while Democracy doesn't, but Democracy just costs one gpt per unit while Republic costs two.) I consider it reasonable to give Democracy the same unit support as Republic in order to take away a significant potential reason not to change, but I would strongly oppose giving Democracy greater support than Republic on top of its advantage of only costing half as much per unit for upkeep above the free support limit.