A core set of balance changes

[KrikkitTwo]

Why? You keep making these assertions without any kind of argument.
The whole point of annexing is to make a short-term temporary happiness hit in order to gain long-term control.
The idea in part was to get you to puppet during a war, then gradually annex and courthouse afterwards (as your happiness allowed).

Then Puppet v. annex is not a Choice....It is always Puppet... and When will I annex (ie build a courthouse)

You might as well have the "courthouse" give additional unhappiness. and remove Puppet penalties

There is nothing broken about this system, so there is no overwhelming need to change it.
This whole thread is supposed to be making the minimal number of changes needed to get the game to play well. Don't fix what isn't badly broken, and work with the existing mechanics whenever possible.

The Definitely Broken part is
Puppets give Culture but not Social Policy costs.

If you give them Social Policy costs, you might as well eliminate Puppets
Having them not give Culture would be much more interesting.

Its not so important now because puppets are a bit too strong. Weaken puppets, and it *will* be an important choice, with short-term costs vs long-term benefit tradeoffs.

No, its not a Puppet v. Annex choice
It is a When to Annex? choice

While this is an interesting choice... it means we can practically eliminate "Annex +Puppet" from the "conquered city screen" and replace it with "Annex now v. Annex Later"

That's nice. But that's not what they are. And they don't need to be "resistance plus" for the game to work. So save that for a separate design project, rather than one focused on the minimal core changes.

Well currently they don't give Social policy Costs and that is a distinct difference from Every other type of city.

Now the options are either
1. Make them more similar to annex (no social Policy costs)... removes uniqueness
2. Make them more different (no Social policy boost from that city's culture)

I could see doing #2, and leaving Gold/Science Mostly alone (just give a % penalty to those)
But in that case, you end up with Building maintenance that you don't want (unless you go through complicated things like Celevin did). Essentially a Random penalty.

But there is no need to make them not give expenses. Buildings in puppet cities shouldn't suddenly stop costing anything.

If you can't control them, then they give a Random benefit/Penalty... Random is bad. (now the ability to Sell buildings can derandomize some of this)

Overall, The minimal change would be
1. Make Puppet culture not count for Social Policies
2. Give Puppet Science, Gold, (and GPP?) a 70% penalty (after all other penalties+bonuses)

 

[Ahriman]

Then Puppet v. annex is not a Choice....It is always Puppet... and When will I annex (ie build a courthouse)

When I annex (if ever) is a strategic decision.

You might as well have the "courthouse" give additional unhappiness. and remove Puppet penalties

I don't understand. Why should a courthouse give unhappiness? Why should a puppet have no penalties? This makes no sense. Building a courthouse makes you... worse off?

The Definitely Broken part is
Puppets give Culture but not Social Policy costs.

I'm still unsure. I think nerfing puppet yields (including culture) might be sufficient, and puppets might need to cause a increase in SP costs.
I would rather have puppets increase SP costs (eg by 1/3 as much as annexed cities, and have a 1/3 culture penalty) than to have puppet-produced culture somehow "not count" for social policy accumulation.
I would need to do some testing of different options.

If you give them Social Policy costs, you might as well eliminate Puppets

Nonsense. Puppets are a way you can capture cities with small short-term happiness consequences.
Eliminating puppets entirely would be a serious engine change, and it isn't necessary.

Having them not give Culture would be much more interesting.

It would be very confusing if your culture for social policies was not the sum of your culture income.

It is a When to Annex? choice

No it isn't. If I have a small, out of the way city that I want only for territory, I would never bother annexing it.

Give Puppet Science, Gold, (and GPP?) a 70% penalty

70%?? Crippling. No way.
Incremental changes, not massive sweeping changes.

 

[Roxlimn]

Why? REX costs you nothing, you just build settlers. War means you have to build a military and actually fight, and expend resources.

And then get cities that already have expanded territories and developed tiles. Fighting is one of the most cost-effective ways to gain land and resources right now - more than REXing. It doesn't hurt that each tile you claim under war moves you closer to the Domination Victory condition, which probably explains why it is the fastest victory condition achievable.

Celevin prefers to raze and resettle because the upkeep offends him on a personal level, but it is not necessarily more efficient to play that way, especially if you are going for speed of conquest.

War should cost, as you say it does, but currently, it is too cost-effective. Annexing is sufficiently costly that it is not always a forgone conclusion. Puppeting should be the same way.

 

[KrikkitTwo]

When I annex (if ever) is a strategic decision.

Is when you annex ever different from when you build a Courthouse?

So Annexing is not a strategic decision, building a Courthouse is the Strategic decision.

I don't understand. Why should a courthouse give unhappiness? Why should a puppet have no penalties? This makes no sense. Building a courthouse makes you... worse off?

A courthouse Currently gives unhappiness while you are building it.
Right now a conquered city goes through 4 phases
1. Resistance
2. Puppet (various penalties+bonuses)
3. Courthouse building Annex (extra unhappiness)
4. Annexed with Courthouse (maintenance)

The only time the city counts a occupied is stages 1+3... so you might as well eliminate the concept of Annexed v. Puppet

Just have
Conquered city.... a Conquered city has certain penalties (same as a Puppet)
Once you build a Courthouse those penalties are gone.

No it isn't. If I have a small, out of the way city that I want only for territory, I would never bother annexing it.

So Puppet is used either for...
1. Early stages of Annexing
or
2. Territory

Then why do you need puppets to give any income??

Just remove the Unhappiness penalty from Occupied cities and instead make it an output penalty (removed by Courthouses)
So options would be
A->Puppet: gives only Territory,and minimal costs
B->Annex: has an output penalty until a courthouse is built

I'm still unsure. I think nerfing puppet yields (including culture) might be sufficient, and puppets might need to cause a increase in SP costs.
I would rather have puppets increase SP costs (eg by 1/3 as much as annexed cities, and have a 1/3 culture penalty) than to have puppet-produced culture somehow "not count" for social policy accumulation.
I would need to do some testing of different options.

It would be very confusing if your culture for social policies was not the sum of your culture income.

It would be very confusing if some cities increased your social policies and some did not
It would be very confusing if some cities increased your city unhappiness (Puppets) and some did not (annexed... they increase Occupied City unhappiness instead)
It would be very confusing if Some cities could be controlled and some could not.
etc.

A tool tip should be enough...
"Puppets do not add to social policy costs, nor do they contribute their culture to the purchase of Social Policies"

There we go... understandable?

Perhaps when you mouse over culture
+X Culture from Non-Puppeted Cities
+Y Culture from City-States

If its still too confusing, then just have puppets unable to produce Culture at all.

Nonsense. Puppets are a way you can capture cities with small short-term happiness consequences.
Eliminating puppets entirely would be a serious engine change, and it isn't necessary.

Making puppets increase social policy costs is a serious game engine change, and eliminating them entirely would be more sensible.

70%?? Crippling. No way.
Incremental changes, not massive sweeping changes.

Crippling? you were proposing 30% right (100-70=30)?... I'd say no less than a 50% penalty to Science+Gold

and Either 0% culture output or Non-Puppet culture only counts.

Because it would be far better to eliminate puppets altogether then to Adding in a Social Policy Cost for them.

 

[Ahriman]

War should cost, as you say it does, but currently, it is too cost-effective.

Sure, but this is mostly because:
a) The military AI performs poorly enough that you can undertake conquest with minimal unit losses
b) Cities are too weak, and too easy to capture
c) Once you build an overwhelming force you can easily sweep through the enemy empire
d) Resistance periods are probably too short, and puppet yields are too high.

Fix these, rather than making sweeping changes.

Annexing is sufficiently costly that it is not always a forgone conclusion. Puppeting should be the same way.

I don't understand what you're saying here. When you capture a city, you always annex, puppet or raze it (and sometimes sell/give it to someone else).
You can't make all of these decisions separately. There is a degrees of freedom constraint here; if you determine conditions when you choose n-1 of these, then the nth condition is already determined.
Supposing that you don't want to raze or give the city away, then choosing when to annex is the SAME decision as choosing when to not keep it as a puppet. It has to be one or the other.

Is when you annex ever different from when you build a Courthouse?

Sometimes, yes, if for example I want to start buying units or buildings in the city, particularly if I'm a golden age where I don't care about surplus happy and so I want to build a few other things first. Rare though.

A courthouse Currently gives unhappiness while you are building it.
Right now a conquered city goes through 4 phases
1. Resistance
2. Puppet (various penalties+bonuses)
3. Courthouse building Annex (extra unhappiness)
4. Annexed with Courthouse (maintenance)

The only time the city counts a occupied is stages 1+3... so you might as well eliminate the concept of Annexed v. Puppet

This makes no sense to me. All 4 of those phases are different.
Cities in resistance have no ranged attack, and provide no income.
Puppets provide income, but cannot build units, do not allow you to buy things, and do not let you control it.
Annexed without courthouse has a big happiness penalty, but allows you to select what to build, to choose what to build (though its usually a courthouse), to micromanage citizen allocation, and to buy things.
Annexed with courthouse lets you control everything and buy, but has a maintenance cost.

Also, I don't think a puppet in resistance counts as occupied, but I am not sure of that.

Just have
Conquered city.... a Conquered city has certain penalties (same as a Puppet)
Once you build a Courthouse those penalties are gone.

But you lose variation here. And you unnecessarily change core gameplay mechanics. Which is not the point of this thread.

Then why do you need puppets to give any income??

Because I use puppets for territory, early stages of annexing, and income. As a reward for conquest.

Why do you need them to not give any income?

If its still too confusing, then just have puppets unable to produce Culture at all.

Or, just work WITH the current system, and give them a culture yield penalty. Ta da!

Making puppets increase social policy costs is a serious game engine change

Its a relatively easy engine change that could be made with a couple of parameter changes.
Eliminating them entirely is not sensible, and is a major engine change that would require major code changes, and graphical changes, and UI changes.
I don't think you understand what an engine change is if you don't see the difference between parameter tweaking and removing a mechanic.

The AI is tuned around an expand and puppet philosophy. If you destroy the validity of that, you're going to have to re-write the AI. I would prefer that AI attention focus on teaching it to better play the game as it is currently designed, rather than trying to teach it ways to play a different game with different mechanics.

Crippling? you were proposing 30% right (100-70=30)?

I proposed a 30% penalty as a starting point, subject to further gameplay testing. So a city that normally produced 10 science produced 7 science instead. You proposed a 70% penalty, so a city that produced 10 science normally produced 3 instead.
Those are very different.

A 70% penalty is going to crash an burn the economy of an expanding AI player.

I'd say no less than a 50% penalty to Science+Gold

This would make puppets useless, they wouldn't even cover their own gold maintenance costs + road costs.

 

[KrikkitTwo]

Or, just work WITH the current system, and give them a culture yield penalty. Ta da!

A Culture Yield Penalty of anything less than 100% is broken.
Because they have 0% Social policy costs, they should produce 0% Social policy culture

Now Ideally they would still produce "Territory Culture" but I guess that could be given up... just accept that Puppet borders won't expand.

Also the current system has no "Puppet Yield Penalty", that would have to be added... there are no parameters to be tweaked (unless they added them... and they might have a "Puppets Allowed" parameter somewhere as well)

Its a relatively easy engine change that could be made with a couple of parameter changes.

see above

I proposed a 30% penalty as a starting point, subject to further gameplay testing. So a city that normally produced 10 science produced 7 science instead. You proposed a 70% penalty, so a city that produced 10 science normally produced 3 instead.
Those are very different.

Sorry thought you were trying to put in a serious penalty for them.

A 70% penalty is going to crash an burn the economy of an expanding AI player.

This would make puppets useless, they wouldn't even cover their own gold maintenance costs + road costs.

Well the penalty Probably wouldn't apply to Trade Route income, and If the Puppet Focuses on Gold (as expected to), then they would almost certainly cover their own expenses. (once sufficient TPs were in)

I guess a slight (30%) penalty to science + gold might not be too bad, especially if they have a 100% penalty to Culture (both Tile and Policy)

 

[Celevin]

Just a comment in the ongoing argument: You can't take away culture from puppets. You think people are complaining now, when puppets build barracks? You're going to make half of the buildings available worthless.

Puppets are already a game of dice with sweeping results depending on what they build. I've had empires change dramatically because all my puppets decided to go happiness buildings, and have had empires go into a huge slump because all my puppets decided to go military buildings.

This is more than just a maintenance cost. This is a huge huge opportunity cost.

 

[KrikkitTwo]

Just a comment in the ongoing argument: You can't take away culture from puppets. You think people are complaining now, when puppets build barracks? You're going to make half of the buildings available worthless.

Puppets are already a game of dice with sweeping results depending on what they build. I've had empires change dramatically because all my puppets decided to go happiness buildings, and have had empires go into a huge slump because all my puppets decided to go military buildings.

This is more than just a maintenance cost. This is a huge huge opportunity cost.

Well either make puppets not build culture buildings Either.... Or for that matter make Puppets Never build any buildngs at all... Puppets build Wealth, and Wealth only. (maybe they build Workers if Wealth can't be built.)

Since all Culture buildings are destroyed on Capture, Puppet will never produce Culture, and the "Puppet Penalty" is the fact that they are undeveloped. (are Culture Buildings destroyed when a city is sold?)

That removes the Randomness (except for the Randomness of which buildings the Puppet initially has)... and it makes it a simple Puppet AI change, the Puppet AI focuses on Gold and only builds Wealth or Workers.


There is still the problem with France/Representation allowing Puppets to get Culture. But in that case, then just add a 0% culture penalty.

So changes
1. Puppet AI: only Workers or Wealth. (puppets build no Buildings, Wonders or Military units)
2. Puppets produce 0 culture, even if they are French, have Representation, World Wonders, or Landmarks, or somehow got a cultural building.

Otherwise they act as they do currently (normal city except uncontrolled and no social policy costs)


[Possibly change it so they don't count in figuring National Wonders]
And then possibly add a science/gold penalty, if the "undeveloped penalty" isn't enough.

 

[Celevin]

I'm going to change the discussion. It's gotten a bit stale, and I think Ahriman can think of a final decision for puppets/annexing that is fair and takes in the recent posts.

I have some suggestions that I've been toying with to make smaller empires more feasible. A big portion of the problem is how much big empires are aided by policies, but even the trees devoted to smaller empires seem a bit weak. Tradition for example is too capitol-centric, meaning it doesn't scale with map sizes, and it doesn't aid your other big cities. Freedom is completely hindered in that big empires actually have more specialist slots.

Revamp Tradition to include more than the capitol
- Change "+33% food surplus in the capitol" to "+20% food surplus in each city". Small cities won't care too much about this.
- Change "-33% unhappiness from the capitol" to "+2 happiness from each city with a wonder". National Wonders will make this really attractive (which smaller empires build more of), and it has synergy with the rest of the tree.

Revamp the Freedom Tree
- Change "+1/2 happiness per specialist" to "Merchants also add +5% gold, engineers add +5% hammers, artists add +5% culture, scientists add +5% science of the city's output". This will make specialists better as the city grows. Happiness gained through specialists is already too powerful, especially for ICS empires.
- Change "-25% cost to all future policies" to "+25% to empire culture". This will remove heavy culture slingshots that dominate culture wins.


If I had a lot of free time on my hands, I'd remove all unlocking of all policy trees by era and change it so all 10 can be unlocked at the start. Rebalance later policies so they are at the same power level as earlier ones. Give each later tree more of a "theme", and not a "requirement for this victory". It would make for a much more sculpted empire from the start, and add more strategy.

 

[Roxlimn]

Here's something that will nerf big empires to smithereens:

Each city after the 5th adds +50% to all technology research costs. This includes puppets.

 

[Rashkaar]

The entire point of puppets, as far as I can see it, is to allow you to conquer in one stage, (re)build in another, while ensuring balance vs less aggressive styles of play, and I believe the *intent* of designers was that you would annex most cities soon after the war ends.

For this reason, I would actually reduce unhappiness caused by puppets to half of population (if garrisoned, perhaps) or more, but significantly, reduce production and other output significantly as well (perhaps 25%), and add some temporary penalties to empire-wide culture production per puppet (each puppet = 5% or whatever).

You could also create penalties based on # of turns a city has been puppet (obviously with a cap).

Annexing should be the obvious long-term choice 9 times out of 10, while puppeting several cities for the short term being a viable option while at war.

You could also go for other routes that would also limit ICS. One thing that comes to mind is limiting certain buildings by population, or benefits scaling by population. As an example, happiness buildings will never provide more than 50% of pop in happiness, e.g. full gain of colluseum requires 8pop), and/or colluseum can not be built until size 6/8/10/whatever.

 

[Celevin]

You could also go for other routes that would also limit ICS. One thing that comes to mind is limiting certain buildings by population, or benefits scaling by population. As an example, happiness buildings will never provide more than 50% of pop in happiness, e.g. full gain of colluseum requires 8pop), and/or colluseum can not be built until size 6/8/10/whatever.

This here is extremely interesting. It solves the problem in pretty much the most blunt way possible. I wonder if it would be enough, though. Often even in an ICS most of my cities are size 11. That's enough to cover a colloseum, circus, and theatre! Plus India might interact weirdly.

I would consider halving all of the happiness bonuses from buildings, and putting in a component "x% of gold + 2x% of production in additional happiness", or something along those lines. (Those numbers chosen because gold is usually valued at 1/2 production). It would make players want to specialize cities again, which from what I've seen is a lot of the complaining.

 

[KrikkitTwo]

Here's something that will nerf big empires to smithereens:

Each city after the 5th adds +50% to all technology research costs. This includes puppets.

Well, You could make it more advanced, and more useful without nerfing large empires to smithereens

Technology costs are incresaed by +10%*(number of cities you have - number of cities of all the civs you know that have the tech) (with it never going less than 0)

 

[Slowpoke]

Well I've finished mega mod (in sig), which covers what I believe are the main balance changes needed. It's a work in progress, but it contains a lot so far. Criticisms more than welcome.

 

[Khaim]

Regarding the puppet discussion, it seems like Krikkitone just doesn't like the idea of puppets. While that's fine, I think most of what he's suggesting goes beyond the small tweaks Ahriman is looking for. With the recent patch, puppets don't build worthless or resource-consuming buildings and have a large bias towards gold. This sort of fixes the research issue: while there's no actual penalty, the research output is much lower than before. The amount of gold generated might be excessive and need a penalty of some sort, but we'll have to play with it some to determine that.

The culture problem is still present. I like the idea of puppets increasing policy costs by some fraction of a normal city, and I think that would be sufficient.

Right now, an empire that has a large proportion of puppets gains social policies faster than an identical empire where all cities are annexed/settled. I do not think this is a mistake. There is a valid argument that puppets are too good at producing culture, but I maintain that a puppet should be "better" than a non-puppet.

- Change "+33% food surplus in the capitol" to "+20% food surplus in each city". Small cities won't care too much about this.

It seems like this benefits all cities equally. Yes, the raw food bonus is larger for large cities, but so is the food bucket. You could rewrite this as "-16% food required for growth" and it would be functionally identical.

- Change "-33% unhappiness from the capitol" to "+2 happiness from each city with a wonder". National Wonders will make this really attractive (which smaller empires build more of), and it has synergy with the rest of the tree.

Love this idea. (Palace counts as a national wonder, right?) If anything, +2 is rather low: compare to Theocracy (~20%) and the original 33%. I think 20-25% of total or population would be good.

- Change "+1/2 happiness per specialist" to "Merchants also add +5% gold, engineers add +5% hammers, artists add +5% culture, scientists add +5% science of the city's output". This will make specialists better as the city grows. Happiness gained through specialists is already too powerful, especially for ICS empires.

I like the idea, but a 5% bonus per specialist is way too good. What if it was a 5% bonus regardless of the number of specialists? In other words, you get +5% gold for having a merchant, but the second merchant only adds the base +2 gold.

- Change "-25% cost to all future policies" to "+25% to empire culture". This will remove heavy culture slingshots that dominate culture wins.

The equivalent bonus would actually be +33%. Good idea though.

If I had a lot of free time on my hands, I'd remove all unlocking of all policy trees by era and change it so all 10 can be unlocked at the start. Rebalance later policies so they are at the same power level as earlier ones. Give each later tree more of a "theme", and not a "requirement for this victory". It would make for a much more sculpted empire from the start, and add more strategy.

This I don't like. The gradual unlocking of the policy trees is part of the era progression, and I think it should remain this way. I've heard good arguments for moving the unlock condition to specific techs rather than just era, but that's a relatively minor change. If we were to do that, I would suggest:

• Piety - Philosophy
• Patronage - Theology
• Commerce - Currency
• Rationalism - Astronomy
• Freedom - Printing Press
• Autocracy - Steam Power
• Order - Biology

All of these are minimum-cost techs for their era (with the exception of Philosophy; Optics is cheaper, 80 to 100, but doesn't fit at all). I realize that Biology doesn't fit well, but Steam Power is the only other valid tech.

 

[Celevin]

It seems like this benefits all cities equally. Yes, the raw food bonus is larger for large cities, but so is the food bucket. You could rewrite this as "-16% food required for growth" and it would be functionally identical.

Yeah, true. The idea is that a large empire with small cities doesn't really care about how fast until their small cities grow, they care more about having the happiness to supply them. Big empires are almost always happiness capped so inflating surplus food doesn't do much. Small empires on the other hand have spare happiness, and growth is hard, so inflating surplus food is good.

Love this idea. (Palace counts as a national wonder, right?) If anything, +2 is rather low: compare to Theocracy (~20%) and the original 33%. I think 20-25% of total or population would be good.

I just tried to go in a different direction than the Piety one. Makes it more flavourful. Besides, if it's the same as Piety, people will stack them, and it might lead to a greatly unbalanced scenario. I don't want them to synergize or be worse when paired, otherwise it will influence people to either avoid or grab Piety when they have Tradition. I think if it's a completely separate way of getting the happiness, it might be better.

It also greatly synergizes with "+33% wonder construction" and National Wonders (which small empires have in spades).

I like the idea, but a 5% bonus per specialist is way too good. What if it was a 5% bonus regardless of the number of specialists? In other words, you get +5% gold for having a merchant, but the second merchant only adds the base +2 gold.

Probably overpowered looking back on it. I thought to myself "you'd need 5 merchants just to get a market", but who knows, multipliers can get out of hand quickly in games. Your idea works but I feel like it's won't stop the problem...

My goal was that "2 merchants in a large city is better than 2 merchants in 2 separate small cities", and it accomplishes it. I don't think your fix does. Maybe we can come up with something else that works?

The equivalent bonus would actually be +33%. Good idea though.

Oops! Messed that up . I'm rather ashamed at myself!

This I don't like. The gradual unlocking of the policy trees is part of the era progression, and I think it should remain this way. I've heard good arguments for moving the unlock condition to specific techs rather than just era, but that's a relatively minor change. If we were to do that, I would suggest:

• Piety - Philosophy
• Patronage - Theology
• Commerce - Currency
• Rationalism - Astronomy
• Freedom - Printing Press
• Autocracy - Steam Power
• Order - Biology

All of these are minimum-cost techs for their era (with the exception of Philosophy; Optics is cheaper, 80 to 100, but doesn't fit at all). I realize that Biology doesn't fit well, but Steam Power is the only other valid tech.

It was more of a personal want, but something I'd never try and implement nor force on others. I dislike things like corporations where you need to wait for most of the game to unlock. I also dislike how a great number of strategies involve saving policies. Policies are meant to be used and enjoyed, not banked and rotting until the Industrial Age!

If I had to have a system where some policies are unlocked later, I'd rather have it unlocked by other policies than science. For example, you get your first 5/10 trees. For you to gain access to the other 5 trees, you must research 10 individual policies previously. This will keep people from saving up all their policies until later on.

(On that note, I also hate hate hate the Cristo Redentor. When I save up 3/4s of my policies until that thing is built, then unlock them all at once and win the game, something feels wrong. Just let me unlock them as I go and don't give me incentives otherwise)


I also have a huge huge rant for the ending of era progressions - I dislike how much the game changes with one tech unlocked. All of a sudden your city states have better units, your Maritime bonuses just went completely off the wall, all those policies you were saving up get spent immediately... I am in support of ending everything that changes with an era switch. I dislike it for the same reason that I dislike unhappiness just going berserk past -10. Don't take this as something I'm trying to suggest for Ahriman's list though. It's got too much personal feelings mixed in compared to what's good for the average civ player.

 

[Khaim]

Probably overpowered looking back on it. I thought to myself "you'd need 5 merchants just to get a market", but who knows, multipliers can get out of hand quickly in games. Your idea works but I feel like it's won't stop the problem...

My goal was that "2 merchants in a large city is better than 2 merchants in 2 separate small cities", and it accomplishes it. I don't think your fix does. Maybe we can come up with something else that works?

What if the 5% bonus replaces their normal bonus? That actually makes it worse for small cities; below a certain base level, you're actually losing output. Going this route might require different bonuses for different specialists. In particular, Artists need a base cpt of 20 to break even at 5%, so they might need to be 10% or even 20%. We should set it so that the break-even is significant enough to discourage use in ICS, but low enough that a large city can get there without focusing. You would see the best results in large, focused cities that have a base well above the break-even point.

Some numbers off the top of my head:
Take a size 10 city. Assume it has a library, market, monument, temple, and workshop. (Also assume the workshop no longer sucks.) On average, its tiles give 1, 1.5, and enough food to avoid starvation. If everyone is on a tile, we get a base output of 16.5, 15, 5 and 11. With normal specialists, we lose -1, -1.9 and some food and get:

Merchant: 2.5
Scientist: 3
Artist: 1
Engineer: 2

Notice that this is a fairly underdeveloped city, and could easily get higher output in one area by better tile use or using better buildings. We want this city to at least break even by switching to Freedom. This gives us the following minimums:
Merchant: 15%
Scientist: 20%
Artist: 20%
Engineer: 20%
This suggests 20% for all types. At first glance this seems ridiculous, but since it's replacing their flat bonuses instead of augmenting them the total effect is smaller. Late-game this can provide a large bonus, but no more than a modern-era building.