Geeky economics Net Present Value question

macrobot

Chieftain
Joined
Jun 5, 2006
Messages
38
I was thinking over the random events in BTS and it occurred to me that they have always been a part of the game since a mine can discover a resource in all versions of the game. In trying to figure out the net impact on a game I tried to calculate the Net Present Value of discovering a gold mine.

My question is what is the equivalent Random event in terms of free money to discovering a gold mine at say turn 60 of a 660 turn game (Epic speed). THe nominal value is easy if you assume the tile will be worked for 600 turns it is worth -600 hammers and +4200 coins (and 1 happiness which is too much for me at midnight). However, if someone offered you 4200 coins at a cost of 600 hammers on turn 60 you certainly would sell the mine because having money now is better than having money later. However, at what price would you sell the mine or how much cash would need to be given to have the same effect on the game?

To start the calculation you first need to determine if the value of the gold compounds over time. Even this is not an easy question. The income stream remains constant and with inflation even effectively goes down. However, the value increases each turn due to your ability to tech faster than you otherwise would. I think the overriding factor is the teching advantage and the value should compound over time but this point is certainly debatable.

Secondly an appropriate discount rate needs to be determined to account for the fact that the value on turn 600 is less than the value now since there is no guarantee you will get to turn 600 and the gold should be a relatively smaller part of your economy then. Because the discount rate compounds over time the rate can at least be constant over time.

Any thoughts on this in the abstract or more specifically on how the random events influence the game?

The following is just some incoherent ramblings about the economics of civ that I was using to clarify some things I have been thinking about. Any comments are welcome.

Net Present Value (NPV) is a way to convert a future stream of income to a specific dollar amount allowing you to make an apples to apples comparison of what are really 2 different things. In the context of a civ game basically every decision made is an NPV calculation.

For example do you work the extra food to grow faster or do you work the mine to get a unit out faster? Basically one will have an immediate benefit that declines over time until it is even at which point the other strategy is even and then becomes more beneficial.

In this context you can think of food money and hammers as essentially the same thing since over time investment in food yields hammers or money and in turn hammers and money allow for more food by allowing more population either from buildings, techs or both. Fundamentally when you are managing your city you are trying to maximize this value.
 
Isn't one of the challenges of playing Civ that these resources (food, hammers, gold) are not always interchangeable - depending on your civics (Civ IV) or the version of Civ that you are playing?

Whilst you may be able to perform an economic analysis (well done, I know little of economics!) does this analysis help your Civ playing?
 
A couple of random comments

you are right to conider the game as a series of game turns rather than a number of years. The number of years a turn takes varies throughout the game but the output that can be achieved in a single turn is constant

On the question of whether the value of gold increases over time - in real life the value of money or of any other commodity does usually decrease over time - it is the value of investment which increases.

you might be right to assume that the main reason why investments grow is that they allow extra resources to be invested in technology (and technology allows you to increase outputs from a fixed land area)

However, there are other forms of investment which produce dividends - buying troops could be an investment if those troops conquered new land (they may even faciltate discovery of a new gold mine).

wonders and buildings are also an investment because they increase outputs

What therefore is the average rate of return for a player ?

I think it depends on the strategy you adopt - if you are determined to have an axe rush - you have a high expectation that investment in axemen will produce a quick victory - you won't invest in other commodities unless they also have high returns

however if you want a cultural victory you might expect a lower rate of return for most of the game

the average return might be 7% (this is a guess)

so £1 promised next year is worth 93p today. £1 in two years is worth 86p today

you need a calculator to work out how much a future income stream of 6 coins per turn for 600 turns is worth at current prices (ie it NPV)

Civilisation is a historical game and it is to be expected that the rate of return varies over time. The game designers clearly understand that growth rates increase in the mediaval and modern periods. This is mainly simulated by ensuring that at certain points each turn takes less years. At the end it is one year per turn - at the start 40 years per turn. So in the modern period growth speeds up - but the effect is not strongly felt because the output in each turn is what is important and that is relatively constant.

finally the player's preference for short term returns will increase as the game progresses - different players aim to finish the game at different points - an axe rush might mean that the game ends fairly early. what this does mean is that players who are about to finish the game are reluctant to embark on building things like historic wonders with a long term payoff. In other words their expected rate of return increases

unfortunately - it may well be that none of this information will radically improve your game playing
 
Are you trying to find an equation like the Black-scholtz (don't remember the name exactly) equation? Something like an exponential decrease in the value of a future return. Or is this just another way of looking at it.

Empirically, you could run simulations of what you could do if you started with a certain amount of gold. Second you could look at various situations of investing:
Normal city growing can be modeled as an asymptotic increase to max production at a certain happy cap.
Having workers is like an asymptotic multiplier on total production, something like A+B*(1-exp(-C*#workers*t/empire size)).
Building a settler initially increases production by 1P,1C, initial tile (say 2 food 1 P, or 3P for another settler). You pay in cost and delayed growth. Upkeep's annoying, though.
Some technologies are instant multipliers, like printing press or biology. Buildings have a delay on their multipliers.
Later on, wars to take a city are unit cost spent to take city (of course, there are strategic gains). Evidently, this is can't be very efficient compared to settling on your own, so there's some cap on how large your empire can be landwise that can be changed by producing military.

So maybe something like worker multiplier*growth equation to happy cap*growth to land, where growth to land is affected by territory size, and that can be increased by production.
 
It's an interesting idea, but I think that it might be overworked. Basically it boils down to long term benefits/costs vs. short term benefits/costs. I think everything in Civ applies to this, and a simple equation, while enlightening in certain situations, isn't likely to really impact the gameplay.

What is the trade off between working a farm to grow your city or a mine to build a unit? What if those units that you've built while stagnating take a rival city? How much does that rival city now contribute to your coffers compared to working that farm or cottage?

It's a good way to look at things for an idea of long term impacts, but just like in the real world, it isn't that cut and dried.

However, I will offer right now 600 hammers to anyone who wants to give me 4200 pieces of gold.
 
Back
Top Bottom