Behavioral economics

ksa76

Chieftain
Joined
Aug 14, 2010
Messages
42
Behavioral economics deals with irrational and emotional decisions and the consequences for returns and resource allocation. The field is primarily concerned with bounded rationality and integrate insights from psychology and neo-classical economic theory. Some common themes of the field which may be of interest to civ-players are 1) the time preference hypothesis, 2) self control mechanisms, 3) the sunk cost fallacy and 4) the winner's curse.

The time preference hypothesis is the notion that people tend to have an ill-conceived preference for immediate over future returns. Experienced civ-players have learned the value of not rushing at first turn (50% penalty), overchopping forests pre-Math and pre-forges (50% and 25% penalty), and building Wealth in cities without forges (25% penalty). With regard to great people, don't start your first golden age prematurely, if you have a valid chance of completing the Mausoleum. Similarly, great people should not be "wasted" on bulbing cheap technologies, to rush cheap wonders (great engineer), for early trade missions (great merchants) or for Academies in secondary science cities (great scientists). Since the cost of subsequent great people is increased, don't commit too many specialists for great leader production too early, i.e. pre-National Epic and pre-Parthenon.

As I used to do, I suspect many players are often working the slider to just grab a new technology. However, if the new tech has no immediate benefits (it is often better to delay it in order to maximize the 20% research overflow bonus for the following tech.

Of course, many civ players may equally overestimate the long-run benefits of certain decisions. Some examples of these are
• spreading missionaries way in advance before shrines have been built. Once the shrines are in place the automatic spread is tripled.
• minimizing city overlap for the possibility of future super high population cities.
• building of happiness and health buildings way ahead than needed. It will often be better to build Wealth (especially if no production/research/commerce multiplier buildings are available)

Self control mechanisms is the notion that people often rely on non-optimal rules of thumb to make decisions easier. However, at higher levels, you need to adapt to the situation rather than adhering to the 60% expansion rule, getting 1,5 workers per city, always granary first and following strict building orders.

The Sunk cost fallacy is the notion that people tend to have strong misgivings about "wasting" resources (loss aversion). Sunk costs have already been incurred and cannot be recovered. The bygones principle argue that historic costs are irrelevant - the only relevant issue concerns future costs and benefits. In other words, don’t throw good money after bad. While fail gold from wonders is a valid strategy (especially for industrious civs), other civs will be better off by shifting to building on wealth or infrastructure if a certain wonder seems out of reach. The sunk cost fallacy can also work psychologically in terms of future committed resources. I suppose some players wrongly decide not to build coal plants if they have already committed themselves to getting the Three Gorges Dam. Even if beelining Plastics (a questionable decision), many cities may benefit greatly from the +25% hammer bonus from coal plants until the Three Gorges Dam is complete.

The winner's curse is a term derived from auctions and describe the tendency for the winning bid to exceed the intrinsic value of what is being offered. In short, the winner's curse says that the winner will tend to overpay. With regard to wonders, people tend to underestimate the benefits required to justify an investment. This is especially the case for small empires where the critical mass of empire benefitting wonders cannot justify the costs.
 
Very interesting read, I had all of your "flaws" and now have none of them :) (not related to your article but to normal developement of my game, so I agree with what you write and don't make fun of it) .

Think this is very well written though a little hard to understand sometimes because it's really condensed, which I value though as I have problems with redundancy.

Maybe some mod will move it over to the Strategy and Tips Forum or even to the Strategy Articles section?

Btw., on which lvl are you playing, we just started on a new project that might interest you.

Regards, Sera
 
Thanks Seraiel.
I am a deity player (on the Earth map at least). I especially linked your contributions for achieving high winning scores . I am also very interested in this new project.
I would be pleased if this post found its place in the strategy and tips forum / the strategy article section.
 
Moving your post I cannot do, but sending you an invitation to the user-grp that works on a current series of new Strategy Articles I can :) Accept it and look aorund if you're interested, don't forget to bring in your ideas :) .

Cya, Sera
 
The time preference hypothesis is the notion that people tend to have an ill-conceived preference for immediate over future returns.

I don't see how you can justify these at all. This game loves immediate returns. When I load a save by someone asking how to salvage their situation, they are usually hiding out among unmolested forests loading their cities up with buildings. Granary is the only thing I see a use for everywhere. Beyond that the game is about workers, settlers, units, and wealth. Wealth is great, and forges are for hammer cities. As for GPs, getting one asap (say, whipping a library in your capital or a food site) will get you out of the early game quicker than waiting for some kind of optimal condition or late wonder. Bleh to that.
 
Hi Tristan. Thanks for your comments on the post.

You make some valid points about the advantages of immediate returns for the early game. My general advise is based on the mere notion that there is a risk of overchopping especially if you are going to build some expensive wonders (say Pyramids og the Great Lighthouse). Equally there is a risk that one for long periods builds wealth rather than key infrastructure.
I always build a forge in all my cities. This, I believe, goes well hand in hand for a long game strategy (Domination/space race win) whereas for a mid-game conquest victory the merits of this approach can be debated.
As for GPs, I agree that whipping an early library for two scientist specialist is a valid strategy I am just stating the obvious that there is a risk of over-commital to great leader points in the early game.
 
I don't see how you can justify these at all. This game loves immediate returns. When I load a save by someone asking how to salvage their situation, they are usually hiding out among unmolested forests loading their cities up with buildings. Granary is the only thing I see a use for everywhere. Beyond that the game is about workers, settlers, units, and wealth. Wealth is great, and forges are for hammer cities. As for GPs, getting one asap (say, whipping a library in your capital or a food site) will get you out of the early game quicker than waiting for some kind of optimal condition or late wonder. Bleh to that.

Sort of. The problem with the OP's hypothesis is that it seems to assume only an immediate set of costs/benefits, not longer-term ones. For instance chopping out your first settler (pre-maths) effectively loses some hammers. But if it also means you get a second city up quickly which secures a strategic resource or boxes in an AI then those gains easily outweigh the lost hammers in the longer term. If you wait longer to do something, you may get more - but by then the game situation may mean that you need more anyway. Time is money (and hammers, and beakers) after all.

IMHO the issue is not so much about immediate returns as newly available ones. Building markets everywhere because you've just researched Currency is dull. Researching at 100% because some tech sounds impressive is silly if you're preparing for war and should be upgrading / rush-buying units. Switching to Universal Suffrage makes no sense if you've few towns and/or need the happy bonuses from HR / Rep. In all cases, pre-existing choices are often better. In other words there's maybe an illusion that newer, more advanced options are better options; that winning is about following a linear path of 'progress' as fast as you can. As you say, it's at least as much about doing a limited and relatively fixed set of things consistently and ruthlessly.
 
Its all about "how long game will go in general".. If I know that game can be done in short time (with correct steps).. than I will try to target for fast benefit... Bulb some key techs to get advantage than chop out army (whatever if there is forge/math) to get some enemy down before it gets too big for fast attack...
Than again.. if you learn map/enemies.. and find out that your current position is too weak for fast moves.. Take long shot and plan ahead to be ready at some "breaking" point (for example set game for rifle-cannon rush.. or even tank-air attack..)..
It will be different game and strategy if play normal size map with 6 AI on 1 big land or Huge with 17 AI on many continents...
 
According to game theory, the optimal play depends on whether or not you expect the other players to react to your move. For example, the prisoner's dilemma has a different solution if you play the game over and over again. As long as the players believe that there will be another round of the game, they will have an incentive to cooperate. They know that if they look out just for themselves, the other players' optimal choice will be to turn on that player.

On the other hand, if one player realizes that this round is the last, she will win more by breaking the agreement.

The time preference hypothesis does apply, but not in the direction that you claim here, in my experience. Economics usually expects an exponential return on investments (i.e. you invest some money, you get it back plus a little something for your trouble. You invest what you invested the first time plus the interest from last time and get a little more interest on the whole thing, and so on).

Humans, however, don't usually think of things accumulating exponentially. They usually think of investments in terms of simple interest--if I invest X dollars and get (X+Y) dollars back, Y is my rate of return each time I invest (forgetting that the next time I could choose to invest X+Y instead of just X)

That means that people tend to underestimate short term advantages. Your arguments for why they might not are artificial controls that the programmers threw in to keep things under control.

Rule of thumbs work to varying degrees, as long as one understands the why's and the wherefore's. Still, good point. The only problem is that they are hard to replace (except by another rule of thumb). There are a lot of unknowns in this game (more if you're just starting out), in which case a good heuristic is invaluable. So guidelines are good while information/ experience is limited.

I can't think of any good examples of the sunk cost fallacy, but loss aversion is a good point. Loss aversion might include things like being unwilling to use your highly promoted unit when you might lose it or being unwilling to chop down a forest. You can become paralyzed by the potential, future losses that you might incur instead of the immediate losses that you will definitely incur.
 
This has been a very illuminating thread, involving multiple wiki searches. From what I can tell, Um seems to be on track with the "underestimating short term gains."
For example, one could extend the OP's idea generally not to chop until post mathematics/forge to "not until factory/power plant for +100% chop yield" and no one would agree with the concept (for many reasons; un-upgraded forest tiles all over the place, a math/forge chop for a maceman might be much more useful than a late chop for most of an infantry etc). If that pre math chop gets my settler out earlier, then it settles earlier, meaning more turns of working tiles. This may produce enough food/commerce/hammers to make up the loss of hammers on its own; however, it also has that hard-to-quantify increase in "power projection on the map" (be it "power" in the form of resources secured, units now produce-able, or increased proximity to rival powers -- which reduces travel time and thus hastens war-actions).
I may agree with GP's though. Saving some for the late game golden ages can be inordinately powerful. Even then though, a golden age's strength is a function of your own empire's strength, meaning it is powerful when you are already powerful, so an earlier bulb might have netted you more. Anyway, the thing to remember when you are throwing away future potential, like early chopping and bulbing, etc, is to "CASH OUT" ('leveraging') when you do. If you are bulbing rather than settling/golden-aging etc, then it should be because you intend to isolate a wonder, make a new unit type no one/enemy has, or make trade bait. If you are chopping to make units, it is for settling or taking cities, not for putting another longbowman on the border of your 'friendly' ai neighbour. As long as you are always "cashing out" on your potential future losses, then you are probably ok.
 
I particularly like Um the Muse's comment about the time value of the different game resources.

Just one minor point I would like to add - as a general rule I try to avoid the "cold whip" or "cold buy" penalty, but one exception that I ALWAYS make now (and feel stupid for not having done so before) is the first Civ executive of any corporation I found (especially if it's Cereal Mills).

The following turn I gladly take the "cold buy" penalty twice (forking out 900 rushbuy gold) and the turn after that, I do it four times.

The effect of one "corp turn" accumulates over the remainder of the game.

============

On another note, at the start of the game, I always switch to Slavery while my first settler is in transit. This effectively puts that particular city (and all "children" of that city) one turn ahead. This early-game decision effectively means that one half of my empire is a turn ahead (compared to an alternate situation where I had not correctly timed my civic change at the start of the game). This simple decision, though seemingly trivial at the start of the game, could potentially represent a swing of 300 hammers and 1000 beakers (i.e. half the "per turn" output of a typical late-game economy).

Fibonacci
 
Top Bottom