Game theory for game practice

So the plan now for this blog is to become a storage space were I post things I find relating to my chosen dissertation topic, and approach what they say and how this is relevant. The first of these is an article from the economist, found here.

The article discusses game theory, and its use in software to attempt to predict the outcomes of many varied situations. The idea behind game theory is to assume that everyone acts in what they consider to be their own best interest, and to combine this information with intelligence on what resources are available to a 'player', to predict the actions they are likely to take, and what effect this will have on others.
Amongst the interesting points raised in this article is the general suggestion that this forecasting software already exists and is in use. It even suggests that "most decision-making advice is political, in the broadest sense of the word". This to me hints that there is an acceptance of a value to the sort of information I hope this game would produce. It also shows that this forecasting software is applicable in many different theaters, it is not limited strictly to the politics of international relations, but has value to a great many different areas. As mentioned in the article "PA Consulting, a British firm, designs bespoke models to helps its clients solve specific problems in areas as diverse as pharmaceuticals, fossil-fuel energy and the production of television shows.". One lesson I believe I can take from this is that in designing the game to be as close to real life as it can, I expect I will find it very easy to neglect certain areas in favour of others. For example, to design the game with as much freedom between players (or player states, one way to make the game more life-like is to organise the players into states, much more on this later) in the field of arms trading would make sense, given the huge preponderance of war games, and games based around conflict. This makes sense in that it is following what is known to be a successful recipe for previous games. However it would be easy to neglect trade in white goods, as I have never heard of a game, let alone a successful one, that revolves around the race to create the best washing machine. What this essentially means is that every compromise I make in making the game less true-to-life in exchange for making it better or more efficient in other ways, degrades the value of the information it will yield, to washing machine-making companies, for example.
One crucial difference between my proposed game, and the game theory software is summed up in the flaw highlighted by "Reinier Van Oosten of Decide... notes that forecasts go astray when people unexpectedly give in to "non-rational emotions" such as hatred, rather than pursuing what is apparently in their best interests.". This difference is the basic use of people to fill the role of people, rather than the game theory assumption that they will follow what they perceive to be in their best interests. This does not solve the problem entirely. There is no reason to believe that people in-game will genuinely feel "non-rational emotions" to the goblin that just shot them, as it is to be supposed (hoped?) that the majority of the user base will be able to differentiate between real life and the game. However the actions of real users in game, it can be argued, will still be more indicative of the actions of people in the real world, than the predictions put forth by the game theory software. This then would have an effect upon the market to which the information produced by the game is valid. It is mentioned in the article that the game theory is especially useful in matters concerning economics, as a person focusing on profit and the acquisition of money is less likely to be waylaid by non-rational emotions. This then suggests that a system based on the input of 'the crowd' will gain some validity through this, increasing its value in many areas, as a result of greater resistance to the difficulty of predicting emotions and their consequences.

Introduction

I don't think I am the sort of person that recognises subtle hints very readily. The reason I think this is because I have decided to write my dissertation about crowdsourcing, yet missed one very obvious conclusion. Crowdsourcing may best be explained by comparison to the term 'outsourcing'. A company is outsourcing when it meets its requirements of information and labour by establishing its new source elsewhere. For example, a British company that sells double glazing over the phone, that moves its call centre to Bangladesh has outsourced that requirement for labour. In the same way then that outsourcing means to gather a resource from outside the local area (be that city, county, country), crowdsourcing means to gather a resource from a crowd. This is something that has been made possible by the spread of the internet. A prime example of this would be Wikipedia, in which a vast amount of information has been gathered together onto one site, freely given by internet users. Those who write articles on wikipedia do not get paid for their labour, rather, it is done to further a personal interest, and it is this free resource that crowdsourcing capitalises upon. The obvious conclusion I mentioned at the beginning, is that despite planning to write my dissertation upon a potential application of crowdsourcing, it has only just occurred to me to blog about the things I learn in researching this idea. If crowdsourcing is the process of using the freely given information from the internet, it seems only right that the information I gain about the subject be put on the internet freely.

Can a computer game be made that is close enough to real life that the information gleaned from players as they play the game be of use in a wide range of areas, yet the game still be an appealing and marketable product

The aim of this blog is to document things I find whilst researching my chosen subject, and to provide a place to keep it all. The precise subject area I want my dissertation to discuss is that written above. To explain it a little better, I'll use some examples from a pre-existing game; World of Warcraft. The first example is an event known as Hakkar's blood plague (video on the right). A glitch in the game allowed a damaging attack used by one of the players most powerful opponents to escape from the confined area in which this opponent existed, and affect the game world as a whole. The attack, called Hakkar's blood curse, caused an amount of damage to a player at timed intervals, furthermore it was spread by proximity, meaning a player standing within a certain radius of an infected player would also contract the disease. Another interesting dynamic of this disease was that it dealt a fixed amount of damage, as a pose to damage proportional to a players level. The levels system within World of Warcraft at the time of this event ranged from 1 to 60, level 1 players being the weakest, and 60 the strongest. What this meant is that the damage caused by the disease was survivable by a high-level player, but was very destructive when the disease was contracted by a low-level player. These characteristics are all comparable with characteristics of diseases in the real world. The spread of the disease by proximity mirrors not simply airborne infections, but also had the effect of making cities and areas of high population hot-spots for the disease, whereas it was easy to avoid in less populated areas, out of the cities. Furthermore the variation in levels in game mirrored the varying levels of health in the real world. A high level character, in the same way as a healthy person, would suffer as a result of the disease, but would not be killed, a low level character would likely be killed outright when infected, which can be compared to at risk groups in real life, such as the elderly, infants and people with pre-existing medical conditions. The effect of this high level of comparability with real life made the information recorded by the game valuable to real world organisations, and information from this in game event has since been used to better understand human reactions to pandemics. This goes so far as anti-terrorism officials studying the case, as some players would intentionally seek to perpetuate the epidemic.

I never experienced the events of the blood curse myself, but one experience I did have as a player that I believe holds similar consequences regards the existence of an in-game economy. It is possible for players to produce items from reagents, these items can then be sold (with some exceptions) or used by the player. I discovered that a certain very low level item could be made from cheap reagents, and then broken down into different resources, that sold for more than it cost to buy the original reagents. This in effect meant that I could buy certain resources, then convert them into more valuable resources to be sold. I started this with a limited amount of money, and succeeded in turning a profit, which I would use to buy more reagents, resulting in more product to be sold. This is a process I maintained until it became apparent that the market for the finished product was being flooded. The amount of product available was greater than the demand for it, meaning the buyers were able to exercise more control over the price they paid for the product, which was driving the price down. This meant the process of changing one resource into another was not as profitable, and caused me to abandon it. What this meant was that the market in the game reacted in the same way to certain criteria as a real life market would do. 

Whilst both of these examples show ways in which an online game mimicked real life, and as a result can be used to provide information useful in understanding how humans would react under certain circumstances, the in game versions of these events are not absolute copies of a similar real world event, and as such, they information they produce cannot be completely reliable. This to me leaves room for a game that seeks to remedy many of the differences between the real world and the game world, as to allow for the generation of information that is of as great a value as possible. To support this with an example, certain items used as reagents in the creation of other items are only available from in game vendors. A vendor is an NPC, or non-player character, it is an AI character that sells goods to the players at a universally constant rate. The limitations this imposes are as follows; a vendor cannot sell out of most of these items (the only exception to this is items that are available from other non-vendor sources), the vendor cannot alter the price of an item according to demand or supply, it is constant. Furthermore, a player cannot create most items available to their profession without giving money to a vendor. These vendors, as far as I know, do not have a personal bank, that is when a vendor takes money in exchange for an item, it is not recorded in a personal bank, and when a vendor gives money for an item sold to them, neither is this. This means that a vendor can never decline to buy items from a player (though it is always at a rate that favours the vendor) and that money can essentially leave the games economy, or be introduced into the games economy by a vendor with apparently bottomless pockets. These in my mind flaw the capacity of the games internal economy to act as a model by which we can take information useful in forecasting human reactions to certain events. It is therefore my aim to create a game that is as true to real life as is possible, whilst allowing for the fact the game must still be marketable, that is to say, the people it relies on to provide this information are still tempted to play the game because it is enjoyable. This is the essence of crowdsourcing, people are inclined towards working, and working hard when it is made available by the internet, as is proven by examples such as wikipedia. the fruits of these labours lie in the information it provides, and I believe a game that fulfills this design brief could be made to gather this information.