To build or not to build?

[stormerne]

This is "must know" information for all civfanatics.

All the following analysis is credited to Andrew Kanaber of Queen's College, Cambridge. I've distilled it to the essential figures civfanatics need to know, and I've added the Wonder note. Andrew's detailed mathematics can be found at http://www.adk24.ucam.org/civ2.html

<u>When to build Libraries, Universities or Research Labs?</u>
Boosting your science output by building these city improvements doesn't always work. The improvements take gold to maintain and if you build them at the wrong time you have to increase your tax rate and decrease your science rate! Of course they do have the advantage of being selective. The following are rules of thumb for when it is advantageous:
* Libraries 3 or more beakers, or when you have the Adam Smith Wonder.
* Universities 12 or more beakers.
* Research Labs 15 or more beakers.

<u>When to build Marketplaces, Banks or Stock Exchanges?</u>
This is simpler:
* Marketplaces 2 or more gold, or when you have the Adam Smith Wonder.
* Banks 9 or more gold.
* Stock Exchange 16 or more gold.

<u>When to build Temples, Colosseums or Cathedrals?</u>
Removing unhappiness by building these city improvements isn't always the best thing. Sometimes it's better to increase the luxury rate, those these improvements do have the advantage of being selective. The figures below assume the a cathedral is a "3" effect, that is any -1 effect of Communism has been negated by the +1 effect of Theology, and that you do not have a Fundamentalist government:
* Temples are always beneficial, especially when you have the Adam Smith Wonder.
* If your city has a Bank: Colusseums and Cathedrals work as well as luxuries.
* If your city has a Stock Exchange: Luxuries work better than Colusseums and Cathedrals!

<u>When to build Factories?</u>
This is simple too:
* Factories 5 or more shields

 

[shadowdale]

Hmm well that's all good and fine to know, but some how I can't see me going over all of this every time I need to build a new improvement in my city - I don't really look at the "when to make a revolution chart" because if I do and I have to wait I simply forget to change my government.

And all of this is great to know but I don't think that I'll be using it very much - I don't want my Civ games to end up being algorithms!!! I want to play and not let all those math guys tell me how to play!!

<IMG SRC="http://forums.civfanatics.com/ubb/snipersmilie.gif" border=0>

Run you bloody mathematician

------------------
Veni Vidi Vici.

Coolbook: Cunobelin Of Hippo, Håkan Eriksson, vladmir_illych_lenin, stellar converter, Stormerne.

[This message has been edited by shadowdale (edited March 09, 2001).]

 

[stormerne]

You don't need to go over it everytime you build an improvement! They're rules of thumb, not algorithms

These things are useful for beginners who wonder why they're worse off when they've built universities in all their cities producing just 4 beakers! And a good player might just play that bit better...

Anyway, I'm certainly not going to tell YOU how to play. You'd have my nomination for current best player versus the AI anyway.

(But I'm still going to knock you off the top of the King level HOF. Just another 170 turns...

)

------------------
"Life shrinks or expands in proportion to one's courage" - Anais Nin

 

[shadowdale]

Well I'm looking forward to seeing that - just so that you know, I did improve that game a little, I just don't know if it will get posted. And I am currently trying to beat Kolumbus myself, I just need to know I I can use the larger map patch in a hall of fame game. You know the 32.767 size, which BTW is 151*217!!

BTW how is your GotM going??? I already finished mine, after a loooong weekend, I think that it took me about 30-35 hours of gaming to do.

------------------
Veni Vidi Vici.

Coolbook: Cunobelin Of Hippo, Håkan Eriksson, vladmir_illych_lenin, stellar converter, Stormerne.

 

[stormerne]

At the risk of going off my own thread topic...

Well if you improve it, surely it wouldn't count, as it would have to have been restarted from a save-reload

Anyway, you'll see mine in due course. It won't be far beyond yours, but it will be beyond!

Why do you need a bigger map? Isn't 75xwhatever (standard large map option) big enough? The problem as I see it is that you can't build more than 255 cities, and the standard large map is plenty big enough for 255 cities.

I've been looking at some of the saved games in the HOF - not yours unfortunately because it's the wrong version - and there's often room for BIG improvements (including on my own very naive Prince entry). In fact I'm thinking of writing an article for the war academy about how to boost your score by 600 to 900 points - easily.

My GotM, hmmm.... I started very badly, and played entirely the wrong strategy for the terrain, map size, level and AI civ placing. By the time I found a decent strategy and got myself out of the mess, it was halfway through the game turns

So I've learnt a lot from that. I think I shall be lucky to win the game, let alone get a decent score. I may have to resort to my last ditch spoiling tactics of wholesale disruption (like massive deliberate pollution, and spies everywhere destroying wonder builds) but it would be a shame to sink to that.

 

[TimTheEnchanter]

I think these are a good start, but I have a lot of problems with some of the simplifying assumptions he made.

The biggest assumption is that you are at the "right" tax rate to begin with. In reality, (at least for the first part of the game) I think we often find ourselves trying to squeeze more science out of our trade even when we have decent gold coming in, but the limitations of our government does not allow us to do so. In these cases, a university might make much more sense earlier than 12 beakers. It allows you to "Buy" a higher science rate than your government allows.

I think it also ignores the value of selectivity of improvements by city. His model looks at tax rates that affect the entire civ to determine whether to build improvements. But if I've got 10 beakers in a couple cities, I'm not changing my entire tax rate to get the equivalent of 2 universities, if I can afford to build them.

He doesn't incorprorate the "value" of luxuries in his study of Bank/Stock exchanges. Just like Tech improvements allow you to reduce your tax rate, you could in theory reduce your luxury rate and make more money to pay for the improvement. I'm not sure why he simplified the model in this way, but it seems inconsistent with his other analysis.

I guess they are good to keep in the back of our minds, but I think he undervalues many of the improvements because of some of these assumptions.

------------------
There are some who call me...Tim

 

[stormerne]

I agree with you Tim. He has used simplifications and that makes his findings rules of thumb at best. And there are many special cases such as you mention where an alternative strategy is best. The secret I suppose is to know when to break away from a general rule and why, and to be able to do that you need both experience and a general rule!

One of the many things he doesn't consider is the "savings value" of the science improvements. For example, at the start of the space race, when you've got all your tech necessary to build all the SS bits, and if you need every bit of speed to make sure you still win, then selling all your science improvements over two or three turns gives you a big cash boost for rush building.

I think these are just a starting point and people may successfully wander off a long way from this point. But I haven't seen any kind of mathematical analysis before of the science and happiness equations. If you know of others, or perhaps can do better yourself starting from scratch, I'll be the first in the queue to read them!


------------------
"Life shrinks or expands in proportion to one's courage" - Anais Nin

 

[shadowdale]

Well the reason I would very much like to play on a bigger map is because then I am sure that all my cities can get a good spot. But also because I want to try some thing out. There has been a few postings about having more than 255 cities, and that has made me wonder a bit, since I know that you can't build more. But what if you can get more cities but conquest - you build 255 cities and then take, say 75 cities from the AI that would give you a considerable advantage in the Hall of Fame competition. And to do this in a real high score game, would be a lot easier in a huge map like 32.767!!!
The thing about my game was simply that I didn't know you could boost you population with celebration, so I loaded the game just as I launched the space ship and played the last few turns with the lux rate at 80%-100% and got out a whole 6% more.
But one thing is sure, after reading a few things in the war academy that I didn't know, I'm sure that I could get a better score on both King and Emperor if I wanted to - So if you beat my record then I'm going to have to beat yours!!

------------------
Veni Vidi Vici.

Coolbook: Cunobelin Of Hippo, Håkan Eriksson, vladmir_illych_lenin, stellar converter, Stormerne.

 

[stormerne]

That's an odd thought about the city limit. If it were true (and I have grave doubts) then it would need a completely different strategy. You'd have to let - even help - other civs grow and not take them over till late in the game. I quite like that strategy anyway, especially on a large map, and the one I'm playing now involved the sudden demise of all other civs - about 90 cities - in just a few turns around the late 18th century. But you'd need to wait even later than that.

<u>If someone knows how to get more than 255 cities, I want to know now!</u>

And as for your HOF High Score...
--- --- "Your excuse for a civilization makes us laugh!" --- ---
Challenge! I shall be very slow and careful and maximise everything and give you a target that is so high you'll take months of game-playing to exceed it. Bwa-ha-ha-ha-ha!

(BTW, I used standard large map, large random archipelago land mass.)

Did you notice the King one is just 1% behind Tap-Long Duong's on Apolyton?

------------------
Falla forsar, flygr orn yfir, sa er a fjalli fiska veiðir
Erne flies over stormy falls, hunting fish upon the fells
- Voluspa

 

[shadowdale]

Yep I noticed the Apolyton, and that was actually one of the reason why I wanted to se if the "lux rate=population growth" would give me a slightly better score, which it did. But if you beat it anyway, then I'm just going to play a new and even better game - I already have a very good idea of what is missing from both my Emperor and King game.

------------------
Veni Vidi Vici.

Coolbook: Cunobelin Of Hippo, Håkan Eriksson, vladmir_illych_lenin, stellar converter, Stormerne.

 

[Andu Indorin]

Andrew Kanaber's analysis is quite good. But we must remember that his analysis does not include the effects of Wonders and Trade Routes, which tend to be more cost effective.

For example, he says that Temples are always beneficial. Yet I have not built one in any my recent games, building Michaelangelo's Chapel instead.

Nonetheless, his analysis does provide good rules of thumb for when to build and when not to build.

 

[Lefty Scaevola]

Originally posted by shadowdale:

And all of this is great to know but I don't think that I'll be using it very much - I don't want my Civ games to end up being algorithms!!! I want to play and not let all those math guys tell me how to play!!

*note to self: accelerate timetable for eliminating illogical carbon based intelligence*

------------------
Gauis Mucius Scaevola Sinistra
Pontificator Pedanticus
Older, richer, and wiser than you.
Did I mention that I love the spellchecker?

[This message has been edited by Lefty Scaevola (edited March 09, 2001).]

 

[Archbob]

Libraries, Universities, and Research Labs are always beneficial when you have more than 2 beakers.

------------------
Fallen Angel Lord,

check out Civworld Forms

 

[vladmir_illych_lenin]

Hmmmmm... I already knew most of this. I take this into careful consideration when my civ is small, but when I have more than 50 cities, I just build something when it has a benefit, eg I build a market when I have 2 gold, library when I have 2 beakers, etc.