Building Wealth vs Building Science

[Zechnophobe]

As usual, I realize I'm quite possibly very late on the scene here, but part of why I write these obvious things down is to get it out of my head and hopefully have tell people what else I can do better.

So, in Civ4 the game rounds down fractions. These rounded down fractions lead you to getting less than you possibly could, and a bit of work can help you get around the problem.

Let's consider a 112 commerce city running at 80% science and 20% income.

This gives us:

.8 * 112 = 89.6 = 89 science
.2 * 112 = 22.4 = 22 income.

So just in the initial conversion we already lose some.

Then we may have a library here:

.5 * 89 = 44.5 = 44 science bonus

So by rounding twice we end up losing a few points of GNP. This is substantially more noticeable on smaller cities where 1 or 2 GNP represents a greater percentage of that cities full yield.

So how can we get that back? The obvious way is to simply set the slider at 100 all the time. Some people propose straight up binary research, where you are either 100% science, or 100% income. That can work, but has the problem of requiring both types of infrastructure (libraries and banks) for you to get the % bonus' on the base GNP the entire time. While you are in 100% science, you have academies, libraries, universities etc giving you bonus, while in 100% income, you likely have far less buildings helping out.

So it is then preferable to 'build wealth' in high production cities. They often build themselves until they don't need any more buildings, and being able to keep the empire afloat at 100% science allows for the greatest efficiency.

And this is the crux of the issue: By judiciously building wealth you also save on another cost: The opportunity cost of building those markets and grocers (You likely want 6 banks for wall street anyway). 'build wealth' and 'build science' both ignore buildings like markets and libraries, so when balancing out the empire, it makes most sense to build the one you have the least infrastructure for.

You can also use it to keep hold of an overflow bonus from previous production for a long period of time. Say if you want to keep it until the tech for that wonder finishes.


Anyhow, probably water under the bridge, but there you have it.

 

[King Pluto]

It seems like the tl;dr version would be: "Build wealth, not science, so that you can make better use of your powerhouse cities with their beaker multipliers."

Which I believe is the commonly accepted answer, yes.

 

[coanda]

As usual, I realize I'm quite possibly very late on the scene here, but part of why I write these obvious things down is to get it out of my head and hopefully have tell people what else I can do better.

So, in Civ4 the game rounds down fractions. These rounded down fractions lead you to getting less than you possibly could, and a bit of work can help you get around the problem.

Let's consider a 112 commerce city running at 80% science and 20% income.

This gives us:

.8 * 112 = 89.6 = 89 science
.2 * 112 = 22.4 = 22 income.

So just in the initial conversion we already lose some.

Then we may have a library here:

.5 * 89 = 44.5 = 44 science bonus

So by rounding twice we end up losing a few points of GNP. This is substantially more noticeable on smaller cities where 1 or 2 GNP represents a greater percentage of that cities full yield.

So how can we get that back? The obvious way is to simply set the slider at 100 all the time. Some people propose straight up binary research, where you are either 100% science, or 100% income. That can work, but has the problem of requiring both types of infrastructure (libraries and banks) for you to get the % bonus' on the base GNP the entire time. While you are in 100% science, you have academies, libraries, universities etc giving you bonus, while in 100% income, you likely have far less buildings helping out.

So it is then preferable to 'build wealth' in high production cities. They often build themselves until they don't need any more buildings, and being able to keep the empire afloat at 100% science allows for the greatest efficiency.

And this is the crux of the issue: By judiciously building wealth you also save on another cost: The opportunity cost of building those markets and grocers (You likely want 6 banks for wall street anyway). 'build wealth' and 'build science' both ignore buildings like markets and libraries, so when balancing out the empire, it makes most sense to build the one you have the least infrastructure for.

You can also use it to keep hold of an overflow bonus from previous production for a long period of time. Say if you want to keep it until the tech for that wonder finishes.


Anyhow, probably water under the bridge, but there you have it.

A few points...

1. The game rounds down empire-wide, not city-by-city. And it does this after library multipliers etc. have been applied. So you can't lose more than 1 beaker due to rounding (I believe this is a change from Civ3). This quickly become negligible (after the first ~50 turns or so).

2. Simply running at 100% research doesn't guarantee no losses due to rounding. If you have just one city making 10 commerce, and it builds a library, at 100% research you get 12.5 beakers, which rounds down to 12.

3. Binary research is no more infrastructure-intensive than regular research. If you have a library, and spend 10 turns at 80% research with 100 commerce, you get 1000 beakers and 200 wealth. If you spend 8 turns at 100% research and 2 turns at 0% research, you get 1000 beakers and 200 wealth. This ties in to your later point, though - if you're building wealth to run a high science slider, you're right that it will let you skip out on wealth-multiplying buildings and focus on research-multipliers.

4. I actually generally find I have important stuff to do with my high-production cities (build an army or chase wonders); after I've whipped a little basic infrastructure it often is the commerce cities that have spare hammers to build wealth.

5. If your sole goal is to maximize research in the short-term (not focusing on long-term things like infrastructure savings), you should bump your science slider up 10% and see how many beakers you gain and how much wealth you lose. If you gain more beakers than you lose wealth, you should build wealth. If you lose more beakers than you gain wealth, you should build research. Most of the time, you'll end up building wealth.

 

[Um the Muse]

I believe your conclusion about getting to 100% science is right, but for the wrong reason. As Coanda pointed out, the rounding is too late in the process to be a problem. The real reason is that there's a natural bonus to specialization.

Run a few numbers or tests, and you'll see that you'll always be rewarded if a city that can more efficiently produce science than other cities does that (almost) exlusively. Btw, the almost is for anything that can't be pooled with the rest of the empire's yields. For example, there's no way for one city to give another food directly, so obviously the city has to produce its own food.

These gains aren't rounding errors but gains from specialization and trade. Note that it does work even if one city is better than another at every thing (but it's much better at one and only a little better at another), but it works even better if both cities are specialized in different things.

One thing that you missed, and is fairly relevant, is that there are other reasons to build markets and grocers. Setting aside the and bonuses (which are actually the deciding factor for me), the market opens up two merchant slots and the market+grocer makes gold production more efficient from commerce or gold sources (shrines, merchants, settled GP) than from building wealth.

Unfortunately, without Rep and if you ignore the GPP, merchants don't beat building wealth from working mines until banks, which is the multiplier building I consider optional (building wallstreet is actually fairly situational). Still, if can't get enough , they're at least something to consider.

 

[ChanceDanger]

There's a few other factors that could play into this as well. First, you can unlock the gold multipliers earlier than you can beaker multipliers (Education is very expensive and Philosophy and paper aren't exactly lighting the world on fire in terms of killer techs).

Of course, you go that way to get tech goodies like liberalism or the free GP techs, but one could make the case that guilds/banking is superior if you plan on war with rifles or muskets and cannons. You spend the beakers that would've gone towards paper/education/philosophy (which is something like 4000 beakers) and I'm pretty sure in the same time frame you can make the guilds/banking/gunpowder plus some into chemistry or replaceable parts, you'd probably already be researching a good military tech at the same time you would have "won" liberalism. You do have to build more expensive infrastructure during this period, but since you don't have to put them anywhere near your hammer areas (Wall Street isn't available immediately, no need to rush out 6 banks) you can continue building an army at leisure. And of course, now you should also have the means to upgrade that army and don't have to go into draft mode and mess with your economy.

Somewhat hypothetical and situation specific, but I'd suggest you try a game where you stay away from the education trap. You might be surprised how little you'll miss the research bonus.

 

[Zechnophobe]

There's a few other factors that could play into this as well. First, you can unlock the gold multipliers earlier than you can beaker multipliers (Education is very expensive and Philosophy and paper aren't exactly lighting the world on fire in terms of killer techs).

Of course, you go that way to get tech goodies like liberalism or the free GP techs, but one could make the case that guilds/banking is superior if you plan on war with rifles or muskets and cannons. You spend the beakers that would've gone towards paper/education/philosophy (which is something like 4000 beakers) and I'm pretty sure in the same time frame you can make the guilds/banking/gunpowder plus some into chemistry or replaceable parts, you'd probably already be researching a good military tech at the same time you would have "won" liberalism. You do have to build more expensive infrastructure during this period, but since you don't have to put them anywhere near your hammer areas (Wall Street isn't available immediately, no need to rush out 6 banks) you can continue building an army at leisure. And of course, now you should also have the means to upgrade that army and don't have to go into draft mode and mess with your economy.

Somewhat hypothetical and situation specific, but I'd suggest you try a game where you stay away from the education trap. You might be surprised how little you'll miss the research bonus.

Library is earlier than marketplace, and universtiry and observatory are often much earlier than Grocer. And heck, once education, you also can immediately go for Oxford, whereas you can't get Wallstreet until Corporation!

I'd say the science ones are all pretty much earlier. Maybe if you are byzantium and want a cataphract rush...?

 

[jwez11007]

I'ld suggest reconsidering chancedanger's advice. I know its the chic thing to chase lib and is the preferred winning strat, you'ld be surprised how easy it is to win without running down the economic line for tech trades. Of course copious warfare is then a necessity and some people aren't fans of that approach. I've recently become a fan of the warmongering tech path, after playing FFH2. Its worked surprisingly well up to emperor and is a refreshing break from the diplo hassles and bulb timing required for the research path...Finally, a quote from good ole Freddy "Don't forget your great guns, which are the most respectable arguments of the rights of kings. " Stop building universities and start building some 'enforcers'

 

[Lalabunny]

There's a few other factors that could play into this as well. First, you can unlock the gold multipliers earlier than you can beaker multipliers (Education is very expensive and Philosophy and paper aren't exactly lighting the world on fire in terms of killer techs).

Of course, you go that way to get tech goodies like liberalism or the free GP techs, but one could make the case that guilds/banking is superior if you plan on war with rifles or muskets and cannons. You spend the beakers that would've gone towards paper/education/philosophy (which is something like 4000 beakers) and I'm pretty sure in the same time frame you can make the guilds/banking/gunpowder plus some into chemistry or replaceable parts, you'd probably already be researching a good military tech at the same time you would have "won" liberalism. You do have to build more expensive infrastructure during this period, but since you don't have to put them anywhere near your hammer areas (Wall Street isn't available immediately, no need to rush out 6 banks) you can continue building an army at leisure. And of course, now you should also have the means to upgrade that army and don't have to go into draft mode and mess with your economy.

Somewhat hypothetical and situation specific, but I'd suggest you try a game where you stay away from the education trap. You might be surprised how little you'll miss the research bonus.

There are alot of reason why Education/Liberlism line is prefered.

1) Although you can "unlock" gold multiplier early, they are not usually worth the hammer. At that stage of the game, you will want your research slider to be as high as possible. Building Wealth or Courthouses will give you more gold then a Market. And both Grocer/market do not have any Leader Traits that reduce their build time (Philo reduce University cost, Cre reduce Library cost).

2) The AI favors researching the Banking line heavily. At higher level, they will beat you to tech like Economic/Gunpowder/Replacable part. If you are researching along the Banking line with them, you will not have any trading bait to offer them.

3) Talking about trade bait, if you have a good trade partner, both Philosophy and Education, although very costly in terms of beakers, are excellent trade bait. If you are first to Philosophy (usually bulb), you can trade it to backfill everything you missed out. Once you have secure Liberlism, Education can be traded away too.

The free tech you get from Liberalism (Nationalism/Steel/Astronomy) can also be traded for the entire Banking line techs. By letting the AI research the Banking line while you research something else, you are killing two birds with one stone. This has a huge snowball effect.

Also when talking about Liberlism, you are not only just getting a free tech, you are also DENYING the AI of one. That is some not to be lightly discounted.

4) AI tends to shun away from the Education tech (for some reason). If you can get Oxford up fast, you can still get ahead in tech even with fewer cities and use that advantage to leverage a few techs from the AI.

5) In terms of military wise, getting Nationalism/Military Tradition/Steel from Liberalism will allow you to steamroll AI with cannons/cuirs while they are still stuck with longbow/maceman. Not to mention Education will enable Gunpower too.

6) I find that the only reason why you would want to go on the Banking line is that you are planning for a Medival war where techs like Machinery/Guild/Engineering are a priority.

 

[Htadus]

The simple fact is if you are building either wealth or flasks, the only multiplier that help you is the forge or other hammer multiplier buildings.

Given that, if your science is at 100% and not losing gold, then you should be building science.
If you are losing gold, then build wealth to allow you to work at a higher science rate. The benefit of this is that the cities with many cottages and has libs, unis and other multipliers give more output.

Now if you are using the wealth slider, then you are really going to need to build both science and wealth multipliers in your cottages cities. The alternative is to have cities with many food tiles to build wealth multipliers (real commerce city) and run merchents and perhaps priests. This way you get the benefit of running high science slider to get the most from your science cities since these commerce cities will make up the deficit with the commerce multipliers.

Most players do not know how to build a commerce city with science slider near 100%. They inevitably work cottages in those cities and unwittingly create another science city. It is the merchants and food that make real commerce cities.

Of course if you are runing Rep, then these commerce cities could use lib and a uni too.

For example a commerce city with all wealth multipliers running 4 merchants will generate 24 (4*3*2) gpt and 12 (4*3) flasks. Lib and a uni will increase this to 18 flasks. However this need to be evaluated based on how many turns may be left in the game.

 

[Um the Muse]

Htadus: When you say commerce city, you mean wealth city, right? Depending on your empire, you might not even want a city dedicated to wealth. I usually use cities that don't have anything else pressing to build accumulate wealth. A more common strategy is to build a forge in a cottage city, then have it build wealth. This second method takes advantage of specialization just as well as having them in seperate cities, with the added benefit that they can actually build something if they have to (like a uni). A forge gives you just as much a wealth advantage as a market, and you won't have any grocers for a long time.

 

[ChanceDanger]

There are alot of reason why Education/Liberlism line is prefered.

1) Although you can "unlock" gold multiplier early, they are not usually worth the hammer. At that stage of the game, you will want your research slider to be as high as possible. Building Wealth or Courthouses will give you more gold then a Market. And both Grocer/market do not have any Leader Traits that reduce their build time (Philo reduce University cost, Cre reduce Library cost).

2) The AI favors researching the Banking line heavily. At higher level, they will beat you to tech like Economic/Gunpowder/Replacable part. If you are researching along the Banking line with them, you will not have any trading bait to offer them.

3) Talking about trade bait, if you have a good trade partner, both Philosophy and Education, although very costly in terms of beakers, are excellent trade bait. If you are first to Philosophy (usually bulb), you can trade it to backfill everything you missed out. Once you have secure Liberlism, Education can be traded away too.

The free tech you get from Liberalism (Nationalism/Steel/Astronomy) can also be traded for the entire Banking line techs. By letting the AI research the Banking line while you research something else, you are killing two birds with one stone. This has a huge snowball effect.

Also when talking about Liberlism, you are not only just getting a free tech, you are also DENYING the AI of one. That is some not to be lightly discounted.

4) AI tends to shun away from the Education tech (for some reason). If you can get Oxford up fast, you can still get ahead in tech even with fewer cities and use that advantage to leverage a few techs from the AI.

5) In terms of military wise, getting Nationalism/Military Tradition/Steel from Liberalism will allow you to steamroll AI with cannons/cuirs while they are still stuck with longbow/maceman. Not to mention Education will enable Gunpower too.

6) I find that the only reason why you would want to go on the Banking line is that you are planning for a Medival war where techs like Machinery/Guild/Engineering are a priority.

All good points, although it is frustrating that every game and every strategy seems to be put up to a deity standard. I will now say that this technique can work quite effectively in Monarch (and apparently Emperor), though even if you'd don't have a tech advantage this approach would still be useful.

Also, just for comparison: Assuming you slow tech (no bulbs, let's assume you got MoM and are trying to GA through those boring middle ages or you're settling or you got terrible GP luck or something), and assuming that you have feudalism before philosophy (Sometimes you want city defense), then here are the up front costs in beakers:

- 2400 for Paper/Education
-1700 for guilds/banking.

That puts you 700 beakers ahead of the schmuck with all the fancy book learning on gunpowder (except you get to education first and you're not going to choose gunpowder). Assuming you can't get the universities in immediately (and even if you can, it's only 25% until Oxford), then the next choice comes:

-Another 2200 beakers into Philosophy and Liberalism
or
-the final 500 beakers to gunpowder, then that final 1700? you're 100 beakers away from finishing either Chemistry or Replaceable parts, one technology away from breaking open the game.

Note: I got those figures from the civilpedia, but I'm pretty sure the prices are higher in an actual game. also, I'll just go right ahead and admit that it's a pretty silly example, but I still think that you don't need to win a straight monopoly on guilds/banking to be effective, whereas obviously if you don't get a monopoly on liberalism it's useless, and if another AI wins education, depending on personality (I'm looking at you Mansa), they might just trade it away to stick it to you.

So maybe the education boost would push you closer to catching up, and you could grab an expensive tech with liberalism and then push towards war (I keep going towards this tangent because the era is about the last time you're going to decide to go to war if you haven't already). One common Liberalism route when it's actually a race is usually to pick up nationalism and get draft an army, or pair it with slavery. My thought is that a guilds/banking allows you to build a "modern" army without having to throw population away. Plus, you have promoted troops against drafted scrubs, and if you don't think that is a difference, ask anyone who's romped with CRIII riflemen.

Also, the best thing about gold is that you're not giving up a technology for whatever you need from a Civ. I'd much rather give 1000 gold and one tech for a technology then 2 techs for one. I'd much rather get trade embargoes and wars started without having to pay in technology, and those opportunities come a little more often than you think. I'd much rather have other civ's civics changed used gold instead of techs.

So I guess my final point is that I think gold is good .

 

[Lalabunny]

All good points, although it is frustrating that every game and every strategy seems to be put up to a deity standard. I will now say that this technique can work quite effectively in Monarch (and apparently Emperor), though even if you'd don't have a tech advantage this approach would still be useful.

Also, just for comparison: Assuming you slow tech (no bulbs, let's assume you got MoM and are trying to GA through those boring middle ages or you're settling or you got terrible GP luck or something), and assuming that you have feudalism before philosophy (Sometimes you want city defense), then here are the up front costs in beakers:

- 2400 for Paper/Education
-1700 for guilds/banking.

That puts you 700 beakers ahead of the schmuck with all the fancy book learning on gunpowder (except you get to education first and you're not going to choose gunpowder). Assuming you can't get the universities in immediately (and even if you can, it's only 25% until Oxford), then the next choice comes:

-Another 2200 beakers into Philosophy and Liberalism
or
-the final 500 beakers to gunpowder, then that final 1700? you're 100 beakers away from finishing either Chemistry or Replaceable parts, one technology away from breaking open the game.

Note: I got those figures from the civilpedia, but I'm pretty sure the prices are higher in an actual game. also, I'll just go right ahead and admit that it's a pretty silly example, but I still think that you don't need to win a straight monopoly on guilds/banking to be effective, whereas obviously if you don't get a monopoly on liberalism it's useless, and if another AI wins education, depending on personality (I'm looking at you Mansa), they might just trade it away to stick it to you.

So maybe the education boost would push you closer to catching up, and you could grab an expensive tech with liberalism and then push towards war (I keep going towards this tangent because the era is about the last time you're going to decide to go to war if you haven't already). One common Liberalism route when it's actually a race is usually to pick up nationalism and get draft an army, or pair it with slavery. My thought is that a guilds/banking allows you to build a "modern" army without having to throw population away. Plus, you have promoted troops against drafted scrubs, and if you don't think that is a difference, ask anyone who's romped with CRIII riflemen.

Also, the best thing about gold is that you're not giving up a technology for whatever you need from a Civ. I'd much rather give 1000 gold and one tech for a technology then 2 techs for one. I'd much rather get trade embargoes and wars started without having to pay in technology, and those opportunities come a little more often than you think. I'd much rather have other civ's civics changed used gold instead of techs.

So I guess my final point is that I think gold is good .

Just so you know, I am still stuck on Monarch level, winning a couple of Emperor once in a blue moon

Your calculation are valid. However in actual game, all these number crunching are very subjective depending on the difficulity.

On a Noble level, a Superscience/Buera Capitol (and later, Oxford) can provide enough flask for you to stay ahead in tech even with fewer cities compare to the AI. You can probably win the Liberalism Race AND still be the first to research Economic. Bulbing Philosophy with a GS gives you no advantages because AI has no tech that you want to trade for. You would rather settle the GS or put beakers into Machinary/Guild which gives you military and Infrastructure advantage. In this type of games, your research would be far ahead of AI that tech trading hold almost no value.

On a higher level however, AI recieve such ridiculous bonus that you need to broker monopoly tech very smartly in order stay on par with them. That means not researching Techs that AI prioritize alot. REXing early to get 6 decent cities before AI takes up all the land can crash your economy and hamper your research greatly. However a well timed GS bulbing Philosophy will put you back in the game. You might spend a GS or 800 beakers, but you will be able to trade for techs that are worth more then 2000 beakers in return.

If those beakers were instead used on say Machinery (700 beakers and on the tech path to Guild/Banking), all you have gained is, well Machinery. You might be able to trade it for some backward techs but at that point, most AI will already have Machinery and it will have very little of trade value. And if you are not planning for a Maceman war, which is why you would want to prioritize Machinery in the first place, what value is Machinery worth? Once Lib race is obtained, you can trade for Machinery/Engineering/Banking/Guild with all the goodie techs that you have researched. And you will still be first to Replaceble part/Rifling. Not to mention Paper is required for Printing Press (not a bad choice for Free tech from Lib) which you need for Replaceble parts.

And with Paper, if you have put in the effort to explore, trading world map for gold can boost your research further. Especially if you find anotehr continent. The true advantage of Nationhood is speed. You do not need to spend 10-20 turns building up your army. You can draft 15 Rifleman/mustketman in the course of 5 turns from 3 low-production cities. Nothing on the Banking Tech line can match that.

On the side note, the only time you will have "terrible GP luck" is when your capital is filled with wonders creating alot of different kind of GPPs. Even if thats the case, wonder should be planned which city they should go. You will be able to get 2 GS, unquestioned, from a city running 2 scientist specialist relatively early in the game.

Gold is good, but in mid games where a 10/20% slider in wealth is not worth the hammer investing in Market/Grocer/Bank, definitely not worth investing beaker in Guild/Banking just for the gold. Land are still improving, Cottages are still growing, Water/Windmill are still not at their best. If you truely need gold at that point, the best way is to trade Tech or manipulate your GPP for a GM (trade Mission). However during late games, when a 10% increase in wealth slider could mean a few hundred in gold, those gold multiplier building starts to get more appealing.

 

[ChanceDanger]

I was sort of implying though that once you got close to the necessary tech (Steel is looking much smarter now that I realized the need for Printing Press, but Cannons can do in a pinch until you can fill up the rest of the techs), you'd lower you're science and build up enough cash for the upgrades. Then you're looking at a mixed force of muskets/macemen/cannons.

Of course, we could just be talking about different styles of games. If I was isolated, or I hadn't been very active on the war front up until the choice I described above (which you are correct, isn't really much of a choice if you didn't have any reason to build Maces), then I would probably go for Education because I won't be able to generate enough wealth to upgrade, or have anything worth upgrading for that matter.

But in a game where you've had an early war or two and build up, I think one could make a compelling argument in its favor.

 

[Lalabunny]

I was sort of implying though that once you got close to the necessary tech (Steel is looking much smarter now that I realized the need for Printing Press, but Cannons can do in a pinch until you can fill up the rest of the techs), you'd lower you're science and build up enough cash for the upgrades. Then you're looking at a mixed force of muskets/macemen/cannons.

Of course, we could just be talking about different styles of games. If I was isolated, or I hadn't been very active on the war front up until the choice I described above (which you are correct, isn't really much of a choice if you didn't have any reason to build Maces), then I would probably go for Education because I won't be able to generate enough wealth to upgrade, or have anything worth upgrading for that matter.

But in a game where you've had an early war or two and build up, I think one could make a compelling argument in its favor.

Yup, that was precisely my point . If you're gearing for an early war, machinery/guild/engineering seems more attractive. Use Treb/Mace/Knights to win an early medivel war. Delaying Philosophy/Education or even losing Liberalism will not cost much provided you won enough land to recover later in the game. Beside your cities will be busy building units to even care about universities. Philosophy does not look as attractive now as Pacifism will murder your economy with all the unit spamming.

Trying to get Economic first for GM looks much appealing now as a Trade Mission will give you enough gold for recovery or to upgrades all your mace veterans to Riflemans (or treb to cannons) should you decide to continue with the war mongering.

My point was that if you plan to forgo Lib race, make sure your alternative plan gives you enough in return. Then again, not everything is set in stone, sometimes it depends on maps/neighbours/diplomacy. Even when you try to play peaceful but you are box in by psycho like Shaka/Alexander, early military techs are indeed essential for survival.

 

[Htadus]

Htadus: When you say commerce city, you mean wealth city, right? Depending on your empire, you might not even want a city dedicated to wealth. I usually use cities that don't have anything else pressing to build accumulate wealth. A more common strategy is to build a forge in a cottage city, then have it build wealth. This second method takes advantage of specialization just as well as having them in seperate cities, with the added benefit that they can actually build something if they have to (like a uni). A forge gives you just as much a wealth advantage as a market, and you won't have any grocers for a long time.

That common startegy will require the cottage city to work a mine or other hammer tile. I would consider this a poor use of a cottage city. I would rather run a scientist in place of the hammer tile since the scientist will get at least 3 beakers and get the multiplier.

Also the commerce cities I refer to will generate a GM or two in the bigining and at the late game when there is a 2X hammer multiplier is available can become a hammer city with workshops and what not. Never the need to bulldoze over mature a town. Also in a science game a grocer is very needed for health.